volume 37 number 2 (issn 0511 5728) january 2015 west ...by ricardo j. rodriguez and winston g lewis...

88
WEST INDIAN JOURNAL OF ENGINEERING Editorial……………………………………………………………………………….…………….….. 2 An Assessment of The UWI’s Faculty of Engineering Capability and Willingness to Engage in Industrial Collaboration for Innovation ………………………………………………………………… 4 Exploring Potential RF Hot Spot Locations in Confined Spaces of Large Wave Guide Dimensions….. 14 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad Gas Condensates …… 23 Simulation of Irrigation Water Requirements of Some Crops in Trinidad Using the CROPWAT Irrigation Software ……………..……….…….……………………………………………………….. 31 Development and Evaluation of Wheeled Long-Handle Weeder ………………………..………..….… 37 A Robust Regression Approach for Excess/Shortage Spare Parts Cost Estimation .. …………….……. .45 A Mechanism for Cutting Coconut Husks ………………. ………………………….………..………....54 Mechanical Properties of Steel-making Slag Reinforced Polyester Composites .……………………… 63 Rheological Study of Cement Modified with a Lignin Based Admixture ………………………………68 Effect of Dynamic and Static Methods of Compaction on Soil Strength …. …………………………… 74 A Strategic Initiative on Enhancing Postgraduate Throughputs at The UWI St. Augustine Campus ….. 79 Published by: Faculty of Engineering, The University of the West Indies St Augustine, Trinidad and Tobago, West Indies Volume 37 Number 2 (ISSN 0511 5728) January 2015

Upload: others

Post on 01-Mar-2021

0 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

WEST INDIAN JOURNAL

OF ENGINEERING

Editorial……………………………………………………………………………….…………….….. 2

An Assessment of The UWI’s Faculty of Engineering Capability and Willingness to Engage in

Industrial Collaboration for Innovation ………………………………………………………………… 4

Exploring Potential RF Hot Spot Locations in Confined Spaces of Large Wave Guide Dimensions….. 14

A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad Gas Condensates …… 23

Simulation of Irrigation Water Requirements of Some Crops in Trinidad Using the CROPWAT

Irrigation Software ……………..……….…….……………………………………………………….. 31

Development and Evaluation of Wheeled Long-Handle Weeder ………………………..………..….… 37

A Robust Regression Approach for Excess/Shortage Spare Parts Cost Estimation .. …………….……. .45

A Mechanism for Cutting Coconut Husks ………………. ………………………….………..………....54

Mechanical Properties of Steel-making Slag Reinforced Polyester Composites .……………………… 63

Rheological Study of Cement Modified with a Lignin Based Admixture ………………………………68

Effect of Dynamic and Static Methods of Compaction on Soil Strength …. …………………………… 74

A Strategic Initiative on Enhancing Postgraduate Throughputs at The UWI St. Augustine Campus ….. 79

Published by: Faculty of Engineering, The University of the West Indies

St Augustine, Trinidad and Tobago, West Indies

Volume 37 • Number 2 (ISSN 0511 5728) • January 2015

Page 2: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

WEST INDIAN JOURNAL OF ENGINEERING

The WIJE Editorial Office Faculty of Engineering, The University of the West Indies, St Augustine

The Republic of Trinidad and Tobago, West Indies Tel: (868) 662-2002, ext. 83459; Fax: (868) 662-4414;

E-mail: [email protected] Website: http://sta.uwi.edu/eng/wije/

The West Indian Journal of Engineering, WIJE (ISSN 0511-5728) is an international journal which has a focus on the Caribbean region. Since its inception in September 1967, it is published twice yearly by the Faculty of Engineering at The University of the West Indies (UWI) and the Council of Caribbean Engineering Organisations (CCEO) in Trinidad and Tobago. WIJE aims at contributing to the development of viable engineering skills, techniques, management practices and strategies relating to improving the performance of enterprises, community, and the quality of life of human beings at large. Apart from its international focus and insights, WIJE also addresses itself specifically to the Caribbean dimension with regard to identifying and supporting the emerging research areas and promoting various engineering disciplines and their applications in the region.

The Publications and Editorial Board Professor Brian Copeland, Chairman, UWI; E-mail: [email protected] Professor Kit Fai Pun, Editor-in-Chief, UWI; E-mail: [email protected] Professor Winston A. Mellowes, Immediate Past Editor, UWI; E-mail: [email protected] Professor Edwin I. Ekwue, Member, UWI; [email protected] Professor Stephan Gift, Member, UWI; E-mail: [email protected] Professor Winston G. Lewis, Member, UWI; [email protected] Dr. Bheshem Ramlal, Member, UWI; E-mail: [email protected] Dr. Gail S.H. Baccus-Taylor, Member, UWI; E-mail: [email protected] Dr. Richard Clarke, Member, UWI; E-mail: [email protected] Mrs. Marlene Fletcher-Cockburn, Secretary (Outreach); E-mail: [email protected]

International Editorial Advisory Committee Professor Andrew Wojtanowicz, Louisiana State University, USA; E-mail: [email protected] Professor Clive Davies, Massey University, New Zealand; E-mail: [email protected] Professor John E.L. Simmons, Heriot-Watt University, UK; E-mail: [email protected] Professor Kulwant S. Pawar, The University of Nottingham, UK; E-mail: [email protected] Professor Peter Hogarth, Bournemouth University, UK; E-mail: [email protected] Professor R. Lal Kushwaha, University of Saskatchewan, Canada; E-mail: [email protected] Professor Rafaella Ocone, Heriott-Watt University, UK; E-mail: [email protected] Professor V.M. Rao Tummala, Eastern Michigan University, USA; E-mail: [email protected] Dr. Albert H.C. Tsang, The Hong Kong Polytechnic University, China; E-mail: [email protected] Dr. Gordon Dodds, Queen’s University of Belfast, UK; E-mail: [email protected] Dr. Jeffery A. Jones, University of Warwick, UK; E-mail: [email protected] Dr. Kwai-Sang Chin, City University of Hong Kong, China; E-mail: [email protected]

Editorial Sub-Committee Professor Kit Fai Pun, Chairman; E-mail: [email protected] Professor Winston G. Lewis, Vice Chairman; E-mail: [email protected] Mrs. Marlene Fletcher-Cockburn, Secretary (Outreach); E-mail: [email protected] Ms. Nancy Ayoung, Secretary (General); E-mail: [email protected] Ms. Crista Mohammed, Technical Support (Copy Editing) E-mail: [email protected] Mrs. Paula John, Technical Support (Web) E-mail: [email protected] Ms. Genna Bowyer, Office Assistant (General); E-mail: [email protected] To order subscriptions or report change of address, simply fill in the required information on the Subscription Order Form provided at the back of the Journal or downloaded from the WIJE website, and/or write to: The Editor-in-Chief, The Editorial Office, West Indian Journal of Engineering, c/o Block #1, Faculty of Engineering, The University of the West Indies, St Augustine, Trinidad and Tobago, West Indies. Fax: (868) 662 4414; Emails: [email protected]; [email protected]

Page 3: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

WIJE; Vol. 37, No.2, January 2015 1

WEST INDIAN JOURNAL OF ENGINEERING

2 Editorial

4 An Assessment of The UWI’s Faculty of Engineering Capability and Willingness to Engage in Industrial Collaboration for Innovation by Cary R. Cameron and Graham S. King

14 Exploring Potential RF Hot Spot Locations in Confined Spaces of Large Wave Guide Dimensions by Ricardo J. Rodriguez and Winston G Lewis

23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad Gas Condensates

by Raffie Hosein and Rayadh Mayrhoo

31 Simulation of Irrigation Water Requirements of Some Crops in Trinidad Using the CROPWAT Irrigation Software by Edwin I. Ekwue, Rebekah, C. Constantine and Robert Birch

37 Development and Evaluation of Wheeled Long-Handle Weeder” by Silas O. Nkakini and Abu Husseni

45 A Robust Regression Approach for Excess/Shortage Spare Parts Cost Estimation by Sunday A. Oke, Desmond E. Ighravwe, and Gholahan Shyllon

54 A Mechanism for Cutting Coconut Husks by Kishan Ramesar, Chris Maharaj and Umesh Persad

63 Mechanical Properties of Steel-making Slag Reinforced Polyester Composites by Isiaka Oluwole Oladele

68 Rheological Study of Cement Modified with a Lignin Based Admixture

by Rean Maharaj, Lebert H. Grierson, Chris Maharaj and Vitra Ramjattan-Harry

74 Effect of Dynamic and Static Methods of Compaction on Soil Strength

by Edwin I. Ekwue, Robert Birch, and Jared Chewitt

79 A Strategic Initiative on Enhancing Postgraduate Throughputs at The UWI St. Augustine Campus by Kit Fai Pun

Volume 37 • Number 2 (ISSN 0511-5728) • January 2015

The Editorial Office West Indian Journal of Engineering

Faculty of Engineering The University of the West Indies

St Augustine The Republic of

Trinidad and Tobago West Indies

Tel: (868) 662-2002, ext. 83459 Fax: (868) 662-4414

E-mails: [email protected]; [email protected]

Website: http://sta.uwi.edu/eng/wije/

Editor-in-Chief: Professor Kit Fai Pun

The Publication and Editorial Board of the West Indian Journal of Engineering, WIJE (ISSN 0511-5728) shall have exclusive publishing rights of any technical papers and materials submitted, but is not responsible for any statements made or any opinions expressed by the authors. All papers and materials published in this Journal are protected by Copyright. One volume (with 1-2 issues) is published annually in July and/or January in the following year. Annual subscription to the forthcoming Volume (2 Issues): US$15.00 (Local subscription); and US$25.00 (By airmail) or equivalent in Trinidad and Tobago Dollars. Orders must be accompanied by payment and sent to The Editorial Office, The West Indian Journal of Engineering, Block 1, Faculty of Engineering, UWI, Trinidad and Tobago. Copyright and Reprint Permissions: Extract from it may be reproduced, with due acknowledgement of their sources, unless otherwise stated. For other copying, reprint, or reproduction permission, write to The Editorial Office, WIJE. All rights reserved. Copyright© 2015 by the Faculty of Engineering, UWI. Printed in Trinidad and Tobago. Postmaster: Send address changes to The Editorial Office, The West Indian Journal of Engineering, Faculty of Engineering, UWI, Trinidad and Tobago, West Indies; Printed in Trinidad and Tobago.

Page 4: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

WIJE; Vol. 37, No.2, January 2015 2

Editorial This Volume 37 Number 2 includes eleven (11) research articles. The relevance and usefulness of respective articles are summarised below.

C.R. Cameron and G.S. King, “An Assessment of The UWI’s Faculty of Engineering Capability and Willingness to Engage in Industrial Collaboration for Innovation”, discuss the importance of University-industry collaboration (UIC) and develop a Conceptual Model of good practice in UIC which has four pillars, namely 1) high quality academic research; 2) a predisposition on the part of academics to engage; 3) a framework of supporting policies and procedures in the university; and 4) an effective office supporting technology transfer. The capability of The UWI Faculty of Engineering for UIC was assessed against the Model. Among the limitations identified, publications data shows that the relatively low international visibility and impact of research undertaken in the Faculty might harm confidence in potential industrial partners. The university policy framework does not create sufficient incentives for definite UIC initiatives to be established.

R.J. Rodriguez and W.G. Lewis, “Exploring Potential RF Hot Spot Locations in Confined Spaces of Large Wave Guide Dimensions”, focus on quantifying location characteristics and intensities of Radio-Frequency (RF) waves propagating within empty rectangular structures likened to those of communication waveguide structures, through the proposal of a theoretical model for predicting the locations and intensities of RF hot zones containing RF hot spots. The significance of this work lies in being a tool for RF safety practitioners who can, without the use of cumbersome equipment and necessary skills for measuring RF, use a less expensive and user friendly method for determining the level of RF safety in a confined space by simply knowing its internal dimensions and material of the surfaces; and the RF source characteristics. The work can provide the IEEE standard body with useful information to include in its guidelines.

R. Hosein and R. Mayrhoo, “A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad Gas Condensates”, discuss the need of accurate description of pseudo-component compositions for Equation of State predictions for gas condensate systems. These extended experimental data are often unavailable and must be generated using mathematical models, of which the exponential and the three-parameter gamma distribution functions are the two most widely used. The Model described in this study resolves the discontinuities in the molar relationships at both SCN8 and SCN12 with an Average Absolute Deviation between the predicted and experimental compositions of less than 10 percent. It is claimed that this model can quite easily be included in Equation of State packages for a more accurate description of compositions for Trinidad gas condensates for performing compositional simulation studies. A partial analysis beyond the C7+ fraction is not required with this new model.

E.I. Ekwue, R.C. Constantine and R. Birch, “Simulation of Irrigation Water Requirements of Some Crops in Trinidad Using the CROPWAT Irrigation Software”, present the design of irrigation schedules for the nine major crops grown in different predominant soils during the dry season for twelve (12) major farming locations in Trinidad. Crop and field parameters were obtained from published texts whereas the climatological data were obtained from the Water Resources Agency in Trinidad. The irrigation schedules using CROPWAT computer software package were planned in such a way that for the convenience of the farmer, the irrigation depth and irrigation interval were kept constant throughout the growing season for each crop and this value depended on the climatological situation or the water consumption pattern of the crops.

In their article, “Development and Evaluation of Wheeled Long-Handle Weeder”, S.O. Nkakini and A. Husseni present the design, construction and testing of a push-type operated wheel weeder with an adjustable long handle. The hoe performance from the tests on a field of Okra plant having an inter-row spacing of 800mm, showed that it could weed satisfactorily, and eliminate the drudgeries associated with the use of the short handle hoe. Field capacity and efficiency of 0.050ha/hr and 87.5% were obtained, respectively. Moreover, the average weeding index and performance index obtained were 86.5% and 1108.48, respectively. At a speed of 0.04m/s, a high efficiency of 91.7% at 0.4m depth of cut was obtained.

S.A. Oke, D.E. Ighravwe and G. Shyllon, “A Robust Regression Approach for Excess/Shortage Spare Parts Cost Estimation”, investigate the development of a predictive model for estimating the amounts of spare parts holding and the cost effects of poor spare parts holding in a system. The model uses an integrated methodology of the penalty cost and the wear technique for the unconstrained optimisation of the excessive spares using big-bang big-crunch (BB-BC) algorithm. It is validated by comparing the results obtained using the in-sample analysis with an out-sample approach. The model is infeasible to track spares and to evaluate the cost of excess/shortage of stockings of multi-items in spare parts inventories.

K. Ramesar, C. Maharaj and U. Persad, “A Mechanism for Cutting Coconut Husks”, describe a conceptual design of a machine for cutting coconut husk halves into pieces for activated carbon production. Alternative interlocking and welded blade arrangements are presented with the potential for scaling up the processing of coconut husks into smaller pieces. Virtual simulations and the experimental testing of a functional prototype are used to validate the conceptual design. It is claimed that the design is functionally acceptable, with directions for further improvements and development.

Page 5: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

WIJE; Vol. 37, No.2, January 2015 3

I.O. Oladele, “Mechanical Properties of Steel-making Slag Reinforced Polyester Composites”, assesses the viability of utilising steelmaking slag for reinforcing polyester matrix to form composites with improved mechanical properties. Slag was prepared by crushing and pulverizing, and then sieved into 75, 106 and 300 µm sizes and, varied masses of the particles were used to develop the composites by reinforcing the unsaturated polyester resin with the steelmaking slag particles. The homogeneous mixtures were poured into the flexural and tensile tests moulds and allowed to cure before being stripped from the moulds. The samples were further allowed to cure for 30 days before carrying out the mechanical tests. The results showed that the composites produced have gained increment in these properties.

In their article, “Rheological Study of Cement Modified with a Lignin Based Admixture”, R. Maharaj, L.H. Grierson, C. Maharaj and V. Ramjattan-Harry investigate with experiments the changes in the rheological properties of Trinidad Portland cement paste blended with Lignosulfonic acid, acetate sodium salt additive. The rheological properties of plastic viscosity (PV) and yield stress (YS) of the cement blend were calculated. It was found that a PV value of 0.7 centipoise obtained with the control sample can be reproduced with the addition of approximately 0.05% admixture using 20% less water. Maximum values of YS generally occur between 0.05% and 0.10% admixture concentrations as a more compact, homogeneous paste system develops. The ability to synthetically alter the rheological properties of Trinidad Portland cement adding a lignin based admixture can serve to optimise the strength, workability and shrinkage due to the reduced water-cement ratio.

E.I. Ekwue, R. Birch and J. Chewitt, “Effect of Dynamic and Static Methods of Compaction on Soil Strength”, investigate the effect of static (hydraulic press) and dynamic (Proctor) methods of compaction on the strength of soils in the laboratory. Soil samples of different densities were obtained by incorporating peat into three agricultural soils. Results indicate that as long as the same soils are compacted statically or dynamically at the similar moisture contents to same bulk densities, similar strength values are expected. The effect of method of soil compaction on soil strength is not important.

K.F. Pun, “A Strategic Initiative on Enhancing Postgraduate Throughputs at The UWI St. Augustine Campus”, reviews the key areas of priority for the Campus for the period 2014-2017, and informs the strategic initiative with a proposed Throughput Enhancement Project (TEP) at The UWI St Augustine Campus. It presents the structure of TEP and a schedule of its implementation. For facilitating the TEP initiative, project leaders and process owners are identified, and resource requirements versus savings are explored. The paper concludes by discussing the evaluation of the TEP efficacy in relation to achieving the Research and Innovation strategic goals of The UWI.

On behalf of the Editorial Office, we gratefully acknowledge all authors who have made this special issue possible with their research work. We greatly appreciate the voluntary contributions and unfailing support that our reviewers give to the Journal.

Our reviewer panel is composed of academia, scientists, and practising engineers and professionals from industry and other organisations as listed below: • Dr. Abrahams Mwasha, University of the West Indies

(UWI), Trinidad & Tobago (T&T) • Professor Chanan Syan, UWI, T&T • Dr. Chris Maharaj, UWI, T&T • Mr. Christopher P. Braden, Halcrow, a CH2M Hill

Company, Tampa, USA • Dr. Christos Braziotis, University of Nottingham, UK • Professor Emeritus David McGaw, UWI, T&T • Professor Edwin I. Ekwue, UWI, T&T • Dr. Gaius Eudoxie, UWI, T&T • Dr. Graham S. King, UWI, T&T • Dr. Isiaka Oluwole Oladele, Federal University of

Technology, Akure, Nigeria • Dr. Jeffery A. Jones, University of Warwick, Coventry,

UK • Professor John E.L. Simmons, Heriot-Watt

University, Edinburgh, UK • Dr. Kiran Tota-Maharaj University of Greenwich, UK • Professor Kit Fai Pun, UWI, T&T • Ms. Man Yin R. Yiu, UWI, T&T • Dr. Musti Sastry, UWI, T&T • Professor Peter Hogarth, Bournemouth University,

Poole, UK • Professor R. Lal Kushwaha, University of

Saskatchewan, Canada • Professor Ramsey Saunders, UWI, T&T • Dr. Reynold Stone, UWI, T&T • Dr. Rupert Williams, UWI, T&T • Dr. Sanjay Bahadoorsingh • Professor Stephan Gift, UWI, T&T • Dr. Umesh Persad, University of Trinidad and Tobago • Dr. William Wilson, UWI, T&T • Professor Winston G. Lewis, UWI, T&T • Professor Yew-Chaye Loo, Griffith University,

Australia

The views expressed in articles are those of the authors to whom they are credited. This does not necessarily reflect the opinions or policy of the Journal.

KIT FAI PUN, Editor-in-Chief Faculty of Engineering,

The University of the West Indies, St Augustine, Trinidad and Tobago

West Indies January 2015

Page 6: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

C.R. Cameron and G.S. King: An Assessment of The UWI’s FoE Capability and Willingness to Engage in Industrial Collaboration for Innovation 4

An Assessment of The UWI’s Faculty of Engineering Capability and Willingness to Engage in Industrial Collaboration for Innovation

Cary R. Cameron a,Ψ, and Graham S. King b

Department of Mechanical and Manufacturing Engineering, Faculty of Engineering, The University of the West Indies, St

Augustine, Trinidad and Tobago, West Indies aE-mail: [email protected] bE-mail: [email protected]

Ψ - Corresponding Author (Received 12 April 2013; Revised 28 July 2014; Accepted 30 January 2015)

Abstract: University-industry collaboration (UIC) has been recognised by numerous authors as engendering innovation and economic development. Despite Trinidad and Tobago’s highly industrialised economy, the efforts of The University of the West Indies’ (The UWI) Faculty of Engineering to foster such UIC has been limited. We use a Conceptual Model of good practice in UIC, which requires that four pillars are in place for effective UIC: high quality academic research; a predisposition on the part of academics to engage; a framework of supporting policies and procedures in the university; and an effective office supporting technology transfer. Our assessment of the capability of The UWI Faculty of Engineering for UIC is against the Conceptual Model. Publications data shows that the relatively low international visibility and impact of research undertaken in the Faculty might harm confidence in potential industrial partners. Although academic members of staff are keen to collaborate and many have industrial experience, the university policy framework does not create sufficient incentives for definite UIC initiatives to be established. The UWI Office of Research, Development and Knowledge Transfer does not have sufficient human or financial resources to fully support UIC. Strategic action by The UWI can alleviate all these limitations and significantly improve capability for UIC. Other universities in the developing world, particularly in Africa, the Caribbean and the Pacific, may be facing similar challenges in UIC and could learn directly from our work.

Keywords: University-Industry Collaboration (UIC); Technology Transfer; Caribbean Universities; The UWI.

1. Introduction Economic development is, in most nations, very dependent on innovations in product or process. The economy of the Caribbean region is heavily dependent on tourism, which has suffered severe reductions in the past years due to the economic slowdown and recession in the USA and Europe between 2007 and 2012. Some Caribbean nations suffer under the crush of International Monetary Fund (IMF) Structural Adjustment Programmes, while others suffer from the depletion or devaluation of commodities on which they have relied for generations. There has been little innovation to establish economic streams.

The primary economic engine of the English-speaking Caribbean is Trinidad and Tobago (T&T), with rich reserves of oil and natural gas, and a more developed manufacturing sector than other nations in the region. However, in the 2012-2013 Global Competitiveness Report, T&T was ranked 104th out of 144 countries for its Capacity for Innovation (Global Economic Forum 2012, 20). The World Trade Organisation (WTO) stated in its 2012 report on T&T:

“Although successive Governments have recognised the need to diversify the economy, and several initiatives have been taken, the low level of

reserves relative to current production of oil and gas, and the increase in gas production in some countries means significant diversification is needed in the medium to short-term.” (WTO, 2011, ix)

The Government of Trinidad and Tobago recognises the need to encourage innovation as a core strategy for the diversification of the economy away from an excessive dependence on oil and gas, and stated in its 2012 Budget, one of the three major policy platforms is: "...creation of entrepreneurial opportunities and an innovation-driven economy to stimulate growth and competitiveness through public/private investment“ (MOF, 2011, 2).

Clarifying and strengthening the national innovation system is a high priority of the Government of Trinidad and Tobago, and significant attention is being paid to the policy framework that will facilitate innovation. According to Etzkowitz and Leydesdorff (2000), an effective system of innovation consists of a series of knowledge-intensive networks between players in the innovation system. Industry, the state and academia are the primary players, with university research providing the primary locus in a triple-helix model.

The vital function of knowledge creation in the

ISSN 0511-5728 The West Indian Journal of Engineering

Vol.37, No.2, January 2015, pp.4-13

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 7: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

C.R. Cameron and G.S. King: An Assessment of The UWI’s FoE Capability and Willingness to Engage in Industrial Collaboration for Innovation

5

triple-helix model is largely fulfilled by the university system. In this model, the so-called Mode 2 University functions as an “amalgam of teaching and research, applied and basic research, entrepreneurial and scholastic” (Etzkowitz et al., 2000). This stands in contrast to the Mode 1 University where the primary functions of the university are education and discovery-focused research. The presence of such a university that is meaningfully engaged in Research and Development (R&D) has a positive impact on high-technology innovation in its environs. In fact, there is a correlation between University R&D expenditure and technology transfer (Varga, 1998). Where there is an existing concentration of economic activities in a locality, the impact of University R&D on local economic performance is even greater. Engineering faculties are of particular importance due to their central role in the technology transfer process.

As the largest and most diverse university in the region, The University of the West Indies (The UWI) has a pivotal role to play in energising the regional innovation system. By application of the Etzkowitz and Leydesdorff (2000) model and the findings by Varga (1998), R&D emerging from The UWI is critical to regional economic development. The UWI Mission Statement recognises and reflects its role in the development of the Caribbean region (The UWI, 2012, 20). Beyond a statement of intent, such as that contained in The UWI Strategic Plan 2012-2017, it is necessary to determine whether, at a faculty level, the capabilities and commitment required for the University to make its necessary contribution, are present. This paper focuses on the Faculty of Engineering. 2. Methodology The aim of this paper is to make a reasonable assessment of the willingness and capability of the Faculty of Engineering at The UWI, located in Trinidad and Tobago, for University-Industry Collaboration (UIC) and to identify ways in which these could be improved. A Conceptual Model of good practice in the facilitation of successful UIC, based on practice reported in the literature and augmented by lessons learnt in a study of a sample of leading UK universities, is used as the basis of assessment of the capability for UIC in the Faculty of Engineering at The UWI. The following factors are assessed:

a) The attitudes of academics in the Faculty of Engineering to UIC;

b) Engineering research quality as indicated by an analysis of publications by engineering faculty members; and

c) The role of the Office of Research, Development and Knowledge Transfer (ORDKT).

2.1 Assessment of Research Quality and Publication An analysis of the publication habits of the Faculty of

Engineering was conducted to evaluate whether research issuing from the Faculty has impact on the wider academic community. A high impact would suggest that “credible and high-quality R&D activities must be taking place in the University”. The vehicle used for this analysis was citation analysis, which is defined by Smith (1981) as “the evaluation and interpretation of the citations received by articles, scientists, universities, countries, and other aggregates of scientific activity, used as a measure of scientific influence and productivity”. Thus, citation analysis is based on the referencing practices of members of the academic community and which are seen as indicative of the regard for the cited work and the impact it has on the scientific community (Van Raan, 2005; Narin, 1976). One of the powerful tools used in citation analysis is the citation index or database, which is a compilation of citation data from sources in a particular area. Though the use of such databases is subject to technical and methodological problems, these databases still greatly facilitate the use of citation analysis (Van Raan, 2005; Smith, 1981; Klinger, 2006; Garfield, 1996; Kostoff, 2002).

Consideration of Kostoff (2002) and Bornmann et al. (2008) makes it clear that attempting to use simple citation counts is not advisable as sub-disciplines have different publication habits as a result of norms within that particular sub-discipline. Therefore, citation frequencies must be situated in a frame of reference. In this case study the frame of reference used was “field of research” aggregated at the level of each department within the Faculty. The method of aggregation used was inspired by the work of Narin (1976) and Schubert and Braun (1996). Thus the reference standard used for each department was derived via weighting based on the publication activity of its researchers in their areas of research as explained below.

Faculty of Engineering (FoE) publication records were obtained from The UWI Annual Report. These records were systematically reviewed and each article which was published in a journal for the 2006/07 to 2010/11 academic years inclusive was cross-referenced with the Thomson Reuters Science Citation Index (SCI). Articles were evaluated on a departmental basis to obtain the following statistics: • Productivity (Number of articles) of each

department each year; • Percentage of journal articles for each department

present in the SCI; • Average Citation Rate per Year for Each

Department; • Comparative Citation Rate for a hypothetical similar

department; • Ranking of Researchers based on number of journal

articles published each year and over the entire examined time period;

• The h-index of each department for the five-year

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 8: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

C.R. Cameron and G.S. King: An Assessment of The UWI’s FoE Capability and Willingness to Engage in Industrial Collaboration for Innovation

6

period examined. The average citation rate for each Department per

year was computed by weighting on the basis of the number of articles published in each subfield with the total number of articles for that year for that department in the SCI. The comparative citation rate is found by Equation 1:

[1]

where is the average citation rate for the

Department, is the citation rate for each subfield

and is the total number of articles published. The comparative citation rate for a hypothetically similar department was computed by obtaining the average citation rate for each subfield in which articles were published by the department in question. The hypothetical department citation rate is found by Equation 2:

[2]

where is the average citation rate for a

hypothetically similar department, is the SCI

average citation rate for each subfield and is the total number of articles published.

Comparing departmental metrics with SCI citation frequency rates gives a proxy for research quality and indicates the perception of the research quality of each Department in the wider academic community (Rudd, 1988). 2.2 Assessment of Academic’s Attitudes A questionnaire was administered to all full-time members of academic staff in The UWI Faculty of Engineering, to gather data on attitudes to different issues that affect UICs. Standardised questions ensured that all respondents were answering the same questions and a Likert scale ensured that respondents answered in a uniform manner (Mills et al., 2010).

The questionnaire was administered electronically. Using a web-based interface, responses remained anonymous, which was beneficial given the sensitive nature of some of the questions. The questionnaire consisted of five demographic questions fields and twenty-five questions related to different aspects of UICs: twenty Likert scale questions; two “Yes/No” questions; and two close-ended multiple-choice questions. Affirmative statements were generally formulated to allow nuanced responses on a five-point Likert scale. Nine general themes were addressed:

a) The role of universities in innovation and technology transfer;

b) Personal orientation with regard to research collaborations;

c) Personal Orientation to Research

Commercialisation and Intellectual Property; d) Willingness to Accommodate Industrial Partners; e) Industrial Experience; f) Government Involvement; g) Incentives; h) Barriers to University-Industry Collaboration; i) Organisational Culture.

After a pilot study, changes were made to the wording of some of the questions and one question added exploring faculty members’ previous experiences with industrial research. 2.3 Assessment of Technology Transfer Support A semi-structured interview was conducted with a Business Development Manager from The UWI Office of Research Development and Knowledge Transfer to gain a more detailed understanding of the operations and challenges of the office. Information was sought about: the organisation of the ORDKT; the processes used by the ORDKT; and the personnel composition of the office. This was combined with personal knowledge derived from close interactions with the ORDKT. 3. Conceptual Model High-quality university research coupled with strong industrial partnerships is a recipe for successful innovation and strong technology transfer (Dooley and Kirk, 2007). Most partnerships between universities and industry that lead to innovation and technology transfers tend to involve either science or engineering university faculties.

For a university to have a significant positive impact on innovation through technology transfer, and hence contribute to economic development in its locale, a number of important factors must be in place. The model that is proposed here is an evolution of work previously reported by King and Cameron (2013). The model presented here has been clarified and is now augmented with good practice data collected on the functioning of UICs in UK universities. A visual representation of the conceptual model is shown in Figure 1.

Figure 1. Graphical Representation of UIC Good Practice Conceptual Model

Qua

lity

Rese

arch

&

Publ

icatio

ns

Academics Administration

Pred

ispos

ed

to

Enga

gem

ent

Supp

ortin

g Po

licie

s &

Proc

edur

es

Effe

ctiv

e an

d Su

ppor

ting

TT

O

Effective University Industry Collaboration

Building relationship leading to

further collaborative

projects

Resources (e.g. equipment,

industry expertise) flow to the University

Revenue for the University

Technology Delivered

Regional/National Economic

Development

Practice and Publications

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 9: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

C.R. Cameron and G.S. King: An Assessment of The UWI’s FoE Capability and Willingness to Engage in Industrial Collaboration for Innovation

7

Effective UIC has numerous benefits to company, national and regional development, to the university, to the participating academics and to students. These have been widely reported elsewhere (King and Cameron, 2013) and some of the most important are highlighted as outputs. Effective UIC requires full support of both academics and the university administration. This is illustrated by these two being the ‘foundation stones’ in Figure 1. 3.1 Quality Research and Publications Credible and high-quality R&D activities must be taking place in the university, so that:

a) knowledge is generated that can lead to innovation and technology transfer; and

b) industrial partners will gain confidence in the capability and value of the university;

c) students and researchers will be attracted to the university, who can then be engaged in industrial research activities.

3.2 Predisposition to Engagement Members of academic staff must value the opportunity to engage with industrial partners in research projects and actively pursue such opportunities (Barbolla and Corredera, 2009). Research in UK universities identified that some academics tend to be much more predisposed to engagement than others, and these are the ones that can make a significant difference in the institution – so-called ‘stars’. For academics to develop a predisposition to engagement requires that:

a) there is a connection between engaging in UICs and the formal assessment and promotion structure of the University (Siegel et al., 2003; Link et. al, 2007);

b) recognition be given for the real-world impact of an academic’s work, beyond recognition of publications;

c) experience in interfacing with, or working in, an industrial context, which sensitises academics to industrial considerations – new formats of doctoral programme and incorporating industrial placements into traditional doctoral programmes are good practices in this regard as seen in UK universities;

d) relationships between academics and industry leaders are established and maintained – the relationships between academics and industrial partners are key to positive outcomes and sustainability;

e) academics are enthused about research work that lies adjacent to, not only directly within, their established sphere of expertise.

3.3 Supporting Policies and Procedures Policies and procedures in a university shape the expectations and behaviour of its academics. If a

university wishes to encourage UIC, then its policies and procedures should be aligned with that objective. This requires:

a) allocation of adequate resources in the University both toward research and support services to facilitate University-Industry Collaborations;

b) a direct systemic link between engaging in UICs and qualification for increased resources for research, whether through industrial partners or public funding (WIPO, 2005; Acworth, 2008. Mustar, 2008) – this can be achieved by requiring that academics demonstrate the impact of their research on real-world scenarios;

c) fair and reasonable sharing of revenues from exercised patents between the inventor academic and the institution;

d) assessment and promotion policies that reward: i. application of research and technology

transfer; ii. generation of research income for the

university from industrial partners; e) recruitment of academics with a predisposition

towards, and track record in, UIC. 3.4 Supporting Technology Transfer Office In UK universities, it is seen that a functional and effective TTO makes a very important contribution to UIC. It serves numerous facilitating purposes and good practice dictates that it should:

a) build social connectivity with potential industrial and other partners – both informally and formally;

b) facilitate academics in matching their research activities to funding partners, whether grant agencies or industrial partners;

c) provide the university with institutional knowledge – breadth of sight of the needs and opportunities in its environment and to which it could respond;

d) enhance the alacrity of the university in responding to emergent needs;

e) ensure that the university has the capability required to deliver desired project results through effective project management;

f) assist in the evolution of university policies and practices to match its ever-changing environment;

g) facilitate the protection of innovations and intellectual property that arise from university endeavours;

h) have programmes that assist in the commercialisation of inventions that are generated by the university, especially through the creation of spin-out companies;

i) integrate the university into industrial and technological agglomerations and partnerships, which must effectively provide a ‘market’ for the research generated and connect the university directly with the economic life of its locale;

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 10: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

C.R. Cameron and G.S. King: An Assessment of The UWI’s FoE Capability and Willingness to Engage in Industrial Collaboration for Innovation

8

If each of the four pillars described in this model is in place, then a university is well able to engage in effective UIC and contribute positively to the national or regional innovation ecosystem. We can use this model as an assessment tool for any university institution. 4. Results and Analysis 4.1 Publication and Citation Data Firstly, let us consider the rate of journal publications as an indication of research productivity. Productivity can be used as a measure of research quality if it is assumed that the funding used to finance research is awarded based on judgements of the research quality and merit of previous research work (Rudd, 1988). Note that this metric does not capture contract research that is fully funded by industrial partners and which may not be readily published, but it is known to the authors that little such research takes place at The UWI.

Therefore, the rate of publications serves as a reasonable proxy for overall research activity. The percentage of Faculty members publishing journal articles in the period 2006-2011 is shown in Table 1. The general trend observed from this data is that the majority of engineering academics are not publishing their research by means of journal articles. Table 1. Percentage of Academic Members of Staff who Published

Journal Articles Year Percentage Publishing 2006-2007 28% 2007-2008 16% 2008-2009 25% 2009-2010 19% 2010-2011 23%

The 2007 edition of the Faculty Scholarly Productivity Index shows that for the ten universities in the US which produced the highest number of academic publications via journal publication, at least 83% of engineering faculty had published a journal article in the last academic year (The Chronicle of Higher Education, 2007). It is acknowledged that these institutions operate in a different context than The UWI. However, it does illustrate that academic faculty at top-tier universities use the medium of journal articles to disseminate their research, and that The UWI Faculty of Engineering falls far short of the levels of involvement of faculty in publishing that is found at these top tier universities. This represents a significant opportunity for improvement for the Faculty of Engineering.

Secondly, the visibility of the articles that are published can be taken into consideration. Figure 2 compares the number of The Faculty’s journal articles that are contained in the SCI with the total published over five (5) academic years. Overall, the ratio of papers contained in the SCI to those published is 40%.

Correlation between the number of articles published by a department and the number of the articles which are included in the SCI is weak, giving a linear regression coefficient of determination (R2) of 0.37. This was not surprising as a number of factors influence the inclusion of an article in a citation, not least the journal in which the article is published (Kostoff, 2002; Cronin, 2001). It could be argued, however, that a high SCI score indicates the usefulness of the research to the wider academic community and implies that the research is of good quality. Consistently high SCI scores would mean good international research visibility, regular citations, and a commensurate research reputation.

Figure 2. Faculty Total Papers Published vs. SCI References

It is noted that many of the Faculty’s publications

are in journals not covered by the SCI; whether in Asian journals, regional journals or even non-English language journals. Such publication practices can reduce the coverage of the Faculty’s publications in the SCI and cause a reduction in the visibility of the research output of the Faculty (Abramo et al. 2010; Bornmann et al., 2008).

Thirdly, and continuing the theme of considering the impact and usefulness of The Faculty’s research, Table 2 shows average citation frequencies for each department (A) in comparison with the average citation frequencies for each subject classification to create a weighted average citation frequency for a theoretical department with similar publication norms (C). These publication norms would be similar with specific regard to the productivity of that theoretical department in the same subject area classifications. The table below shows the citation frequencies for each department in the Faculty and for their respective comparison theoretical departments.

In almost all cases, in the five academic years examined, the average citation rates are far below the respective citation rates of the theoretical departments.

0

10

20

30

40

50

60

2006-2007 2007-2008 2008-2009 2009-2010 2010-2011

Faculty Total

Papers Published Papers Contained in SCI

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 11: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

C.T. Benjamin and K.F. Pun: An Exploratory Study to determine Archetypes in the Trinidad and Tobago Fashion Industry Environment 9

Table 2. Average Citation Frequencies by Department

Department Variable Academic Year 2006-2007 2007-2008 2008-2009 2009-2010 2010-2011

Chemical Engineering A 4.75 2.29 2.33 1.33 1.08 C 7.06 8.78 5.93 4.16 1.90

Civil Engineering A 3.00 - 0.50 0.00 0.33 C 9.39 - 2.22 2.95 1.02

Electrical Engineering A 2.00 3.38 2.50 2.14 0.00 C 8.35 6.81 6.98 4.87 1.37

Geomatics Engineering and Land Management A 3.50 - 2.00 0.00 0.00 C 9.63 - 5.71 4.16 1.45

Mechanical Engineering A 6.00 - 1.00 2.00 0.00 C 6.78 - 4.94 4.43 3.12

This may indicate either or both of: (i) poor regard amongst the wider academic community for the research coming out of the Faculty; and (ii) poor visibility of the research reported by the Faculty. 4.2 Questionnaire Responses Response rates for the questionnaire were 20% of the target group. While this is a bit too low for the establishment of statistical significance, the results can be used qualitatively and indicatively. In the description and analysis presented here, key themes are identified and inferences made as to the dominant attitude among academic staff on that particular issue.

The first issue that will be considered is willingness of Faculty of Engineering academics to engage in UICs. The responses to a number of the questionnaire items give positive indications with regard to this issue. Respondents unanimously indicated that they thought that universities such as The UWI should be involved in supporting regional innovation and economic development. They also were generally supportive of knowledge transfer between The UWI and industry and perceived little ethical conflict with those types of activities. Furthermore, most respondents recognised the potential benefit that collaborative research with industry could have for their own research, and most were willing to modify their research and publication habits to accommodate the needs of industrial partners. These responses are important because ethical conflicts, differences in research priorities and lack of accommodation of industrial partners can hinder the success of UICs (Dooley and Kirk, 2007; Barnes et al., 2002). Sensitivity to these issues can be regarded as a positive indicator for the potential success of any future UICs.

Further supporting the impression that Faculty of Engineering academics are willing to engage in UICs. Most respondents had previous industrial experience or would be interested in industrial attachments. In addition, the majority of respondents maintain networks of industrial contacts. This willingness to engage with industrial partners is a positive factor that can contribute to the success of UICs as academics with industrial

experience can be effective “boundary spanners” who by virtue of their experience and contacts in both the industrial and academic worlds can facilitate the creation and operation of UICs (Thune, 2011; Pertuzé et al., 2010; Philbin, 2009; Wright et al., 2008). Industrial experience furnishes the academic with knowledge of the possible applications of research carried out in the university environment, and industrial contacts can also help to keep him up to date with the needs of industry.

The reminder of the questionnaire considered other issues that can affect the process of knowledge transfer and therefore affect the success of UICs. The first of these was the attitude of the surveyed academics to commercialisation of their research. While some of the respondents indicated interest in commercialisation of their research, there was a large proportion that expressed indifference towards commercialisation. This suggests that academics in the Faculty of Engineering may not be particularly entrepreneurially oriented. This does not bode well for the potential of commercialising the faculty’s research as such attitudes suggest a low level of motivation among the faculty. In addition, a perceived lack of interest in entrepreneurial activities within a faculty has been seen to dampen the level of interest of other academics in pursuing such activities (Tartari et al., 2010; O’Shea et al., 2007).

The next issue considered was attitudes to intellectual property (IP). The majority of respondents indicated that they would have no issue with the university holding the rights to intellectual property they create using university resources. What was also clear was that researchers expect to be remunerated fairly for any revenue-generating IP that they produce. This is in line with recognised practice in other countries where universities hold the rights to the research of their employees. This has been found to be a more efficient and successful way of managing revenue-generating technology transfer (Verspagen, 2006). These findings are supported by previous research that shows that fair compensation should be built into universities’ IP policies as this can improve the commitment of academic researchers to UICs (Mustar et al., 2008).

The final issue considered from the questionnaire

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 12: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

C.T. Benjamin and K.F. Pun: An Exploratory Study to determine Archetypes in the Trinidad and Tobago Fashion Industry Environment 10

data is barriers to involvement in UICs. The major barrier highlighted by the responses is a lack of suitable government funding programmes for UICs in specific areas. In some countries where UICs are more common, government financial support of innovation has been provided often through multiple agencies and programmes set up to fund particular scientific areas. Examples of this are the Biological Sciences Research Council (BBSRC) and the Engineering and Physical Sciences Research Council (EPSRC) in the UK (Acworth, 2008). In many cases, research proposals that include UICs are considered more favourably. The provision of local or regional government funding for engineering projects that included UICs could catalyse engagement with industry by academics in the faculty.

Another potential barrier highlighted was the perception that the majority of respondents had that The UWI’s organisational culture does not support innovation. O’Shea et al. (2007) highlighted in examining the contributing factors to the success of the Massachusetts Institute of Technology as a hotbed of innovation and entrepreneurialism highlighted that the development of an organisational culture that supports innovation is key. These responses suggest that greater institutional support is required for innovative research by the Faculty, and that the support services available must be marketed to academics. 4.3 Supporting Policies and Procedures Although a detailed examination of policies and procedures is not the primary concern of this paper, it is worth mentioning that some are in place at The UWI. Specifically:

a) An IP Policy, in line with the recommendations of the World Intellectual Property Organization (WIPO), is in place;

b) The IP Policy includes specific provision for the splitting of revenues between the academic or inventor and the university in the case that a patent generates income;

c) An invention disclosure process that allows for the assessment of an invention for patenting support by the University; and

d) Recognition of patents as being worth three journal papers for the purpose of academics’ assessment and promotion.

There is no apparent deliberate attempt made on the part of The UWI to recruit academics with experience in or disposition towards UIC. Apart from regulations for academics engaging in consultancy with private enterprises, there are no policies or procedures that specifically address wider issues of UIC. In addition, there is no institutional effort towards building functional relationships with industrial partners. Beyond IP and patenting, the policy framework addressing commercialisation of research findings is very weak.

4.4 Technology Transfer Support The ORDKT provides the typical TTO activities of patenting, licensing and liaising with industry (Lee et al., 2010). Apart from patenting and licensing activities, the ORDKT assists in UIC through administering consultancies and sponsored research. Consultancy is the most common form of UIC that the university engages in. Patents are occasionally pursued, and The UWI currently has a handful of US patents pending.

Barnes et al. (2002) highlighted that engaging in a variety of modes of UICs helps to strengthen relationships between universities and industrial partners. In order to foster UICs, the ORDKT has assisted the Faculty of Engineering in the staging of industry consultations. The aim of such consultations is to ascertain the research needs of industry, but it also served to help build social networks. Being apprised of the needs of industry has been identified as a contributor to successful UICs and activities such as industrial consultations have been recommended in previous research on the subject (Dooley and Kirk, 2007; Siegel et al., 2003). While the activity in these areas is encouraging and shows that there is interest in strengthening UIC, the activity is still immature.

Funding is another issue of vital importance for the success of UICs. The ORDKT is actively involved in seeking financial support for university research projects. Funding comes from a mixture of government and multi-lateral funding agencies, although government funding has been decreasing in recent years. Funds are disbursed through various grants for which researchers can apply. Unfortunately, these budgetary restraints may be a limiting factor for academics seeking to engage in research. The low level of funding for operations of the ORDKT itself, particularly compared with the annual patenting and licensing budgets of TTOs in developed countries, is also a challenge.

Staffing represents a challenge for the ORDKT. While there are a number of professionals with expertise in legal matters and business development there are still gaps in the available skill sets in the ORDKT, which need to be met. The interviewee suggested that one such gap exists in the area of technology marketing. There is also a need for more technical support staff to aid in research and evaluating patent applications. Staff headcount is low compared to the volume of work that could potentially be undertaken, and this limits the extent to which proactive relationship building can take place with external partners and potential partners. 5. Discussion The conceptual model proposed earlier in this paper posits that if certain factors are present then the institution or faculty will have the capability to engage in meaningful UIC. Therefore, to determine the capability of The UWI’s Faculty of Engineering for UIC it is necessary to compare the present situation of The

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 13: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

C.T. Benjamin and K.F. Pun: An Exploratory Study to determine Archetypes in the Trinidad and Tobago Fashion Industry Environment 11

Faculty with the conceptual model’s four pillars of ‘good practice’. Pillar 1: Quality Research and Publications Comparing the SCI citation rates for each of the department in The Faculty to a theoretical academic department that publishes in the same areas, it is seen that, generally, the rate of citations of the Faculty is significantly lower than those of theoretical departments that publish in the same areas.

Research published by members of The Faculty is not being cited as often as work from researchers in equivalent departments in other institutions that publish in the same areas. This low citation frequency could be indicative of:

a) the scientific community’s lack of regard for the researchers and research of The Faculty of Engineering; and

b) a lack of visibility of research from The Faculty; c) the publication habits of the researchers in the

faculty, whereby regional or national journals are targeted rather than international journals;

d) a generally low level of productivity among researchers within The Faculty, with relatively few academics publishing regularly.

Points (c) and (d) above combine to diminish the profile of the university to other members of the scientific and academic communities. On the basis of these findings it can be said that the Faculty of Engineering does not meet the requirement of ‘quality research and publications’ based on the collected data.

In order to boost the confidence of industrial partners in The UWI for UIC, members of faculty must continue to be encouraged to publish relevant papers in top-tier academic journals. The university should also profile its research nationally and regionally to build respect in the eyes of potential partners. Pillar 2: Predisposition to Engagement Based on the responses to the questionnaire, it was clear that the majority of academic staff members who responded recognised the value of engaging with industry in their research, and were willing to do so. The respondents also indicated that for the most part they perceived no difficulty with pertinent issues such as ethical conflicts, difference in research priorities and lack of accommodation of industrial partners. The academics' flexibility with these issues is a positive indicator for the success of UICs. Furthermore, the majority of respondents indicated that they had previous industrial experience and maintained networks of industrial contacts. Both of these points add to the positive predisposition to engagement of the academics of the faculty. The data also revealed that many of the respondents were indifferent to commercialisation and entrepreneurial activity, which would tend to detract from the predisposition to engagement.

Based on our research data, it seems that this pillar for UIC is in place and could be strengthened if The Faculty or the ORDKT were to assist academic members of staff in efforts to build relationships with industrial partners. Pillar 3: Supporting Policies and Procedures Supporting policies and procedures have been established at The UWI that could potentially assist in the technology transfer process. However, the shortfall lies in the lack of utilisation of those policies. For instance, the invention disclosure process is generally utilised less than five times per year. Relatively few patents are filed by the university, so despite the revenue sharing algorithms that are defined and the benefit to an academic’s assessment and promotion prospects, research that produces patents seems to be scant. Furthermore, there is limited budget allocated to defending patents, and without the means to defend a patent, its value becomes negligible. In its history, The Faculty of Engineering has only produced one spin-off company, which reflects the lack of policies and procedures that incentivise such innovation activities.

It would be in the University’s interest to more actively pursue UIC to support the academics and capitalise on the benefits that it can afford. This work indicates that there is a predisposition towards UIC among academics, so the reason for the low numbers of collaborative may be due to university culture. Policies and procedures help to form the culture, so amending the policy framework to positively reward involvement in UICs increases capacity. Such a move would be in line with university policies in other parts of the world (Polt et al., 2001; Perkmann and Walsh, 2007). Pillar 4: Technology Transfer Office Industry consultations held by the Faculty of Engineering and supported by the ORDKT address consultations in part points 'a' and 'b' in the fourth pillar of the Conceptual Model and make small steps toward fulfilling point 'f'. The consultations assist in relationship building, assessing the needs of local industry, and enlighten potential industrial partners about the capabilities of The Faculty.

The activity of the ORDKT in facilitating the protection of IP by means of filing for patents where necessary addresses point 'h' in fourth pillar. On the other hand, the interviewee from the ORDKT stated that one of the deficiencies of the ORDKT is staffing including areas such as business development. This suggests that currently the ORDKT cannot fulfill point 'g' of the conceptual model – protection of the IP of innovations.

At present, The UWI does not fulfil the fourth pillar of the Conceptual Model, since the ORDKT has insufficient financial and human resources to provide strong support for UIC, or even, for that matter,

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 14: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

C.T. Benjamin and K.F. Pun: An Exploratory Study to determine Archetypes in the Trinidad and Tobago Fashion Industry Environment 12

extensive technology transfer support. Increasing the funding of the ORDKT could be seen as an investment if the return is obtained in the form of incoming research funding, revenues from patent royalties, dividends from spin-out companies, and sponsorship from industrial partners. Implications for Other Universities The challenges faced by The UWI in engaging in UIC, as a regional university in the developing world, are not unique. It is quite likely that many other institutions, especially those in Africa, the Caribbean and the Pacific (ACP), also have difficulty in competing globally in research quality, might have an orientation towards teaching rather than research, suffer from an underdeveloped policy framework and lack an effective TTO. The Conceptual Model and findings from this study at The UWI might be very useful as a means of comparison for universities across the ACP region. This could be pursued as further work. 6. Conclusion We have proffered a Conceptual Model for UIC which has as pillars for its success: quality research and publications; predisposition to UIC on the part of academics; appropriate policies and procedures in the institution; and an effective supporting technology transfer operation.

The publication habits of academics in The UWI Faculty of Engineering have not served to elevate the research reputation of the institution, with the result that industrial partners are less likely to have confidence in the value of The Faculty as a partner.

Questionnaire results showed that the majority of the academics who responded are willing to be involved in UICs and to engage with industry in a variety of ways. There is also a willingness to accommodate the needs of industrial partners in collaborations. These findings bode well for future attempts to initiate UICs. However, a low level of social interaction between academics and industry, and an absence of commercial or entrepreneurial orientation in the research agendas of academics are factors that need to be addressed to increase the probability of success of UICs involving The UWI.

Some policies and procedures to support UIC are in place, but incentives are required to encourage academics to engage more actively in UIC. Hence, the ORDKT is significantly under-resourced which limits the range of TTO functions that it can effectively perform. The UWI would be well advised to remedy this situation so that the office can provide much stronger support for UIC. Financial investment in the ORDKT should generate a clear and positive return on investment in terms of research revenue generated.

Recommendations to improve the performance of the Faculty of Engineering with regard to UICs are:

a) Tie the granting of research funds to academics more closely to an assessment of the track record of their research impact.

b) Ongoing dialogue and consultations with industrial leaders to learn the significant issues that their companies face that could become the focus of research studies.

c) Recruiting and retaining faculty in sub-disciplines that are identified as industrially relevant.

d) Government, The UWI and Industry to create and fund research grants in subfields that have particular industrial and national significance for economic development.

e) Establishment of a policy for collaboration with industry. This policy would provide financial incentives for academic inventors for their involvement in UICs. Promotion and tenure policies should also be modified to reward involvement in UICs.

f) Additional financial and human resources should be allocated for the ORDKT taking into consideration best practice and the resources needed to support the University’s goals for the ORDKT.

This evaluation has made it clear that achieving strong and productive links and viable UICs between the Faculty of Engineering and local industry requires restructuring and a significant investment of energy by both the Faculty and the University Administration. This situation is probably not unique to The UWI, but may be reflected in many other developing world universities, especially those in the ACP regions. References Abramo, G., D’Angelo, C. A., and Di Costa, F. (2010),

“University-industry research collaboration: a model to assess university capability.” Higher Education, Vol.62. No.2, pp. 163–181.

Acworth, Edward B. (2008), “University–industry engagement: The formation of the Knowledge Integration Community (KIC) model at the Cambridge-MIT Institute”, Research Policy, Vol.37, No.8, pp. 1241-1254.

Barbolla, A., Bernardos, M. and Casar Corredera, J.R. (2009), “Critical factors for success in university–industry research projects”, Technology Analysis and Strategic Management, Vol.21, No.5, pp.599-616

Barnes, T., Pashby, I. and Gibbons, A. (2002), “Effective University - Industry Interaction: A Multi-case Evaluation of Collaborative R&D Projects”, Science, Vol. 20, No. 3, pp. 272-285.

Bornmann, L., Mutz, R., Neuhaus, C., & Daniel, H. (2008), “Citation counts for research evaluation: standards of good practice for analyzing bibliometric data and presenting and interpreting results”, Ethics in Science and Environmental Politics, Vol.8, pp. 93–102.

Cronin, B. (2001), “Bibliometrics and beyond: some thoughts on web-based citation analysis”, Journal of Information Science, Vol.27, No.1, 1–7.

Dooley, L., and Kirk, D. (2007), “University-industry collaboration: Grafting the entrepreneurial paradigm onto academic structures”, European Journal of Innovation

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 15: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

C.T. Benjamin and K.F. Pun: An Exploratory Study to determine Archetypes in the Trinidad and Tobago Fashion Industry Environment 13

Management. Vol. 10, pp. 316-332. Etzkowitz, H. and Leydesdorff, L. (2000), “The dynamics of

innovation: from National Systems and ‘Mode 2’ to a Triple-Helix of university-industry-government relations”, Research Policy, Vol. 29 (2), No. 2, pp. 109-123.

Garfield, E. (1996), “The significant scientific literature appears in a small core of journals”, The Scientist, Vol.10, No.17, pp.13-16.

Global Economic Forum (2012), The Global Competitiveness Report 2012-2013: Full Data Edition, available at: http://www3.weforum.org/docs/WEF_GlobalCompetitivenessReport_2012-13.pdf. (Dated 31 March 2013)

King, G.S. and Cameron, C.R. (2013), “An enhanced model for university-industry collaboration for innovation in Trinidad and Tobago”, West Indies Journal of Engineering, Vol.36, No.1, pp.86-94.

Klinger, B., & Lederman, D. (2006), Diversification, innovation, and imitation inside the Global Technological Frontier, World Bank. Retrieved on 8th June 2012 from http://elibrary.worldbank.org/doi/book/10.1596/1813-9450-3872

Kostoff, R. (2002), “Citation analysis of research performer quality”, Scientometrics, Vol.53, No.1, pp.49–71.

Mills, A.J., Durepos, G., and Wiebe, E. (2010), Encyclopedia of Case Study Research, Los Angeles: SAGE Publications.

MOF (2011), Budget Statement 2012: From Steady Foundation to Economic Transformation. Trinidad and Tobago. Ministry of Finance (Accessed March 28, 2012. http://www.finance.gov.tt/content/Budget-Statement-2012.pdf

Narin, F. (1976), Evaluative Biliometrics: The Use of Publication and Citation Analysis in the Evaluation of Scientific Activity, Computer Horizons Inc., Washington, D.C.

O’Shea, R. P., Allen, T. J., Morse, K. P., O’Gorman, C., & Roche, F. (2007), “Delineating the anatomy of an entrepreneurial university: The Massachusetts Institute of Technology experience”, R&D Management, Vol.37, No.1, pp.1–16.

Perkmann, M., and Walsh, K. (2009), “The two faces of collaboration: impacts of university-industry relations on public research”, Industrial and Corporate Change, Vol.18, pp.1033-1065.

Pertuzé, J., Calder, E., Greitzer, E., & Lucas, W. (2010), “Best practices for industry-university collaboration”, MIT Sloan Management Review, Vol.51, No.4, pp.83–90.

Philbin, S.P. (2010), “Developing and managing university-industry research collaborations through a process methodology /industrial sector approach”, Journal of Research Administration, Vol.41, No.3, pp.51-69.

Polt, W., Rammer, C., Gassler, H., Schibany, A., and Schartinger, D. (2001), “Benchmarking industry-science relations: the role of framework conditions”, Science and Public Policy, Vol.28, pp.247-258.

Rudd, E. (1988), “The evaluation of the quality of research”, Studies in Higher Education, Vol.13, pp.45-57.

Siegel, D. S., Waldman, D., and Link, A. (2003), “Assessing the impact of organizational practices on the relative productivity of university technology transfer offices: an exploratory study”, Research Policy, Vol.32, pp.27-48.

Smith, L. (1981), “Citation analysis”, Library Trends, Vol.30, No.1, pp. 83–106.

Tartari, V. S., D'Este, P., and Perkmann, M. (2010), “Come engage with me: The role of behavioral and attitudinal cohort effects on academics’ levels of engagement with industry”, Paper presented at the DRUID-DIME Academy Winter 2010 PhD Conference on ComwellRebild Bakker, Aalborg, Denmark,

January 21-23. The Chronicle of Higher Education (2007), Faculty Scholar

Productivity Index, available at:http: //chronicle.com/stats/productivity/.(Dated 8 March 2013).

The UWI (2010), Strategic Plan 2012-2017, The University of the West Indies, available at: http://www.uwi.edu/libraries/planningdocs/uwi_strategic_plan_2012-2017_final.sflb.ashx (Dated 8 March 2013).

Thune, T. (2011), “Success factors in higher education-industry collaboration: A case study of collaboration in the engineering field”, Tertiary Education and Management, Vol.17, No.1, pp.31–50.

Van Raan, A. (2005), “Fatal attraction: Conceptual and methodological problems in the ranking of universities by bibliometric methods”, Scientometrics, Vol.62, No.1, pp.133–143.

Varga, A. (1998), “University research and regional innovation: A spatial econometric analysis of academic technology transfers”, Norwell, MA: Kluwer Academic Publishers.

Verspagen, B. (2006), “University research, intellectual property rights and European innovation systems”, Journal of Economic Surveys, Vol. 20, pp. 607-632.

WTO (2011), Trade Policy Review: Trinidad and Tobago, World Trade Organisation, available at: http://www.wto.org/english/tratop_e/tpr_e/tp360_e.htm. (Dated 8 March 2013)

Wright, M., Clarysse, B., Lockett, A., and Knockaert, M. (2008), “Mid-range universities’ linkages with industry: Knowledge types and the role of intermediaries”, Research Policy, Vol. 37, pp.1205-1223.

Authors’ Biographical Notes:

Cary R. Cameron is a Master’s student in Engineering Management at The University of the West Indies. He received his B.Sc. in Chemical and Process Engineering from The University of the West Indies St. Augustine in 2008. Following his undergraduate career, Mr. Cameron worked as a Process Engineer in the bauxite industry. Here at UWI, his research interests have centred on university-industry collaborations and innovation policy in the Caribbean. Graham S. King is presently Lecturer in Mechanical Engineering at The University of the West Indies and a Chartered Engineer. He graduated from Loughborough University with a First-Class Honours degree in Automotive Engineering and from The University of Warwick, England, with a Masters degree in Engineering Management and an Engineering Doctorate. Dr. King accumulated 16 years of experience in the automotive industry before returning to academia and has been involved in multiple collaborative projects between universities and industry. Alongside teaching, Dr. King is presently serving as the Coordinator of the MMERC, and in this capacity builds and oversees collaborative projects with industrial, governmental and academic partners. ■

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 16: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

R.J. Rodriguez, and W.G. Lewis: Exploring Potential RF Hot Spot Locations in Confined Spaces of Large Wave Guide Dimensions 14

Exploring Potential RF Hot Spot Locations in Confined Spaces of Large Wave Guide Dimensions

Ricardo J. Rodriguez a,Ψ, and Winston G Lewis b

Department of Mechanical and Manufacturing Engineering, Faculty of Engineering, The University of the West Indies, St

Augustine, Trinidad and Tobago, West Indies aE-mail: [email protected]

bE-mail: [email protected] Ψ - Corresponding Author

(Received 8 April 2014; Revised 15 August 2014; Accepted 05 October 2014) Abstract: Radio-Frequency (RF) waves belong to the non-ionizing region of the electromagnetic spectrum and as such do not possess sufficient energy to ionize particles. They may however cause heating especially of human tissue which can therefore render them a potential hazard for workers and the public at large. International authorities have recognised the potential threat and have developed RF exposure guidelines to ensure some level of safety. These guidelines are developed for the electromagnetic spectrum but with a special interest in the Specific Absorption Rate (SAR) region (i.e., 30 to 1,000 MHz) where the strictest limits are set. In this part of the spectrum, a standing man can absorb most RF energy causing tissue within the body to resonate and generate heat. Therefore, while RF waves are used in modern technology in manufacturing, medical diagnosis and treatment, communications and navigation, there is some danger associated with its use that needs further investigation. The standards generally address RF propagating in free space and set limits for these but there also needs to be consideration for propagation in confined spaces where the waves store and build up their energies after reflecting off of inside surfaces and interfering with each other. The RF in the SAR region of the spectrum is of concern since most of our broadcasting towers operate between these frequencies and are within the public domain which contain confined spaces that may act as waveguides for these wavelengths. This paper focuses on quantifying location characteristics and intensities of RF propagating within empty rectangular structures likened to those of communication waveguide structures, through the proposal of a theoretical model for predicting the locations and intensities of RF hot zones containing RF hot spots. The significance of this work lies in being a tool for RF safety practitioners who can, without the use of cumbersome equipment and necessary skills for measuring RF, use a less expensive and user friendly method for determining the level of RF safety in a confined space by simply knowing its internal dimensions and material of the surfaces; and the RF source characteristics. The work can provide the IEEE standard body with useful information to include in its guidelines.

Keywords: RF hazard, SAR, confined space, reflection, waveguide

1. Introduction Waveguide technology has managed to constrain microwaves to remain within set boundaries with fixed frequencies that force the waves to behave differently to if they were moving in free space. This technology has offered the ability to efficiently transport energy for communication purposes other than transmission lines. The size of a waveguide may be a few centimetres in width and height and can guide microwaves with wavelengths of the order of centimetres, but what if spaces of larger dimensions are capable of being waveguides and can guide waves of larger wavelengths such as those that correspond to frequencies between 30 to 300 MHz? Such radio-frequency (RF) waves belong to the specific absorption rate (SAR) region of the spectrum where strict exposure limits are set for health and safety protection since a standing man acts as a dipole antenna for this radiation and can absorb

maximum energy from propagating RF waves. This review visits the likelihood for potential energy

build-up from RF propagation in rectangular confined spaces that are of waveguide design but with larger dimensions, by investigating the propagation characteristics within these confined spaces. Such confined spaces include silos, tanks, pipes, man-holes, air-condition ducts, tunnels, wells, engine rooms and operator rooms on board vessels. In these confined spaces waves reflect off of the walls and combine constructively or destructively with incident waves producing reinforcement or cancellation respectively. Where there is reinforcement, the intensity of the wave for a particular distance in accordance with the standard may be higher than expected for this distance from the source thereby exposing the worker to larger intensities than the proposed limit.

Standards are not fault proof and may from time to

ISSN 0511-5728 The West Indian Journal of Engineering

Vol.37, No.2, January 2015, pp.14-22

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 17: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

C.T. Benjamin and K.F. Pun: An Exploratory Study to determine Archetypes in the Trinidad and Tobago Fashion Industry Environment 15

time not fully address all the risks of hazards present since hazards may change depending on their location, intensity, source and presence with other hazards or other perilous conditions and interaction with artefacts. The standards therefore give a probability of safety but do not guarantee it, which should prompt safety practitioners to engage in methods of determination and analysis of the risks posed by hazards in a more inclusive and holistic manner (IEEE, 1999). 2. Literature Review There has been from time to time, public outcry that electromagnetic waves coming from the various sources cause persons to have headaches, develop rashes, cancer, nervous disorders, irritability, moodiness, breathing difficulties, muscle pain, mental depression, among other claims (Cember, 1996). To address the claims, a compromising position is taken in the IEEE C 95.1 standard with strict limits set at 30 to 300 MHz where the SAR (Rodriguez, 2008). The standards set guidelines generally for RF propagation in free space for plane waves, it is however difficult to accept that RF exposure in open or free space offers the same risk as within a confined space where the energy is confined within boundaries. The standards have not strictly focused on RF propagation in this range, inside of rectangular and cylindrical life-size structures likened to communication waveguides for microwave propagation. According to the Institute of Electrical and Electronics Engineers (IEEE, 1999), a RF hot spot is a highly localised area of relatively more intense radio-frequency radiation that manifests itself in localised areas in which there exists a concentration of radio-frequency fields caused by reflections and/ or narrow beams from antennae. Of special interest to worker safety therefore, is work performed in confined spaces (NCDOL, 2012).

Confined spaces pose potential threats to the health and safety of workers that are not readily visible, (latent hazards), and may lead to fatal injuries because there is failure to recognise and treat the hazard and the suitability of the emergency response (NCDOL, 2012). The OSHA of the United States has set a maximum exposure level of 10 mW/cm2 for frequencies between 10 MHz and 100 GHz averaged over a 6 minute period, while IEEE sets at 1mW/cm2 and E (electric) field at 61.4 V/m over the same period.

Given the input of these authorities there seems to be some level of concern and it would be prudent for interested parties to determine in advance the location and intensities of RF as it propagates through a confined space so as to set confined space design parameters and exposure guidelines to protect workers and the public. The literature reveals some of the work done by researchers in the development of propagation models for communication loss through buildings without paying attention to the harmful effects that may accompany this propagation. The models generally

introduce free-space propagation without considering restriction of wave movement due to building dimensions. This is justified since most models consider higher frequency or very small RF wavelength for losses in communication networks.

RF from urban communication antennas can reach confined spaces as they bounce off of high rise buildings, reflect off of walls and diffract off the roof tops to the road level (Walfisch and Bertoni, 1988). A study conducted by Iskander et al. (2002) was concerned with the deterioration of communication signals within confined spaces such as buildings and sites due to the multiplicity and complexity of constraints situated inside the building such as walls, furniture, humans, constructive and destructive interference. Therefore, before any planning of implementation of wireless communication there must be thorough understanding of path loss and propagation characteristics.

To predict propagation path we consider three models of interest which have been proposed: i) empirical model which uses a set of equations arranged according to field measurements, ii) site specific (ray tracing) model which uses ray tracing techniques but very computational and difficult to work with and iii) the theoretical model which uses assumed conditions in an ideal setting. Accuracy of ray tracing models depends on trajectory of launching ray, wall parameters, accuracy of locations and knowledge of field orientations which makes the model difficult to work with.

The shoot and bounce ray model (SBR) is one of a few ray models for predicting propagation after scatter but has drawbacks in determining whether rays are reflected rays from hitting something or whether the ray came straight or whether have in fact been hit or not and if so, which have been hit. The ray cone launching model was constructed to capture the spherical wave-front of the freely propagating wave at the receiving location where an observer will position himself to determine if the receiving point is within the cone. Ray cones are not the answer however since they overlap thereby creating ray doubling errors (Catedera et al., 1998). Interestingly, there have been several methods proposed to mitigate this effect. For instance, Durgin et al. (1997) use the method of distributed wavefront where the contribution of all the rays close to the receiver are taken into account and their sum is the total power received. The distributed wavefronts method is purported to improve the calculated fields but is found to be complex and inherently inaccurate (Porrat and Cox, 2003).

The ray trajectory models are significant but troubling to define so an image method is set to get the trajectory between the receiver and transmitter by placing reflector plates in the path of two transmitters. Lines connecting the receiver to the transmitter and the points of reflection and finally intersection are constructed and used to get the trajectory. A hybrid of image and SBR models are then used to identify the

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 18: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

C.T. Benjamin and K.F. Pun: An Exploratory Study to determine Archetypes in the Trinidad and Tobago Fashion Industry Environment 16

probability trajectory from transmitter to receiver (Tan et al., 1996).

Another propagation model is that of Liang and Bertoni (1998), who assume specular reflection at buildings, diffraction from roofs, building corners where buildings are put into groups of polyhedrons with similar characteristics, reflection coefficients at walls and diffused scattering coefficient. There is much disparity since architecture plays a major role in concrete refinement, windows, and glass, etc. The model identifies additional limitations such as phase of field, position accuracy of building, wall construction, local scattering from street lights, vehicles, people, fading signals. For smaller wavelengths surface roughness is of more concern than for larger waves since there is a tendency for greater diffused reflection.

Erceg et al. (1992) use a line of sight (LOS) model and addresses multiple reflections from walls before reaching the receiver. Attenuation, power of each ray using classical square law with a reflection coefficient that is dependent on the angle between the ray and reflecting surface out of sight (OOS), street. The received power reaching the antenna is then calculated using the square power law, path lengths, reflection coefficient and phase difference between rays reaching receiver.

Despite the advantage stated in the use of ray optics to determine location of waves, it falls short in describing such phenomena as diffraction and interference of waves and under these circumstances, wave optics will have to be used to fully address the wave behaviour. The extent to which interference is a concern for propagating RF is seen inside of smooth metal buildings where ‘radio dead spots’ occur during communication and signals become virtually non-existent.

These dead spots arise because of almost perfect, lossless reflections from smooth metal walls, ceilings or fixtures that interfere with the direct radiated signals. The dead spots exist in 3-dimensional space within the building and motions of only a few inches can move from no signal to full signal. In practice, not only metallic materials cause reflections, but dielectrics (or electrical insulators) also cause reflections. The actual signal levels reflected from insulators depend on a very complicated way on the above characteristics as well as the geometry of the situation.

Suffice it to say, that insulators are not as good at reflecting radio signals as metal surfaces, but even common insulating materials do cause some reflection of radio waves. Multipath occurs when all the radio propagation effects combine in a real world environment. In other words, when multiple signal propagation paths exist, caused by whatever phenomenon, the actual received signal level is the vector sum of all the signals incident from any direction or angle of arrival. Some signals will aid the direct path, while other signals will subtract (or tend to vector

cancel) from the direct signal path. The total composite phenomenon is thus called multipath. Two kinds of multipath exist: specular multipath -- arising from discrete, coherent reflections from smooth metal surfaces; and diffuse multipath -- arising from diffuse scatterers and sources of diffraction (the visible glint of sunlight off a choppy sea is an example of diffuse multipath).

Ray tracing however has been found to be a tedious task for both indoor and outdoor propagation given the obstacles encountered which scatter electromagnetic radiation. Site-specific models have been proposed that give predictions of propagation and scatter from walls and use statistics to predict those from people and chairs etc. Algorithms for model prediction fall into two main categories: ray shooting and method images. The first sets rays in all directions from a transmitter and the second deals with setting each obstacle as a virtual source from which radiation emanates (Porrat and Cox, 1993).

The literature generally takes into account attenuation of RF signals in the communication range of about 900 MHz and above and deals categorically with wall material, obstacle blockage such as machinery and LOS. Measurements on narrow and wide band propagation of RF in buildings have been classified into categories 1-8 describing the environment of propagation such as: residential houses and offices in sub urban, residential and office buildings in urban areas, open factory, open environment such as railway stations, airports and underground such as subways (Molkdar, 1991). The methods all used do not take into account the safety of the individuals in these RF infested areas but rather is concerned with signal losses. There is therefore room for model building that seeks out the propagation characteristics of RF in confined spaces to further improve the safety exposure guidelines for such waves. 3. Research Objectives The following objectives have been set for RF propagation within confined spaces acting as a transverse electric (TE10) mode waveguide:

a) determination and calculation of hot zone areas using simple trigonometry

b) determination and calculation of hot spot locations and intensities within hot zones using vector analysis and electromagnetic field theory

c) obtaining probability of exposure for single mode and multiple mode operations for TE10 confined spaces

4. Methodology A waveguide ray tracing model has been proposed in this research that avoids some of the problems encountered in the earlier models that use expanding spherical wave-fronts and measurements. The model

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 19: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

C.T. Benjamin and K.F. Pun: An Exploratory Study to determine Archetypes in the Trinidad and Tobago Fashion Industry Environment 17

proposed has waves confined to one plane making calculations simpler and does not use field measurements which have accompanied uncertainties and errors.

As RF waves propagate down a guide, they reflect off of the walls in a zig zag manner setting up standing waves with the majority of the energy from the electric field travelling down the centre of the space and tailing off at the walls, while the magnetic fields are tangential to the walls.

While energy within the guide does not fall off as rapidly as in open space, there are losses within the guide due to eddy current heating especially at higher frequencies (i.e., smaller wavelengths) where skin depth is present and is comparable to surface roughness. According to Mathew and Stephen (1968), any closed structure such as a pipe or confined space can propagate TE or TM modes. Waveguide transmission becomes more efficient at increasing frequencies which are suitable for microwave communication since the waves are confined and do not expand or diverge in accordance with an inverse square law for RF propagation. (Whitaker, 2002).

Figure 1. Picture of typical waveguide for communication Source: http://en.wikipedia.org/wiki/File:Waveguide.svg

An electromagnetic field can propagate along a

waveguide in various ways. Two common modes are known as transverse-magnetic (TM) and transverse-electric (TE). In TM mode, the magnetic lines of flux are perpendicular to the axis of the waveguide. In TE mode, the electric lines of flux are perpendicular to the axis of the waveguide (Chatterjee, 1968). At any frequency above the cut-off (the lowest frequency at which the waveguide is large enough), the feed line will work well, although certain operating characteristics vary depending on the number of wavelengths in the cross section (Mathew and Stephenson, 1968).

Waves of frequencies between 20-3,000 MHz are used for line of sight communications in aircrafts, FM, TV and amateur radio broadcasting. Wavelengths at these frequencies are comparable to the dimensions of trees, buildings, hills and as such constitute good reflecting surfaces for setting up constructive and destructive interference between reflected and incident of generated signals. The same is true for propagation of signals between two layers of air at different temperatures (tropospheric ducting) acting as a duct with dimensions that propagate certain modes of transmission. This duct acts as a waveguide for certain frequencies depending on the dimensions and wavelength of the signals (Whitaker, 2002).

Table 1 shows typical waveguide dimensions for microwaves (Wikipedia, 2014). The last column, column 5, shows the inner dimensions of width and height and it can be observed that the ratio is 2:1. For this condition, propagation of waves can exist without much loss in energy. For such dimensions, there can be a number of possible modes by which energy can be transferred with each being characterised by a distinctive field configuration.

Waveguide propagation can be hypothesised for empty, rectangular spaces of dimensions similar to the wavelength of the propagating RF. For such spaces, there will be a cut-off frequency below which waves will not propagate, all frequencies above may pass. If the guide is however operating in single mode operation, then only one frequency is allowed to pass. According to ECE (2014), the formulae for wave propagation inside a rectangular wave guide are as shown:

1. c = 3x108 m/s (c= speed of light in vacuum) (1) 2. λcmn = 2/ √[(m/a)2 + (n/b)2]: (cut-off wavelength

for m, n modes for m=1,2,3., n = 0,1,2,3. (2) 3. Sin θ = λ/λc = fc/f: θ is angle of incidence at

entrance of guide (3) 4. fcmn = c/2 {√ [(m/a)2 + (n/b)2]}: cut-off

frequency for m, n modes (4) For any wave propagating down a waveguide, it

does so in the Z direction with a group velocity given by Ug. In so doing, a stationary wave pattern is set up as the waves bounce off of the walls of the guide in a ‘zig-zag’ manner in the XZ plane with velocity in direction of wave travel, Uu, (see Figure 2).

Table 1. Standard sizes of waveguides * - Radio Components Standardisation Committee

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 20: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

C.T. Benjamin and K.F. Pun: An Exploratory Study to determine Archetypes in the Trinidad and Tobago Fashion Industry Environment 18

There is therefore a component of wave motion in both Z and X directions. The Y-axis addresses the magnitude of the electric field E only in the TE10 mode and there is no component of E in either Z or X directions. The E field spreads across the width of the guide and is half wavelength of the propagating RF for the TE10 mode with maximum field strength at the centre and falls off to zero at the walls. For the TE20 mode, the width is two ½ wavelengths and the E field distribution is E+ and E- across the width corresponding to maxima at top and bottom (ECE, 2014).

Should RF waves enter a rectangular confined space at some angle θ to the Z axis (angle of attack), then the entrance would be filled with parallel rays all striking different points on the inside walls of the confined space and reflecting. If the confined space is operating as a waveguide in single mode operation (only one frequency allowed to propagate), then the ray pattern looks like Figure 2, where triangular zones (hot zones) are set up with defined boundaries of propagating and reflected rays from within the guide. Hot spots are found within the confines of the hot zones and occur if the path difference between the reflected and incoming ray is a whole number multiple of the wavelength (i.e., n = 0, 1, 2, 3, 4, etc.). It is not necessarily critical to know exactly where these hot spots are but may be of more importance to know where their probabilities lie within the confined space. This is achieved by first detecting the hot zone locations and the areas of the hot zones. Figure 2. Confined space of width ‘a’, as a single mode operation

waveguide for cut-off frequency in TE10

To illustrate this hypothesis let us set up a simple

scenario of a propagating RF from an external antenna source for both single and multiple mode operations for a confined space acting as a waveguide. Consider that a rectangular high reflectance air-condition duct of width a = 1.75m and height b = 0.88m is in the line of sight of a RF broadcasting antenna. The incident plane waves propagating some distance from the antenna strike the open rectangular duct (waveguide) at 60 degrees (angle of attack) to the Z-axis of the duct. Determine which frequencies can propagate through this duct and thence

the sizes of the hot zone areas so produced if the space is to operate as a waveguide for the TE10 mode in single and multiple modes of operation. 4.1 Single Mode Operation The cut-off frequency fcmn = c/2 {√[(m/a)2 + (n/b)2]}; m=1 n=0 for TE10 mode which is 85.7MHz for this guide dimension. For width ‘a’ = 1.75m and angle of incidence 60o, then f = fc / sin 60 which is equal to 98.9 MHz. If the antenna emits at 98.9 MHz, then since this frequency is higher than the cut-off, it should be able to propagate down the guide, air-condition duct in this case, once excited. For guide operating as a single mode operation, i.e., for the TE10 mode only, waves with frequency 98.9 MHz that strike the entrance at 60o will propagate and fill up between the points A and Z. Incident waves strike the wall of the duct at Z and G and at 30o to the normal (n0) at the wall’s surface and reflect at the same angle 30o and continue to E and F. These will then reflect from this surface to strike the first surface at H and L (see Figure 2).

Interference occurs within triangles: OZG, PEF and RHL where incident and reflected rays cross. These are considered areas of potential RF hot spots. The hot zones are all of same area and repeat periodically and give the probability of RF hot spot accumulation, where the waves interfere constructively. These hot spots will have higher field intensities than the waves that do not interfere constructively and are therefore potential points of higher exposure. 4.2 Multiple Mode Operation For the fixed waveguide width, a = 1.75m, the cut-off frequency is 85.7 MHz for the dominant TE10 mode. Varying frequencies of RF can propagate outside the single mode operation but their angles of attack will be different. Theoretically, using geometric construction or ray tracing technique, the smaller the angle then the larger the hot zone area as seen in Table 2 and Figure 3.

We can infer that the probability or likelihood of finding hot spots decreases with increasing angle of attack, i.e. the areas get smaller and smaller so interference patterns decrease. Based on this it should also be noted that the allowed propagating frequency decreases with increasing angle of attack (see Figure 4).

Figure 5 shows that as the guide width increases the cut-off frequency decreases, broadcasting frequencies are expected to be between the 50 and 150 MHz frequency band generally. Confined spaces with dimensions corresponding to values between 1m and 3m in width may therefore be waveguides in the TE10 mode for these signals.

One deliverable so far based on our discussions is the construction of a SAR structural guide chart that predicts whether there is potential threat to RF exposure within a structure from a nearby source (see Figure 6). This of course is for the TE10 mode and can be extended

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 21: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

C.T. Benjamin and K.F. Pun: An Exploratory Study to determine Archetypes in the Trinidad and Tobago Fashion Industry Environment 19

to other modes as well as multiple mode operations and can be useful in building design drawings.

Table 2. Zone area and propagating frequency (fc= 85.7MHz) for varying angles of θ.

θ/ o

Tan θ Area of zone A = .7656/ tan θ

Frequency MHz (f = fc/ sin θ)

Sin θ

2 0.35 21.87 2857 .03 3 .052 14.72 1714 .05 4 .070 10.94 1224 .07 5 .0875 8.75 952 .09 6 0.105 7.29 779 .11 7 .123 6.22 714 .12 8 .141 5.43 612 .14 10 .176 4.35 504 .17 20 .36 2.13 252 .34 30 .58 1.32 171.4 .5 35 .70 1.09 150.3 .57 50 1.19 .649 112.3 .77 60 1.73 .443 98.5 .87 70 2.75 .279 91.1 .94 75 3.73 .205 88.3 .97 80 5.67 .13 87.0 .985 85 11.40 .067 86.0 .996 86 14.3 .054 85.957 .997 87 19.08 .04 85.786 .999 88 28.6 .027 85.7857 .999 89 57.3 .013 85.717 .9998

Figure 3. Graph of hot zone area versus angle of attack, θ Figure 4. Graph of propagating frequency versus angle of attack, θ

Figure 5. Graph of guide width versus cut-off frequency

Figure 6. SAR structural guide chart (Cut-off frequencies in TE10 mode for varying rectangular confined space widths ‘a’ in Single

Mode Operation). Confined space name

a/ m b/ m fc/ MHz: fcmn = c/2

{√[(m/a)2 + (n/b)2]

Single mode

freq. at 600 /MHz

SAR

CS300 3.00 1.50 50.0 57.7 Yes CS275 2.75 1.35 54.5 62.9 Yes CS250 2.50 1.25 60 69.3 Yes CS225 2.25 1.12 66.8 77.1 Yes CS200 2.00 1.10 75 86.6 Yes CS175 1.75 1.40 85.7 98.9 Yes CS150 1.50 0.75 100 115.5 Yes CS125 1.25 0.65 120 138.5 Yes CS110 1.10 0.55 136 157.0 Yes

5. Calculation and Analysis 5.1 Hot Zone Areas The area of a hot zone is given by a2/4 tan θ, where θ is the angle of attack of the incoming waves to the xz plane of the space and ‘a’ is the width of the space. As θ decreases the zone area increases up to the guide length otherwise the hot zone area within the guide begins to decrease. In this regard, the effective area of the hot zones found for a given width within the guide is therefore restricted by the length of the guide and defined for: θ < tan-1 a/l, where l is the length of the guide.

A = ½ (a/2)l, tan θ = a/2 /l/2 = a/l Therefore, l = a/ tan θ, and A = a2/ 4tan θ The maximum area AO of the hot zones is therefore

a2/4 tan-1a/l. So for a guide of width 1.75m as pointed out in our scenarios earlier, with a guide length of say, 4m, the minimum angle for optimal area is 23.6 degrees and the max hot zone area is therefore A0 = a2/ 4 tan 23.6 = 1.76m2. Larger angles of attack will see smaller frequencies, as long as above the cut-off frequency, propagate and hence smaller and less zone areas which means smaller probabilities of exposure. Figure 7 shows smaller angle giving smaller area inside guide.

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 22: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

C.T. Benjamin and K.F. Pun: An Exploratory Study to determine Archetypes in the Trinidad and Tobago Fashion Industry Environment 20

Figure 7. Grid with zone areas

5.2 Hot Spot Location The hot spot locations are identified by tracing any two rays from their origin at the entrance of the guide to their point of intersection and if the path difference ∆ between them is a whole number of wavelengths, nλ, then we have constructive interference and a hot spot. We first divide the guide space into a grid of equal squares. Using vector analysis, vector r3 is the resultant of vectors r1 and r2 (see Figure 8). Note that (r4 – r3) is the path difference ∆, which is the vector along the entrance width. From Figure 8, r3 = 8i+j and r4 = 8i+9j, so r4 – r3 = 8j from which the magnitude ∆ becomes 8 units as seen on the grid.

Figure 8. Vector analysis for guide hot spots

For example, from earlier guide width ‘a’ = 1.75m,

which is divided into 14 units on the grid so each unit = 1.75/ 14 = 0.125m, which implies that 1m = 8 units. See Table 2 for f = 98.5 MHz which gives λ= c/f = 3x108/ 98.5x106 = 3.1m, for angle of attack of 60 degrees. Is the point Q a hot spot for the two rays leaving the guide entrance? We must equate this to nλ, where λ=3.1m. As we see; n =1/3.1 < 1,2,3 i.e., whole numbers so this is not a hot spot. If we allow for a higher frequency and

hence larger area by making the angle of attack lower, then say for angle 30 degrees, frequency = 171.4MHz and λ= 1.75m (see Table 2). This implies that n = 1 so we have a hot spot for the lower angle and wider area. There is a potential hot spot then in the guide provided that the length is long enough to accommodate the zone area. We can therefore set up a table for the TE10 guide for different frequencies above the cut-off of 85.7 MHz and see what path differences and frequencies would lead to hot spots.

Our earlier notion of the number of hot spots increasing as the frequency increases or as θ decreases has been verified theoretically. This does not discount that there may already be an existing threat by its normal presence. For this guide then, there is greater probability of higher exposure at 952 MHz (5 hot spots) than there is for 504 MHz (3 hot spots) and similarly for 171.1 MHz (1 hot spot).

Table. 3. The number of hot spots with increasing frequency Frequency,

f/ MHz λ/m θ/

degrees ∆/ m n Frequency,

f/ MHz 98.5 3.1 60 1 <1 no 171.4 1.75 30 1.75 1 yes 504 0.59 10 0.59 1 yes 504 0.59 10 1.75 3 yes 504 0.59 10 1.18 2 yes 952 0.315 5 0.315 1 yes 952 0.315 5 1.575 5 yes 952 0.315 5 0.63 2 yes 952 0.315 5 0.945 3 yes 952 0.315 5 1.26 4 yes

5.3 Hot Spot Intensities Earlier we obtained a path difference between a resultant ray and incident ray, r3 and r4 respectively which set the stage for determining whether a hot spot was present or not. Once a hot spot is found to be present, we must now calculate the net intensity of the superimposed waves at the point. Accompanying the path difference is a phase difference of 2π∆/λ, which has been left out here but could be easily placed in the reflected wave as rE = rE0 ej(ωt – [β+2π∆/λ]z) . ax

The field intensity of the incoming vertically polarised wave is Ei = E0 ejωt. e-γz, (unbounded) where γ = α + jβ and γ is imaginary if (attenuation factor) α = 0. i.e. there is no attenuation of the wave. Inside the confined space, iE = iE0 ej(ωt - βz) . ax and rE = rE0 ej(ωt - βz) . ax , the sum of the two rays arriving at Q: E(z,t) = iE + rE = iE0 ej(ωt - βz) . ax + rE0 ej(ωt - βz) . ax. (1) Both E field intensities still travelling in the forward ‘z’ direction.

Equation (1) gives the time varying E field in the z direction and the stationary wave set up in the x direction, where ω is the frequency, β is the phase constant of the waves and ax is the unit vector in the x direction. The waves are vertically polarized with no

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 23: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

C.T. Benjamin and K.F. Pun: An Exploratory Study to determine Archetypes in the Trinidad and Tobago Fashion Industry Environment 21

change upon reflection. Fresnel relationship for waves striking a boundary

with E field normal to the plane of incidence (xz plane) and parallel to the boundary (zy plane):

rEo/ iEo = [θ i cos θ t – θ tcos θ t]/ [θ i cos θ i + θ tcos θ t]. (2)

where θ i and θ t are the intrinsic impedances of medium 1 and medium 2 which could be air and aluminium respectively and are calculated from: θ = √(µω/σ) and θ i and θ t are the angles of incidence and transmission of the ray striking the boundary interface (Hecht, 1975).

Assuming perfect reflection and no transmission through the boundary, then θ t = 0, and eqn 2 becomes:

rEo/ iEo = [θ i cos θ i – θ t]/ [θ i cos θ i + θ t] (3) Therefore, rEo = [θ i cos θ i – θ t]/ [θ i cos θ i + θ t] . iEo which can be written as:

rEo = θ iEo (4) Equation (1) becomes:

E(z,t) = = iE0 ej(ωt - βz) ax + θ iE0 ej(ωt - βz) . ax. (5) So, E(z,t) = iE0 ej(ωt - βz) (1 + θ) . ax (6)

This can be further written as: E(z,t) = iE0 ej(ωt - βz) (1 + θ) Sin kxx. (7)

which gives the travelling part of the electric field and the stationary part of the standing wave which essentially is the electric field intensity of the constructive wave at the point Q. this can be further simplified using the imaginary part of (7) to get

E(z,t) = iE0 Sin (ωt – βz) . (1 + θ) Sin kxx (8) 6. Results From above, for a hot spot with path difference, ∆ = x due to two intersecting waves, the electric field at some point Q say, within the hot zone of these intersecting waves is given by (9)

E(z,t) = iE0 Sin (ωt – βz) . (1 + θ) Sin kx∆ (9) The paper does not specifically deal with diffraction

effects at the entrance of the guide, some diffused diffraction at the surfaces, phase change at Q and that the guide will not be a perfect conductor. There was also the assumption that there would be total reflection and no transmission or very little at the boundary interface. 7. Conclusion This article presents a thought provoking notion of large waves in SAR region propagating down rectangular confined spaces just as microwaves would in waveguides setting up standing waves due to interference of incident and reflected waves within the space. These RF hot spots possess the potential to contain higher field intensities than if the waves were propagating in free space. When we therefore investigate safety hazards due to noise, lighting, toxic and gas build-

up within confined spaces, accumulation of RF along the confined space must also be studied.

While the discussion is purely speculative and requires experiment for validation, it seems justifiable theoretically, that RF hot zones can exist in confined spaces. The determination of these RF hot zones can be of importance in setting building code guidelines for construction of antennae and RF towers close to communities. The work presented here is limited in that only rectangular, high reflective, empty confined spaces have been brought to the fore. Confined spaces are expected to be used by people and contain artefacts such as furniture, tools, etc which makes a very complex situation of multiple reflections and diffraction effects. The intention however was to deliver the trivial case and then extend to the more complex.

Moreover, other modes of operation (such as TE11, TE21, and TE31) were not discussed since the dominant mode was thought to be sufficient and does not infer that the other modes are less important. It is hoped that with experimental work and further research that the contribution offered here would find resting place in RF safety guidelines and RF propagation hot zone management in confined spaces. References: Bhattacharya. A., Gupta S. and Chakraborty, A. (1999), “Analysis

of rectangular waveguides and thick windows as EMI sensors”, Progress in Electronics Research, Vol.22, pp.231-258

Catedra M.F, J.Perez, F.S. de Anana and O.Gutierrez. (1998), “Efficient ray tracing techniques for three dimensional analyses of propagation in mobile communications: Application to picocell and microcell scenarios”, IEEE Antennas Propagation Magazine, Vol.40, April, pp.15-28.

Cember, H. (1996), Introduction to Health Physics, McGraw Hill, New York

Chatterjee, R. (1968), Elements of Microwave Engineering. Ellis Horwood series in electrical and electronic engineering. John Wiley, New York.

Cleary J.G, and Wyvill, G. (1998), “Analysis of an algorithm for fast ray tracing using uniform space subdivision”, Image and Vision Computing, Vol.4, pp.65-83.

Durgin, G., Patwari, N., and Rapport, T.S. (1997), “Improved 3D ray launching method for wireless propagation prediction”, Electronics Letters, Vol.33, pp.1412 - 1413.

ECE (2014), Waveguides, Accessed July 2014 from: http://www.ece.msstate.edu/~donohoe/ece3323waveguides.pdf

Erceg V., Ghassemzadeh, S., Taylor, M., Li, D., and Schilling, D.L. (1992), “Urban/ suburban out of sight propagation modelling-2GHz”, IEEE Communication Magazine, June1992, Vol.30, pp.56-61

Fortune S.F, Gay, D.M., Kerninghan, B.W., Landron, O., Valenzuela, R.A., and Wright, M.H. (1995), “WISE design of indoor wireless systems: Practical computation and optimisation”, IEEE Computing in Science and Engineering Magazine, Spring, pp58-68.

Hecht, E, (1975), Optics. McGraw Hill, New York IEEE (1999), IEEE Standard for Safety Levels with Respect to

Human Exposure to Radio Frequency Electromagnetic Fields, 3 KHz to 300 GHz (CP5.1-1999), Institute of Electrical and Electronics Engineers, New York

Iskander. F, Magdy, F. and Zhengqing,Y. (2002), “Propagation prediction models for wireless communication systems”, IEEE

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 24: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

C.T. Benjamin and K.F. Pun: An Exploratory Study to determine Archetypes in the Trinidad and Tobago Fashion Industry Environment 22

Transactions on Microwave Theory and Techniques, Vol.50, No.3, pp.1-5

Liang, G. and Bertoni, H.L. (1988), “A new approach to 3-D ray tracing for propagation prediction in cities”, IEEE Transactions on Antenna and Propagation, Vol.46, June, pp.853-863

Mathews, P. and Stephenson, I. (1968), Microwave Components, Chapman and Hall Ltd, 11 New Fetter Lane, London (Chapter 1, p.3-27).

Molkdar, D. (1991), “Review on radio propagation into and within buildings”, Proceedings of Institution of Electrical Engineers, Part. H, Vol.138, pp.61-73

NCDOL (2012), A Guide to Radio Frequency Hazards with Electric Detonators, N.C. Department of Labour Occupational Safety and Health Division, 1101 Mail Service Center Raleigh, NC 27699-1101, Cherie Berry, Commissioner of Labour

Okumura, Y., Ohmori, E., Kwano, T., and Fukuda, F. (1968), “Field strength variability in VHF land mobile service”, Review of the Electrical Communication Laboratory, Vol.16, September-October, pp.825-873.

Porrat, D. and Cox, D.C. (2003), UHF Propagation in Indoor Hallway, Accessed July 2014 from: http://wireless.stanford.edu/papers/UHFpropagation_hallwaysPorrat.pdf

Rappaport, T.S. (1996), Wireless Communications: Principles and Practice, IEEE Press, p.71, 130-131

Rodriguez, R. (2008), A HSE Radiation Management Model for the Workplace, (Unpublished thesis), Department of Mechanical and Manufacturing Engineering, The University of the West Indies, St Augustine, Trinidad and Tobago.

Seybold, J. S. (2005), Introduction to RF Propagation, John Wiley, New Jersey, p.11

Tan, S.Y. and Tan, H.S. (1996), “A microcellular communications propagation model based on the uniform theory of diffraction and multiple image theory”, IEEE Transactions on Antennas Propagation, Vol.44, October, pp.1317-1326

Walfisch, J. and Bertoni, H. (1988), “A theoretical model of UHF propagation in urban environments”, IEEE Transactions on Antennas and Propagation, Vol.36, No.12, December, pp 1788-1796

Whitaker, J. (2002), The RF Transmission Systems Handbook, CRC Press LLC, Boca Raton, Florida 33431, p.15-13.

Wikipedia (2014), Waveguide (electromagnetism): Standard Sizes of Rectangular Waveguide, Accessed July 2014 from http://en.wikipedia.org/wiki/Waveguide_(electromagnetism)

Authors’ Biographical Notes:

Ricardo Rodriguez is currently pursuing a PhD degree in Mechanical Engineering under the supervision of Professor Winston Lewis. His research concerns itself with the modeling of RF in confined spaces. Mr. Rodriguez holds a BSc degree in Pure and Applied Physics, MSc in Engineering Management and MPhil in Mechanical Engineering from The University of the West Indies. He was also awarded a special prize for his contribution on Opportunities Management at the Prime Minister’s Award Function, 2004 in the Bright Solutions category. He has worked at the Trinidad and Tobago Bureau as a Standards Officer III where he developed standards on gasoline, diesel and Health and Safety Standards in support of the Occupational Safety and Health Act. Mr. Rodriguez was also the Co-Chair with a Telecommunications Authority of Trinidad and Tobago representative on a committee mandated to review the local radiation exposure guidelines for Trinidad and Tobago. Winston G. Lewis is presently Professor of Industrial Systems Engineering of the Faculty of Engineering at The University of the West Indies. He is a registered professional engineer and Fellow of the Association of Professional Engineers of Trinidad and Tobago (APETT), and a member of the Order of International Fellows in the Association for Iron and Steel Technology and the American Society of Mechanical Engineers (ASME). He was a member of the steering committee appointed for the establishment of The University of Trinidad and Tobago (UTT), and Chairman for the development of national standards on workplace design at the Trinidad and Tobago Bureau of Standards. Professor Lewis’ research and development work is in metallurgical and Industrial engineering, sheet metal forming, manufacture of the steel-pan musical instrument, applied ergonomics and workplace design, engineering quality management and nano-technology. He has authored 77 academic papers with 39 Journals and 38 Conference Papers and has successfully supervised one PhD candidate, three MPhil, and forty-four graduate students with their MSc, final year projects. ■

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 25: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

R. Hosein and R. Mayrhoo: A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad Gas Condensates 23

A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad Gas Condensates

Raffie Hosein a,Ψ, and Rayadh Mayrhoo b

Petroleum Studies Unit, Department of Chemical Engineering, Faculty of Engineering, The University of The West Indies,

St Augustine, Trinidad and Tobago, West Indies aE-mail: [email protected]

bE-mail: [email protected] Ψ - Corresponding Author

(Received 18 April 2014; Revised 16 July 2014; Accepted 18 October 2014) Abstract: An accurate description of pseudo-component compositions is required for Equation of State predictions for gas condensate systems. These extended experimental data are often unavailable and must be generated using mathematical models, of which the exponential and the three-parameter gamma distribution functions are the two most widely used. The development of these two techniques was based on the assumption of a continuous molar relationship for pseudo-components. However, experimental compositional data for gas condensate systems show discontinuities in this relationship at Single Carbon Number (SCN) 8 and 13. The models when applied to extend the heptanes plus (C7+) fraction for Trinidad gas condensates, under-predict the SCN8 mole percent and over-predict the SCN12 mole percent due to the aforementioned discontinuities. The Average Absolute Deviation between the predicted and experimental SCN8 and SCN12 data were both greater than 25 percent. The two-coefficient method described by Ahmed et al. (1985), when applied to extend the C7+ fraction, reduced the discontinuity at SCN8 to less than 12 percent. However the SCN12 group still had a deviation greater than 18 percent. These results show that existing models were not designed to take care of these discontinuities and should be used with caution when extending experimental data beyond SCN 7. The Model described in this study resolves these discontinuities in the molar relationships at both SCN8 and SCN12 with an Average Absolute Deviation between the predicted and experimental compositions of less than 10 percent. This model can quite easily be included in Equation of State packages for a more accurate description of compositions for Trinidad gas condensates for performing compositional simulation studies. A partial analysis beyond the C7+ fraction is not required with this new model.

Keywords: Trinidad, gas condensate, plus fraction, Single Carbon Number (SCN), extended analysis, coefficient model

Nomenclature AAD = average absolute deviation A, B = constants in equation 1 bbls = barrels C = Carbon CGR = Condensate Gas Ratio MW = Molecular Weight P = Pressure, psia PVT = Pressure, Volume, Temperature S = Coefficient scf = Standard Cubic Feet SCN = Single Carbon Number SG = Specific Gravity

1. Introduction An accurate description of the compositions of the Single Carbon Number (SCN) groups (pseudo-components) is an integral part of the reservoir fluids characterisation process (Ahmed, 1989; Danesh, 1998). For gas condensate systems these data are applied with Equations of State (EOS) to evaluate gas and condensate reserves and production for field development and surface facility design. Very often the required extended

compositional data are unavailable experimentally and are generated from mathematical models historically known as “splitting schemes”.

Models to extend the composition beyond the measured plus fraction are included in EOS simulation packages (Ahmed, 1984, 1989; Danesh, 1998; Whitson and Brule, 2000). The two most extensively used are the exponential (Pedersen et el., 1984, 1985, 1989) and the three-parameter gamma distribution function (Pearson,

ISSN 0511-5728 The West Indian Journal of Engineering

Vol.37, No.2, January 2015, pp.23-30

UWI = University of the West Indies zn = mole percent of SCN fraction

Subscript

i = Component i n = SCN fraction number + = plus or last fraction

Greek α, β, η = parameters in the gamma distribution function.

Г = gamma function > = greater than < = lesser than

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 26: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

R. Hosein and R. Mayrhoo: A Model for Extending the Analysis of the Heptanes plus Fraction for Trinidad Gas Condensates 24

1895). The use of the latter for this purpose was suggested by Whitson (1983). These models (Yarborough, 1978; Pedersen et al., 1989; Whitson, 1983) are applied to gas condensate systems with the assumption that there is a continuous relationship between pseudo-component composition and molecular weight. This assumption was based on compositional data limited to North Sea gas condensate systems and more recently with data from a wider region (Al-Meshari and McCain, 2007).

Literature on simulation models recommends using experimental compositional data beyond C7+ when applying extended models to obtain a more accurate description of pseudo-component compositions. It was not until recently that the limits of the required experimental data were defined for improved predictions (Hosein and McCain, 2009) when using the aforementioned models. These limits were based on discontinuities observed in the molar relationships at SCN8 and at SCN13 from experimental compositional data for gas condensates measured with samples with worldwide origin.

In this study, the limitations of the exponential and three-parameter gamma distribution functions were reviewed and a new model for a more accurate description of pseudo-component compositions is described for Trinidad gas condensate systems. The initial condensate-gas ratios of the samples tested were less than 50 stock-tank barrels per million standard cubic feet of separator gas.

2. Compositional Data Sets used in this Study A total of twelve (12) sets of compositional data (Appendix A, Table A-1) were used in this study. The compositions for samples PL1 to PL6 were generated experimentally by a commercial laboratory whereas the compositions for samples CL1 to CL6 were generated experimentally in the UWI PVT Laboratory (Hosein, 2004). The samples were taken from separators for various producing gas reservoirs located offshore the Southeast coast and North coast of Trinidad, (Hosein, 2004). The compositions of the C7+ fractions were less than 4.0 mole % which would classify Trinidad gases as lean gas condensates (McCain, 1990). 2.1 Chromatographic Experiment Analyses of separator gas and condensate samples were conducted by a commercial laboratory on a VARIAN gas chromatograph (GC). The instrument was custom designed with sampling valves for injection, packed columns connected to a thermal conductivity detector (TCD) and a capillary column connected to a flame ionisation detector (FID) as shown in Figures 1 and 2.

The arrangement provides the requirements of GPA Methods 2286-95 and 2186-95 for extended analyses of Natural Gas and Natural Gas Liquids respectively. The analysis of the separator gas was combined with the

analysis of the separator liquid to yield the composition of the reservoir fluid or well stream (Hosein, 2004).

Figure 1. Photograph of Gas Chromatograph

Figure 2. Schematic of Gas Chromatograph

2.2 Observed Relationship between Experimental SCN Composition and Molecular Weight.

Yarborough (1978) and Pedersen et al. (1984) observed that plots of mole fraction versus molecular weight (or SCN group) for North Sea gas condensates usually exhibit a continuous exponential distribution function as shown in Figure 3, or a linear relationship as shown in Figure 4 (i.e., log of mole % versus molecular weight or semi-log plot). Hosein (2004) and Hosein and McCain (2009) provided experimental data on samples from worldwide locations that show evidence of distinct discontinuities at SCN8 and SCN13 (see Figures 5 and 6). They suggested that these discontinuities should be taken into consideration for a more accurate description of SCN compositions when using extended models for gas condensates.

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 27: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

R. Hosein and R. Mayrhoo: A Model for Extending the Analysis of the Heptanes plus Fraction for Trinidad Gas Condensates 25

Figure 3. Molar Distribution of SCN Composition for

North

Figure 4. Molar Distribution of SCN Composition for North (Source: Pedersen et al. 1989)

3. Models tested for extending the Plus Fraction for

Gas Condensate Systems 3.1 The Exponential Distribution Function Pedersen et al. (1984) expressed the observed continuous exponential relationship into a continuous linear one with a logarithmic expression for mole percent as a function of molecular weight as follows:

log zn =A + B (Mn) …..(1) where: zn = composition of SCN group n, mole percent Mn = molecular weight of SCN group n. A, B = constants determined by the least squares fit to

the experimental data. This generally accepted representation of a single

straight line relationship is shown in Figure 4. With this model, the Average Absolute Deviation (AAD) obtained between the predicted and experimental compositions for the twelve data sets is shown in Table A-2 and Figure A-1. The compositions of the SCN8 groups were under predicted by more than 25 percent whereas the compositions of the SCN12 groups were over predicted by more than 30 percent due to the discontinuities in the molar relationship at SCN8 and SCN13 as shown in Figures 5 and 6, respectively. Hosein and McCain

(2009) suggested that extended experimental data up to C20+ is required when applying this model. This would provide a minimum of seven experimental data points to define the discontinuity at SCN13 and beyond as shown in Figure 6. Hence, this scheme is more suitable for predicting composition beyond the SCN19 group.

Figure 5. Log Mole % versus Molecular Weight for Trinidad Sample PL6

Sources: Hosein (2004); Hosein and McCain (2009)

Figure 6. Log Mole % versus Molecular Weight for Trinidad Sample PL6

Sources: Hosein (2004); Hosein and McCain,)2009) 3.2 The Three-Parameter Gamma Distribution

Function The three-parameter gamma probability function (Pearson, 1895) is used to characterise molar distribution (i.e. mole percent and molecular weight relation of pseudo components) as follows:

P(x) = (x - η) α -1 exp [- (x - η) / β] / β α Г (α) …. (2) where: α, β and η are parameters defining the

distribution (Whitson 1983). The basic assumption that is made when applying this

model to gas condensate systems is also a continuous (exponential) relation between SCN composition in mole

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 28: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

R. Hosein and R. Mayrhoo: A Model for Extending the Analysis of the Heptanes plus Fraction for Trinidad Gas Condensates 26

percent and molecular weight (Whitson, 1983). This occurs when the parameter α = 1. When extending the C7+ fraction, the value of η = 86.177, which is the molecular weight of n-C6H14 (Al-Meshari and McCain, 2007). The Gamma model as described by Al-Meshari and McCain (2007) was applied to predict SCN compositions for the twelve data sets in Table A-1 of Appendix A. The end points for the integral for each frequency of occurrence calculation used to calculate the SCN compositions were the molecular weights of the successive normal paraffins.

The AAD obtained between the predicted and experimental compositions of the SCN groups for the 12 data sets are given in Table A-2 and Figure A-1 of Appendix A. These results show that the compositions of the SCN8 groups were under-predicted by more than 25 percent whereas the compositions of the SCN12 groups were over-predicted by more than 25 percent, illustrating that this model does not take care of the discontinuities in the molar relationship at the SCN8 and the SCN13 groups when extending the C7+ fraction. Hosein and McCain (2009) demonstrated that extended experimental data up to the C14+ fraction are required for best prediction with this model. 3.3 Ahmed et al. (1985) two Coefficient Splitting

Scheme The model devised by Ahmed et al. (1985) was based on observation that hydrocarbon systems exhibit a molar distribution relative to the average molecular weight of the plus fraction. They described a “marching technique” from which the average molecular weights of the plus fractions (Mn+) are calculated from experimental compositional data (see Appendix A). They used plots similar to Figures 7 and 8 to prepare generalised coefficients (S) for a two-segment (two-coefficient) relationship to calculate mole percent of SCN groups as follows:

S =15.5 for SCN = 8 and S = 17.0 for SCN > 8 The plots were expressed by the generalised equation:

M n + = M 7 + + S (n-7) …..(3) The Ahmed et al. (1985) splitting scheme was tested

using the 12 sets of compositional data from Table A-1. The most significant observation was that the two segment relationship improved the prediction of the SCN8 and SCN12 compositions when compared to the exponential and gamma distribution functions as shown in Table A-2 and Figure A-1 of Appendix A. The AAD between the predicted and experimental compositions of the SCN8 groups was under 12 percent. However, this value for the SCN12 groups was just over 18 percent. Also there was an over-prediction of the composition of the SCN7 group. The Average AAD for this group was as high as 23 %. This result indicated that the coefficient S = 15.5 for the SCN8 group is too high for Trinidad gas condensates.

Figure 7. Molecular Weight of Plus Fraction versus SCN Source: Ahmed et al. (1985)

Figure 8. Molecular Weight of Plus Fraction versus SCN Source: Ahmed et al. (1985)

4. A Proposed “Four Coefficient” Model (4CM) for

Splitting the C7+ Fraction. The observed discontinuities at SCN8 and SCN13 and change in slope at SCN13 (Hosein and McCain, 2009) as shown in Figure 5 and 6 suggest that Ahmed’s splitting scheme should be modified to four (4) segments and hence four (4) coefficients, instead of two as follows:

1) Segment 1from SCN7 to SCN8 due to the observed discontinuity at SCN8 and segment 2 from SCN8 to SCN12 (see Figures 9 and 10).

2) Segment 3 from SCN12 to SCN`13 due to the observed discontinuity at SCN13 and segment 4 from SCN13 and beyond (see Figures 11 and 12).

4.1 Coefficients for the New 4CM Model The “marching technique” described in Appendix A was applied to the 12 sets of compositional data given in Table A-1. The calculated average molecular weights of the plus fractions (Mn+) were plotted against SCN number (7, 8, 9…..) as shown in Figures 9 and 10 and Figures 11 and 12 for sample PL6, respectively (Hosein, 2004). The first two segments shown in Figures 9 and 10 provide the coefficients S for n = 8 and for 8 < n ≤ 12. The y intercept is the molecular weight of the C7+ fraction.

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 29: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

R. Hosein and R. Mayrhoo: A Model for Extending the Analysis of the Heptanes plus Fraction for Trinidad Gas Condensates 27

Figure 9. Molecular Weight of Plus Fraction versus SCN for Sample PL6 (Coefficient S = 11.0 for n = 8)

Figure 10. Molecular Weight of Plus Fraction versus SCN for Sample PL6 (Coefficient, S = 14.8 for 8 < n <13)

Figure 11. Molecular Weight of Plus Fraction versus SCN for Sample PL6 (Coefficient S = 10.4 for n = 13)

Figure 12. Molecular Weight of Plus Fraction versus SCN for Sample PL6 (Coefficient, S = 12.3 for n > 13)

Equation 3 is applied with the “marching technique” to compute the C7 composition with the coefficient S derived for n = 8 and compositions from C8 to C11 with the coefficient S derived for 8 < n ≤ 12. Predictions from these first two segment relationships (Figures 9 and 10) take care of the discontinuity at C8 (as shown in Table 1 and Figure 13 for sample PL6).

Table 1. Deviation between Predicted and Experimental SCN Composition for Sample PL6

SCN Group

Experimental Mole %

Predicted Mole % 4CM

Deviation %

7 0.307 0.306 -0.2 8 0.380 0.371 -2.3 9 0.205 0.209 2.0

10 0.157 0.158 0.8 11 0.113 0.121 6.8 12 0.071 0.071 -0.7 13 0.079 0.081 2.5 14 0.062 0.057 -7.3 15 0.053 0.049 -7.0 16 0.039 0.042 7.1 17 0.032 0.034 7.3 18 0.028 0.027 -2.3 19 0.021 0.021 -1.1

Figure 13. Predicted and Experimental SCN Composition for Sample PL6

M n + = M12 + + S (n-12), where n ≥ 12 …..(4) The second two segments shown in Figures 11 and

12, which has been mathematically expressed by Equation 4 from this study, provide the coefficients S for n = 13 and n >13. The y intercept is the molecular weight of the SCN12 plus group which is calculated from the “marching technique” (as outlined in Appendix A). Equation 4 is applied with the “marching technique” to compute the composition of SCN12 with the coefficient S derived for n = 13 and the compositions of SCN groups greater than SCN12 with the coefficient S derived for n >13.

Predictions from these second two segment relationships (see Figures 11 and 12) for SCN

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 30: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

R. Hosein and R. Mayrhoo: A Model for Extending the Analysis of the Heptanes plus Fraction for Trinidad Gas Condensates 28

compositions take care of the discontinuity at SCN13 and change in slope from SCN13 and beyond (as shown in Table 1 and Figure 13. The deviation between the predicted and experimental compositions for sample PL6 was less than ±8.0 percent as shown in Table 1.

The coefficients S, for each of the four (4) compositional segments described above were determined for the twelve (12) compositional data sets (see Table A-1 of Appendix A) and are shown in Table 2. The averages of these coefficients (shown in Table 3) were applied with the Four Coefficient Model (4CM) described above to split the C7+ fraction of the twelve Trinidad gas condensate samples in Table A-1. Table 2. Coefficients S for each Data Set from the Four Segment

Approach

Sample n = 8 S

8<n<13 S

n = 13 S

n > 13 S

PL1 PL2 PL3 PL4 PL5 PL6

10.7 12.2 11.4 10.3 10.9 11.0

15.1 16.1 15.4 14.2 15.6 14.8

10.1 11.6 11.1 10.3 10.6 10.4

11.5 13.6 13.3 12.0 12.1 12.3

CL1 CL2 CL3 CL4 CL5 CL6

12.1 11.4 11.8 12.2 12.7 10.2

16.6 15.7 15.8 16.6 17.0 13.9

11.4 12.1 11.5 10.8 11.5 10.3

13.3 13.8 13.5 12.5 13.4 11.8

Table 3. Coefficients for Predicting SCN Compositions, S by the

Four Coefficient Model (4CM)

SCN, n n = 8 8<n<13 n = 13 n > 13 Coefficient S = 11.5 S = 15.5 S = 11.0 S = 13.0

4.2 Comparison of Results The results obtained from the proposed 4CM model were compared with those obtained by the Ahmed et al. (1985) splitting scheme. Table 4 and Figure 14 show that this new 4CM model gives better prediction of SCN compositions than the Ahmed et al. (1985) two coefficient model. The AAD between predicted and experimental compositions of the SCN groups was less than 10 percent. It is also important to note that experimental compositional analysis beyond C7+ is not required when applying this 4CM model. 5. Conclusions For gas condensates there are discontinuities in the relationship between compositions of the SCN groups and molecular weights at SCN8 and SCN13. Existing splitting schemes such as the exponential and gamma distribution function were developed based on the assumption of a continuous relationship.

Table 4. Average Absolute Deviation between SCN Compositions Predicted by Ahmed et al. and the 4CM Splitting Models and

Measured SCN Compositions from Table A-1

SCN Group

No. of Data AAD, % Ahmed et al.

AAD, % 4CM

7 12 23.0 8.2 8 12 11.6 6.5 9 12 7.0 7.7

10 12 4.7 4.5 11 12 5.4 6.5 12 12 18.4 8.8 13 12 8.3 6.3 14 12 7.4 5.5 15 12 9.7 6.0 16 12 7.1 8.0 17 12 9.2 7.1 18 12 13.3 7.6 19 12 15.2 8.5

Figure 14. Average Absolute Deviation between Predicted and

Measured SCN Compositions from Table A-1

As a result these schemes under predict the SCN8 compositions and over-predict the SCN12 compositions by more than 25 percent. The use of the gamma probability distribution function does not recognise either of these discontinuities.

The method advocated by Ahmed et al. (1985) takes care of the discontinuity at SCN8 but not at SCN13. Therefore these splitting schemes require experimental extended analysis beyond SCN 14 for a more accurate description of the compositions of the SCN groups. The proposed new “Four Coefficient” model takes these two discontinuities into account and does not require a partial experimental analysis beyond C7+. This new “Four Coefficient” model can be used to predict the compositions of SCN groups for Trinidad gas condensates.

The four coefficients generated from this study can be applied to extend the C7+ fraction for gas condensate systems from any region. Improved predictions can be obtained for samples from a particular region or for multiple samples by generating four new coefficients by

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 31: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

R. Hosein and R. Mayrhoo: A Model for Extending the Analysis of the Heptanes plus Fraction for Trinidad Gas Condensates 29

applying the equations and method described for this “Four Coefficient” model. Acknowledgements The authors would like to thank the Campus Research and Publication Fund Committee of The University of the West Indies for providing the financial support for this Research Project. References: Ahmed, T. (1989), Hydrocarbon Phase Behaviour, Gulf

Publishing Company, Houston, Texas. Ahmed, T. H., Cady, G. V. and Story A. L. (1985), “A generalised

correlation for characterising the hydrocarbon heavy fractions”, SPE 14266, Society of Petroleum Engineers.

Ahmed, T. H., Cady, G. V., Story A. L., Verma, V., and Banerjee, S. (1984), “An accurate method for extending the analysis of C7+). SPE 12916, Society of Petroleum Engineers.

Al-Meshari A. A. and McCain W. Jr. (2007), “Validation of splitting the hydrocarbon plus fraction: First step in tuning equation of state”, SPE 104631, Society of Petroleum Engineers

Danesh, A. (1998), PVT and Phase Behaviour of Petroleum Reservoir Fluids, Elsevier, Amsterdam

Hosein, R. (2004), Phase Behaviour of Trinidad Gas Condensates, (Unpublished PhD Thesis), Faculty of Engineering, The University of the West Indies, St Augustine, Trinidad.

Hosein, R. and McCain W. Jr. (2009), “Extended analysis for gas condensate systems”, (SPE 110152-PA), Reservoir Evaluation and Engineering, Vol. 12, No. 1, pp.159-166.

Katz, D. L. and Firoozabadi, A. (1978), “Predicting phase behaviour of condensate/crude oil systems using methane interaction coefficients”, Journal of Petroleum Technology and Alternative Fuels, Vol.30, No.11, pp.1649-1655.

Katz, D. L., Hekim, Y. and Firoozabadi, A. (1978), “Reservoir depletion calculations for gas condensates using extended analyses in the Peng-Robinson Equation of State”, Canadian Journal of Chemical Engineering, Vol. 56, pp.1649-1655.

McCain, W. D. (1990), The Properties of Petroleum Fluids, Second Edition, Penn Well Publishing Co., Tulsa, Oklahoma

Pearson, K. (1895), “Contributions to the mathematical theory of evolution 11. Skew variations in homogeneous material”, Philosophical Transactions, Royal Society of London, Series A. 186, pp.343-414.

Pedersen, K. S., Thomassen, P., and Fredenslund, A.A. (1984), “Thermodynamics of petroleum mixtures containing heavy hydrocarbons 1. Phase envelope calculations by use of the Soave-Redlich-Kwong Equation of State”, Industrial & Engineering Chemistry Process Design and Development, Vol.23, pp.163-175.

Pedersen, K. S., Thomassen, P., and Fredenslund, A.A. (1985), “Thermodynamics of petroleum mixtures containing heavy hydrocarbons. 3. Efficient flash calculation procedures using the SRK Equation of State”, Industrial & Engineering Chemistry Process Design and Development, Vol.24, pp.948-954

Pedersen, K. S., Fredenslund, A.A. and Thomassen, P. (1989), Properties of Oil and Natural Gases, Gulf Publishing, Houston.

Whitson, C. H. (1983), “Characterising hydrocarbon plus fractions”, Journal of Petroleum Science and Engineering, Vol.23, No.4, pp.683-694.

Whitson, C. H. and Brule, M. R. (2000), Phase Behaviour, Monograph Series, SPE, Vol.20, pp.35-42.

Whitson, C. H. and Torp, S. B. (1983), “Evaluating constant volume depletion data”, Journal of Petroleum Technology pp.610-620.

Yarborough, L. (1978), “Application of a generalised equation of state to petroleum reservoir fluids”, Journal of the American Chemical Society, Vol.182, pp.386-439.

Appendix A: Splitting Scheme of Ahmed et al. (1985) – “Marching Technique” The “marching technique” devised by Ahmed et al. (1985) was based on observation that hydrocarbon systems exhibit a molar distribution that is relative to the average molecular weight and specific gravity of the plus-fraction. It involves calculating the composition zn at a progressively higher SCN fraction as follows:

n -1 z n = ( z7+ – ∑ zi ) [ ( MW(n + 1) + – MWn + ) / (MW(n + 1) + – MWn)] ×100 ..A-1 i = 7 where:

z n = mole percent of the extended SCN fraction MWn = molecular weight of the SCN fraction as outlined by Katz

and Firoozabadi (1978). MWn + = molecular weight of the n plus fraction (C8+, C9+……), as

calculated by the following expression:

MW n + = MW 7 + + S (n-7) …..A-2 Where the subscript n is the number of the SCN fraction and

the coefficient S for gas condensate systems = 15.5 for n = 8 and 17.0 for n > 8.

MW (n + 1) + in Equation A-1 is the molecular weight of the next plus fraction and is also calculated from Equation A-2, but with the next n value, i.e. n = n + 1.

Procedures for obtaining molecular weight and specific gravity of the n+ fractions, i.e., MW n + and SG n + were summarised by Ahmed et al. as follows:

1) Given the composition; specific gravity; and molecular weight for a hydrocarbon system, up to C7+, all components heavier than C7 are grouped into a “plus” fraction, i.e. C 8+, which is characterised by an average molecular weight MW 8 +, a specific gravity SG 8 + and a total mole percent z 8+.

2) The average molecular weight and specific gravity of C8+ are than calculated from the following relationship:

MW8+ = [ z7+ MW7 + – z7 MW7] / z8 + ….. A-4

and SG8+ = z8+ MW8+ / [ (z7+ MW7+ / SG7 +) - (z7 MW7 / SG7)] ….. A-5

where MW7+, SG7 + = measured molecular weight and specific

gravity of heptanes plus respectively MW8+, SG8 + = calculated molecular weight and specific

gravity of octanes plus respectively MW7, SG7 = average molecular weight and average specific

gravity of heptanes as recommended by Katz-Firoozabadi (1978)

3) The physical properties of the next hydrocarbon “plus” fraction i.e. C9+ are calculated following the procedures outlined in steps 1 and 2, thus,

MW9+ = [ z8+ MW8 + – z8 MW8] / z9 + ….. (A-6)

and SG9+ = z9 + .MW9+ / [ (z8+ MW8+ / SG8 +) - (z8 MW8 / SG8)] ….. (A-7)

This “marching technique” for calculating average molecular weight and specific gravity is repeated until the last component in the hydrocarbon system is reached.

The Equations of Average Absolute Deviation (AAD) and Deviation (Dev) are expressed below: n

Average Absolute Deviation, AAD = 1 × ∑ Zcalc. – Zexpt. n i=1 Zexpt.

Deviation, Dev. = Zcalc. – Zexpt. Zexpt.

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 32: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

R. Hosein and R. Mayrhoo: A Model for Extending the Analysis of the Heptanes plus Fraction for Trinidad Gas Condensates 30

Table A-1. Compositional Data and Properties of C7+ for the 12 Trinidad Gas Condensate Samples Used in This Study SCN

Group PL1

Mole % PL2

Mole % PL3

Mole % PL4

Mole % PL5

Mole % PL6

Mole % CL1

Mole % CL2

Mole % CL3

Mole % CL4

Mole % CL5

Mole % CL6

Mole % C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 C18 C19

C20+

0.561 0.789 0.491 0.354 0.267 0.176 0.197 0.170 0.144 0.119 0.104 0.099 0.080 0.373

0.458 0.583 0.319 0.245 0.196 0.127 0.148 0.105 0.093 0.074 0.066 0.054 0.045 0.240

0.508 0.652 0.353 0.286 0.194 0.134 0.146 0.121 0.103 0.077 0.065 0.052 0.041 0.187

0.346 0.444 0.308 0.231 0.165 0.108 0.122 0.095 0.084 0.068 0.059 0.049 0.040 0.149

0.368 0.471 0.297 0.208 0.141 0.089 0.101 0.079 0.069 0.052 0.045 0.037 0.031 0.134

0.307 0.380 0.205 0.157 0.113 0.071 0.079 0.062 0.053 0.039 0.032 0.028 0.021 0.069

0.588 0.729 0.491 0.327 0.246 0.156 0.169 0.151 0.121 0.098 0.085 0.076 0.064 0.344

0.353 0.457 0.304 0.222 0.162 0.118 0.129 0.101 0.088 0.068 0.059 0.052 0.044 0.232

0.390 0.486 0.322 0.225 0.171 0.115 0.123 0.106 0.093 0.069 0.059 0.054 0.043 0.222

0.523 0.633 0.352 0.238 0.196 0.110 0.123 0.099 0.086 0.066 0.055 0.047 0.039 0.184

0.414 0.504 0.310 0.215 0.162 0.100 0.112 0.090 0.079 0.064 0.056 0.047 0.039 0.218

0.294 0.367 0.234 0.159 0.112 0.070 0.080 0.056 0.046 0.039 0.032 0.027 0.019 0.057

C7+ SG7+

MW7+

3.294 0.8031

160

2.753 0.8004

157

2.919 0.7939

150

2.268 0.7967

153

2.122 0.7918

148

1.616 0.7869

143

3.645 0.8021

159

2.389 0.8049

162

2.478 0.8022

159

2.751 0.7925

148

2.410 0.8005

157

1.592 0.7849

141

Table A-2. Average Absolute Deviation between SCN Compositions Predicted by Ahmed et al. and the 4CM Splitting Models and Measured SCN Compositions for the 12 Samples in

Table A-1 SCN

Group No. of Data

AAD, % EXPON.

AAD, % GAMMA

AAD, % Ahmed et al.

7 12 4.5 8.7 23.0 8 12 26.2 26.8 11.6 9 12 8.7 10.6 7.0

10 12 4.2 6.4 4.7 11 12 14.1 14.2 5.4 12 12 35.1 31.6 18.4 13 12 7.8 11.3 8.3 14 12 7.6 13.6 7.4 15 12 3.2 10.9 9.7 16 12 3.1 19.0 7.1 17 12 4.9 15.5 9.2 18 12 9.9 9.5 13.3 19 12 8.2 8.1 15.2

Authors’ Biographical Notes:

Raffie Hosein is Senior Lecturer in the Petroleum Studies Unit at The University of the West Indies (UWI) in Trinidad and Tobago. Previously he worked as a Petroleum Engineer with the Ministry of Energy in Trinidad and late, as a Senior Associate Professor in the Department of Petroleum Engineering at Texas A&M University at Qatar. He received his B.Sc., M.Phil and Ph.D degrees in Petroleum Engineering from The University of the West Indies, Trinidad and Tobago.

Figure A-1 Average Absolute Deviation among Predicted and Measured SCN Compositions for the 12 Samples in Table A-1

Rayadh Mayrhoo is a Chemical and Process Engineer Trainee with Atlantic LNG. Previously, he worked as an Associate Professional with the Ministry of Planning and Sustainable Development in Trinidad. He received his BSc in Chemical and Process Engineering and MSc in Petroleum Engineering from The University of the West Indies, Trinidad and Tobago. ■

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 33: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

E.I. Ekwue et al. Simulation of Irrigation Water Requirements of Some Crops in Trinidad Using the CROPWAT Irrigation Software 31

Simulation of Irrigation Water Requirements of Some Crops in Trinidad Using the CROPWAT Irrigation Software

Edwin I. Ekwue a,Ψ, Rebekah, C. Constantineb and Robert Birch c

Department of Mechanical and Manufacturing Engineering, Faculty of Engineering, The University of The West Indies,

St Augustine, Trinidad and Tobago, West Indies aE-mail: [email protected]

bE-mail: [email protected] cE-mail: [email protected]

Ψ - Corresponding Author (Received 13 June 2014; Revised 07 October 2014; Accepted 18 October 2015)

Abstract: The Crop Water Requirements (CROPWAT) computer software package was used to design irrigation schedules during the dry season (February to May) for twelve (12) major farming locations in Trinidad. The irrigation schedules are for the nine major crops grown in different predominant soils in the selected locations. Crop and field parameters were obtained from published texts whereas the climatological data were obtained from the Water Resources Agency in Trinidad. The irrigation schedules using CROPWAT were planned in such a way that for the convenience of the farmer, the irrigation depth and irrigation interval were kept constant throughout the growing season for each crop and this value depended on the climatological situation or the water consumption pattern of the crops.

Keywords: Irrigation, Scheduling, Crop, Soil, Trinidad

1. Introduction Scientific irrigation scheduling is the best management practice needed for improving farm irrigation management (Simon et al., 1998). One of the objectives of irrigation management is to plan irrigation schedules, that is, to determine the correct water depth and interval of future irrigations in a locality so as to achieve optimal crop productivity (Evans et al., 1996). If water deliveries are untimely or not in desired volume, irrigation efficiency decreases. In Trinidad, the process of determining the artificial amounts of water needed by crops is seen as unimportant. The interaction of the authors with the farmers showed that they were either over-irrigating or under-irrigation their crops. Limited supply will lead to a reduction in yield whereas excess water results in percolation losses, leaching of nutrients and may cause temporary water-logging (Ekwue and Rigg, 2001).

Optimal irrigation practice requires accurate planning of the irrigation schedule. The challenge in irrigation scheduling is to achieve optimum yield with minimum water losses. Irrigation scheduling is normally achieved by observing the plants, keeping a water balance sheet, measuring the soil moisture content and using computer software (Evans et al., 1996; Goldsmith et al., 1988). The common computer software used in irrigation scheduling includes California Irrigation Management Information System, CIMIS (Allen et al., 2010), Crop Irrigation Water Requirements, CRIWAR

(Lenselink and Jurriens, 1993), Irrigation Scheduling Information System, IRSIS (Raes et al., 1988a) and the Crop Water Requirements, CROPWAT (Smith, 1992) of the Food and Agriculture Organisation of the United Nations. Goldsmith et al. (1988) also developed a computer spreadsheet model for scheduling irrigation.

Information on the required depths and time of carrying out irrigation practice in different soils for different crops in many areas in Trinidad and the Caribbean region as a whole, is not readily available. Smith (1959) computed the irrigation needs of many Islands of the Caribbean region using the rainfall and evapotranspiration data. He did not actually produce irrigation schedules in these locations which require information about the weather, crop and field data. Ekwue and Rigg (2001) used IRSIS to schedule and evaluate irrigation patterns in the St. Mary Banana Estates in Jamaica. There is the need to extend this evaluation to other parts of the Caribbean region and beyond.

This paper reports the result of a study undertaken to collect, collate weather, crop and field data from 12 major farming locations in Trinidad in the Caribbean region in order to produce optimal irrigation schedules which will be used for the major crops grown in different soils. 2. Materials and Methods A computer programme for irrigation planning and

ISSN 0511-5728 The West Indian Journal of Engineering

Vol.37, No.2, January 2015, pp.31-36

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 34: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

E.I. Ekwue et al. Simulation of Irrigation Water Requirements of Some Crops in Trinidad Using the CROPWAT Irrigation Software 32

management, CROPWAT, was developed to plan irrigation schedules at field level (Smith, 1992). The version of CROPWAT used, CROPWAT 8.0 for Windows (FAO, 2013) calculates crop water requirements and irrigation requirements based on soil, climate and crop water using the procedures of Allen et al. (1998) and Dorrenbos and Kassam (1979).

For a given climate, crop and field, it offers the possibilities of computing the net irrigation requirements and the optimal water distribution resulting in the highest yield under conditions of limited water. CROPWAT shows the consequences of the irrigation schedule in terms of water application efficiency and expected yield response. The main strength of the software is its ability to ease the calculation of alternative irrigation schedules, their consequences and the effect on the different entities of the soil water balance. The software was used to plan irrigation schedules of twelve (12) irrigation areas in Trinidad (see Table 1). 3. Simulation of Irrigation Schedules To run CROPWAT, the following data has to be specified: a. The climatological data which consist of the reference

crop evapotranspiration (ET0) and the rainfall data. For irrigation scheduling, Allen et al. (1998) suggested the use of monthly climatological normals of ETo and monthly levels of rainfall that will be exceeded in 4 out of 5 years (80% dependable rainfall), termed effective rainfall. For this study, ETo (see Table 2) was

computed by CROPWAT using the Penman-Monteith method. The Penman-Monteith method is the one recommended by the International Commission for Irrigation and Drainage (ICID) and the Food and Agriculture Organisation (FAO) of the United Nations. Its use has also been found promising within the Caribbean region (Simon et al., 1998). The weather data (air temperature, wind speed, humidity and radiation) used for the ETo computation as well as mean monthly rainfall data were collected for the four weather stations from the Water Resources Agency in Trinidad. The four (4) weather stations are in close proximity to the 12 irrigation areas (see Table 1). The effective rainfall values (see Table 2) were computed from the rainfall values using empirical linear equation derived by Ekwue et al. (1997) for Trinidad.

b. The crop data consist of information about the length of crop development stages of the crop, the critical crop depletion factor (p-factor), the crop coefficient (Kc) and the crop yield response factor to water (Ky). The p-factor expresses the fraction of the soil water that can be extracted at the potential rate by the roots. It is given as the ratio between the readily available and the total available soil water. The data for the nine major crops used in this study (see Table 3) were extracted from the data presented by Allen et al. (1998) and Doorenbos and Kassam (1979). Planting date for the crops was put as February 1st, which is during the dry season when supplementary irrigation is needed in Trinidad.

Table 1. Weather stations and the irrigation areas

Weather Stations Elevation (m) Longitude (W) Latitude (N) Irrigation areas covered Hollis Reservoir 150 61o 11’ 10o 41’ Plum Mitan; Fishing Pond/Oroupouche UWI Field Station 15 61o 23’ 10o 58’ Aranguez; Macoya; Bamboo Settlement;

Orange Groove Navet Reservoir 122 61o 11’ 10o 24’ Rio Claro; Tabaquite; Biche/Cuche Piarco Met Office 11 61o 21’ 10o 35’ Cunupia; Felicity; Maloney/Arouca

Table 2. Reference crop evapotranspiration, ETo (mm/day); rainfall (mm) and effective rainfall (mm) of the various months for four weather stations near to the irrigation areas

Weather stations

January Feb. March April May June July August Sept. Oct. Nov Dec.

Hollis Reservoir

*3.61 **191

***131

3.87 132 84

4.71 107 64

4.20 134 85

4.45 195 134

3.97 334 245

4.38 349 257

4.68 346 255

4.37 218 152

3.95 285 206

3.50 380 282

3.44 355 262

UWI Field Station

3.79 85 46

4.38 51 19

4.79 32 4

5.03 51 19

4.87 93 52

4.34 160 106

4.15 210 146

4.14 237 168

4.28 161 107

4.03 183 124

3.72 203 140

3.57 134 85

Navet Reservoir

3.61 154 101

3.79 82 44

4.36 41 11

4.27 48 17

3.75 113 69

3.30 246 175

3.69 269 193

4.01 198 137

3.99 130 82

3.72 145 94

2.97 207 144

3.31 216 157

Piarco Met Office

4.27 81 43

4.88 60 26

5.33 35 6

5.60 53 21

5.27 125 78

4.50 200 138

4.41 217 152

4.47 231 162

4.56 142 92

4.23 162 107

3.90 197 136

3.77 148 97

Notes: *Reference crop evapotranspiration, ETo **Rainfall, R ***Effective Rainfall = -22.1 + 0.75 R. Values are averages for the periods 2001 to 2012.

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 35: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

E.I. Ekwue et al. Simulation of Irrigation Water Requirements of Some Crops in Trinidad Using the CROPWAT Irrigation Software 33

c. The crop data consist of information about the length of crop development stages of the crop, the critical crop depletion factor (p-factor), the crop coefficient (Kc) and the crop yield response factor to water (Ky). The p-factor expresses the fraction of the soil water that can be extracted at the potential rate by the roots. It is given as the ratio between the readily available and the total available soil water. The data for the nine major crops used in this study (see Table 3) were extracted from the data presented by Allen et al.

(1998) and Doorenbos and Kassam (1979). Planting date for the crops was put as February 1st, which is during the dry season when supplementary irrigation is needed in Trinidad.

d. The field data comprise the soil water content at field capacity and at the permanent wilting point together with the basic infiltration rate. The field data (see Table 4) for the 10 typical soil textures found in the different locations used in this study were taken from data presented by Brown and Bally (1976).

Table 3. Crop data for the nine major crops grown in Trinidad Crops Critical crop

depletion fraction, p

Rooting depths (cm)

Crop growth periods (days) Initial Crop

development Mid-season Late-season

Water melons 0.40 130 Days: 25 *Kc: 0.50

**Ky:0.50

35 -

0.60

40 1.05 1.10

20 0.75 0.80

Pumpkin

0.35 130 Days: 20 Kc: 0.60 Ky: 0.20

30 -

0.40

30 1.00 0.45

20 0.80 0.60

Egg plant 0.40 110 Days: 30 Kc: 0.60 Ky: 1.40

45 -

0.60

40 1.05 1.20

25 0.90 0.60

Cucumbers 0.50 100 Days: 20 Kc: 0.60 Ky: 0.20

30 -

0.40

40 1.05 0.45

15 0.90 0.60

Tomatoes 0.40

100 Days: 30 Kc: 0.60 Ky: 0.50

40 -

0.60

45 1.15 1.10

30 0.80 0.80

Sweet peppers 0.30 80 Days: 30 Kc: 0.60 Ky: 1.40

35 -

0.60

40 1.05 1.20

20 0.90 0.60

Cabbage 0.45 50 Days: 40 Kc: 0.70 Ky: 0.20

60 -

0.40

50 1.05 0.45

15 0.95 0.60

Cauliflower 0.45 50 Days: 35 Kc: 0.70 Ky: 0.20

50 -

0.40

40 1.05 0.45

15 0.95 0.60

Lettuce 0.30 40 Days: 20 Kc: 0.70 Ky: 0.20

30 -

0.40

15 1.00 0.45

10 0.95 0.60

Notes: * Kc is crop factor; ** Ky is crop yield response factor

Table 4. Field data for the soil types in different irrigation areas Weather stations

Irrigation areas Soil types Field capacity, % by mass

Permanent wilting point, % by mass

Bulk density (g/cm3)

Infiltration rates (mm/d)

Hollis Reservoir

Fishing Pond/Oroupouche

Sangre Grande Silty clay

36 24 1.32 264

Plum Mitan Navet clay 33 24 1.20 228 UWI Field Station

Orange Groove Sandy clay loam 29

13

1.28

547 Macoya

Bamboo Settlement Aranguez Clay 33 14 1.25 204

Navet Reservoir

Rio Claro Ecclesvile clay loam

36 18 1.14 298

Biche/Cuche Navet clay 33 24 1.20 228 Tabaquite Brasso clay 35 17 1.15 228

Piarco Met Office

Cunupia Cunupia silty clay 36 24 1.32 254 Felicity Felicity clay loam 30 22 1.38 298 Maloney/Arouca Clay loam 22 10 1.38 298

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 36: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

E.I. Ekwue et al. Simulation of Irrigation Water Requirements of Some Crops in Trinidad Using the CROPWAT Irrigation Software 34

With the weather, crop and soil data specified for the different locations, the irrigation schedules were determined for the 80% dependable rainfall. In order to obtain an easy and manageable schedule for farmers, the time criterion in the CROPWAT software was set to “fixed interval” while the depth criterion was set at “fixed depth” for periods of the irrigation season. The following approach was used. First a constant irrigation depth was selected based on the soil type and the rooting depth of the crop. In all the cases, it was found that the interval between irrigations could be kept constant for the whole irrigation season, without inducing deep percolation or crop stress.

As an example, for the scheduling of irrigation of tomato grown on clay soil in Aranguez (see Table 5), after entering the constant depth and interval of irrigation, the irrigation simulation was carried out. The diagram obtained for the root zone depletion of water for the crop was examined. Irrigation is ideal when the water depletion levels fall between the readily available water and the field capacity as shown in Figure 1. The table of expected yield response within the software was checked to ensure that at the different growth stages of the crop, the yield response was 100% or very close to it. The table on irrigation efficiency was also checked to

make sure that irrigation efficiencies were very close to 100%. Depending on the results of the simulations, the schedule could be adapted and tested again until ideal root zone depletion pattern similar to that in Figure 1 was obtained. 4. Discussion of Simulation Results The irrigation schedules designed for the crops grown in different soils in the twelve locations in Trinidad are summarised in Table 5. For all schedules derived, the irrigation depth was kept constant throughout the growing season and the irrigation interval was also kept constant. The depth of water chosen for simulation was a function of the root zone depth of the crop as well as the water storage capacity of the soil.

The crops with deep root systems like the tomatoes with maximum root depth of 100 cm (see Table 3) could be applied larger depths of water (up to 17 mm of water, see Table 5) without causing water-logging, especially if they are grown in clay soils which have large water storage capacity. This was shown by the larger difference between the values of field capacity and the permanent wilting point (see Table 4). The irrigation interval in such cases is large (i.e., 6 days), meaning that irrigation would be applied less often.

Table 5. Irrigation scheduling parameters for the different irrigation areas Weather stations Irrigation areas Crops Irrigation water

requirements (mm) Irrigation scheduling parameters

Depth (mm) Interval (days) UWI Field

Station Aranguez clay Tomato 422 17 6

Sweet pepper 380 10 3 Cabbage 390 9 3

Macoya sandy clay loam Egg plant 380 17 5 Tomato 422 17 5 Cabbage 390 9 3

Orange Grove sandy clay loam

Pumpkin 326 17 5 Tomato 422 17 5

Egg plant 381 17 5 Bamboo Settlement sandy

clay loam Sweet pepper 380 10 3

Cabbage 390 9 3 Hollis Reservoir Plum Mitan clay Tomato 111 6 6

Pumpkin 94 9 6 Water melon 92 9 6

Fishing Pond/North Oroupouche silty clay

Tomatoes 111 9 6 Pumpkin 94 9 6

Navet Reservoir Rio Claro clay loam Water melon 255 14 6 Pumpkin 250 14 6

Cucumber 259 14 6 Biche/Cuche clay Water melon 255 15 6

Cucumber 259 16 6 Pumpkin 250 18 6

Tabaquite clay Cauliflower 272 8 3 Pumpkins 250 15 6

Piarco Met Office Cunupia silty clay Sweet pepper 419 11 3 Pumpkins 375 20 5

Felicity clay loam Tomatoes 460 20 5 Cucumber 389 17 4

Maloney/Arouca clay loam Cauliflower 429 10 3 Lettuce 393 9 3

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 37: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

E.I. Ekwue et al. Simulation of Irrigation Water Requirements of Some Crops in Trinidad Using the CROPWAT Irrigation Software 35

Notes: RAM: Readily available moisture; TAM: Total available moisture Figure 1. Simulation results for irrigation of tomatoes grown in Aranguez clay soil

On the other hand, shallow rooted crops like

cabbage with maximum root depth of 0.5 m (see Table 3) could only be applied 9 mm of irrigation water at a time if grown in the same clay soil (see Tables 5). Irrigation in this case is more often (i.e., 3 days). 5. Conclusions It has been shown that the CROPWAT software package is an easy tool to generate the optimal irrigation pattern of crops in Trinidad. Another similar software, Irrigation Scheduling Information Systems (IRSIS), has been used to schedule irrigation projects in France, Brazil and Jamaica by Raes et al. (1988b), De Goes et al. (1992) and Ekwue and Rigg (2011), respectively. Goldsmith et al. (1988) also described the use of a computer spreadsheet model to produce irrigation schedules for smallholder farmers in Zimbabwe and Thailand. The main strength of the CROPWAT software is that each of the designed schedules could be evaluated in terms of crop response and deep percolation losses. The criterion used for the planning of the irrigation schedules is a constant irrigation depth, since farmers are more apt in managing a constant depth rather than a variable depth (Sagardoy, 1982). The irrigation interval was also kept constant since the variations in the weather and water consumption pattern of the crops were not enough to make the need to vary if necessary.

Future studies should test the efficiency of these optimally derived irrigation schedules in the field. Meanwhile these produced schedules could be utilised and may be further refined for specific situations of the “dry year” or “wet year” as specified by Ekwue and Rigg (2001). Smith (1986) and Goldsmith et al. (1988) stated that computer based simulations similar to the

ones produced in this paper using CROPWAT could be utilised in the training of irrigation and soil management professionals. Hypothetical simulations of an irrigation system could be used to show managers in the Caribbean and other regions, the consequences of their decisions and how different groups of crops, water and weather situations interact with one another to affect crop yields and soil conditions (Goldsmith et al., 1988). References: Allen, L, Christian-South, J., Cohen, M.J., Schulte, P., and Smith,

C. (2010, California Farm Water Success Stories: Final Report, Pacific Institute, California

Allen, R.G., Pereira, L.S., Raes, D. and Smith, M. (1998), “Crop evapotranspiration - Guidelines for computing crop water requirements”, FAO Irrigation and Drainage Paper, No.56, Rome, Italy.

Brown, C. and Bally, G. (1976), Land Capability and Survey of Trinidad and Tobago: No. 4 - Soils of the Northern Range of Trinidad, Government Printery, Port of Spain, Trinidad

De Goes, C.M., Gaal Vadas, R., Calasans Rego, N. and Raes, D. (1992), “Computer support systems for irrigation scheduling - Case study: Pirapora Project, Brazil”, ICID Bulletin, Vol.41, No.2, pp.19-26.

Doorenbos, J. and Kassam, A.H. (1979), “Yield response to water”, FAO Irrigation and Drainage Paper, No.33, Rome, Italy

Ekwue, E.I. and Rigg, R.V. (2001), “Computer-aided irrigation scheduling - a case study: Design of irrigation schedules for St. Mary Banana Estates, Jamaica”, West Indian Journal of Engineering, Vol.23, No.2, pp.1-8.

Ekwue, E.I., Stone, R.J., and Smith, R. (1997), “Statistical analysis of Caribbean rainfall data: Formulating linear models relating dependable rainfall to mean monthly rainfall”, West Indian Journal of Engineering, Vol.19, No.2, pp.49-58.

Evans, R., Sneed, R.E. and Cassel, D.K. (1996), “Irrigation scheduling to improve water and energy-use efficiencies”, North Carolina Co-operative Extension Service, Published Number

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 38: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

E.I. Ekwue et al. Simulation of Irrigation Water Requirements of Some Crops in Trinidad Using the CROPWAT Irrigation Software 36

AG 452-4, available at: www.bae.ncsu.edu/programs/extension/evans/ag452-4.html (Accessed June 4, 2014).

FAO (2013), CROPWAT 8 for Windows, Food and Agriculture Organisation, available at: http://www.fao.org/nr/water/infores_databases_cropwat.html (Accessed May 27, 2014).

Goldsmith, H., Bird, J.D. and Howarth, S.E. (1988), “Computerised irrigation scheduling using spreadsheet models”, Irrigation and Drainage Systems, Vol.2, pp.211-227.

Lenselink, K.J. and Jurriens, M. (1993), An inventory of Irrigation Software for Microcomputers, International Institute for Land Reclamation and Improvement, Wageningen, The Netherlands.

Sagardoy, J.A. (1982), “Organisation, operation and maintenance of irrigation schemes”, FAO Irrigation and Drainage Paper, No.40, Rome, Italy.

Simon, C.M., Ekwue, E.I., Gumbs, F.A. and Narayan, C.V. (1998), “Evapotranspiration and crop coefficients of irrigated maize in Trinidad”, Tropical Agriculture (Trinidad), Vol.75, No.3, pp.342-347.

Smith, G.W. (1959), Irrigation Needs of Antigua, W.I., Imperial College of Tropical Agriculture, Soils Research and Survey Department, Regional Research Centre, Trinidad.

Smith, L.E.D. (1986), The role of simulation exercises in training irrigation managers, In Sheng, T.S. and Smith, L.E.D: Computers in irrigation management, ODI/IIMI Irrigation Management Network, 86/2c, ODI, London (ISSN 0260-8596).

Smith, M. (1992), “CROPWAT: A computer program for irrigation planning and management”, FAO Irrigation and Drainage Paper, No. 46, Rome, Italy.

Raes, D., Lemmens, H.; van Aelst, P., Bulcke, M.V. and Smith, M. (1988a), IRSIS: Irrigation Scheduling Information Systems Manual, Katholieke Universiteit Leuven, Belgium

Raes, D., Gullentops, D., Vanden Bulcke, M. and Feyen, J. (1988b), “Planning irrigation schedules by means of the IRSIS software package - A Case study: Chateau Porcine, France”. ICID Proceedings, Yugoslavia, Vol.4, pp.263 - 271.

Authors’ Biographical Notes

Edwin I. Ekwue is the Head of the Department of Mechanical and Manufacturing Engineering and Professor in charge of the Biosystems Engineering program at The University of the West Indies, St Augustine, Trinidad and Tobago. He is a member of the Editorial Board of the West Indian Journal of Engineering. His specialty is in Water Resources, Hydrology, Soil and Water Conservation and Irrigation. His subsidiary areas of specialisation are Structures and Environment, Solid and Soil Mechanics, where he has teaching capabilities. Professor Ekwue has published widely. He had served as the Deputy Dean (Undergraduate Student Affairs), the Deputy Dean (Post-graduate Affairs and Outreach), the Chairman of Continuing Education Committee, and the Manager of the Engineering Institute in the Faculty of Engineering at UWI. Rebekah, C. Constantine is a graduate of the Faculty of Engineering, The University of the West Indies in 2014 majoring in Mechanical Engineering with Biosystems. She has completed an Internship as a Mechanical Engineer Trainee at Nu-Iron Unlimited, Point Lisas Industrial Estate, Point Lisas. Her major interests are reliability and maintenance of mechanical devices. Robert Birch is an Instructor in the Department of Mechanical and Manufacturing Engineering at The University of the West Indies, St Augustine, Trinidad and Tobago. He is a registered Professional Engineer (R.Eng) and Project Management Professional (PMP) with over sixteen years of industrial and teaching experience. He has a BSc. (Eng) and MPhil in Agricultural Engineering from The University of the West Indies and is presently pursuing a PhD in Mechanical Engineering. Mr. Birch is a member of the Institution of Agricultural Engineers (UK). His interests are in Field Machinery and Heavy Equipment Design, Fluid Power Technology and Soil-Machine interaction. ■

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 39: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

S.O. Nkakini and A. Husseni. Development and Evaluation of Wheeled Long-Handle Weeder 37

Development and Evaluation of Wheeled Long-Handle Weeder

Silas O. Nkakini a,Ψ, and Abu Husseni b

Department of Agricultural and Environmental Engineering, Faculty of Engineering, Rivers State University of Science and

Technology, Nigeria aE-mail: [email protected]

bE-mail: [email protected] Ψ - Corresponding Author

(Received 30 April 2014; Revised 24 November 2014; Accepted 30 January 2015) Abstract: A push-type operated wheel weeder with an adjustable long handle, was designed, constructed and tested. The hoe performance from the tests on a field of Okra plant having an inter-row spacing of 800mm, showed that it could weed satisfactorily, and eliminate the drudgeries associated with the use of the short handle hoe such as backache, pains at the spine and lower waist region. Field capacity and efficiency of 0.050ha/hr and 87.5% were obtained respectively. Furthermore, the average weeding index and performance index obtained were 86.5% and 1108.48, respectively. At a speed of 0.04m/s, a high efficiency of 91.7% at 0.4m depth of cut was obtained. The developed wheeled long- handle weeder was found efficient.

Keywords: Push-type, adjustable long handle weeder, weeding index, performance index

1. Introduction Weeding is an important but equally labour intensive agricultural unit operation. Weeding accounts for about 25% of the total labour requirement ranging from 900-1200 man hour/hectare during cultivation season (Nag and Dutta, 1979). Its delay and negligence reduces crop yield from 30 to 60% (Singh, 1988). Weed control has become a highly specialised activity employing thousands of people especially in developing countries. This activity involves industries providing the necessary chemicals (herbicide), and individuals engaging in the practices of weed control.

Many weeding implements have been developed, amongst which are the traditional hoes, spades and the cutlasses. Their effectiveness is still very low with high energy demand. The average energy demand of the traditional tillage hoe ranges from 7 to 9.5kJ/min when compared with 4.5 kJ/min (75 watts) which is optimum limit of continuous energy output of man (Nwuba, 1981).

2. Literature Review Kamal and Babatunde (1999) developed a push type Oscillatory power weeder with the following operational parameters; weeding efficiency, field capacity, depth and width of cut, amplitude and frequency of vibration. Field capacity was found to be 0.036 ha/hr and efficiency 81.34%. Jadhav and Turbatmath (1991) developed a bullock drawn multipurpose hoe. This was designed to suit any row spacing between 300mm and 450mm for inter- cultural operation. The actual field capacity of a

pair of the hoe varied from 0.15-0.25ha/hr. A ridge profile weeder was developed by Odigboh

and Ahmed (1980). The weeder consists of two bicycle wheels welded together to a common hub, with the bicycles spokes replaced with 6mm diameter mild steel rods, rear and front sprockets, roller chains, shaft, two gangs of rotary hoe weeders and handle.

Yadav and Pund (2007) developed a wheeled hoe based on ergonomic factors. The performance of the developed weeder was evaluated in the field of groundnut crop. The field capacity of the weeder was found to be 0.048ha/hr. It was observed that the cutting width was proportional to the field capacity. Further evaluation revealed that, the weeding efficiency was 92.5%. The performance index was found to be 2611.7. It was observed that the developed weeder was not only suitable for groundnut crop but could also be used for other crops by adjusting the row spacing.

The use of the common short handle weeding hoe involves the application of much force with little output, the operator experiences backache, pains in the spine and lower waist, as a result of the working posture (Singh, 1992). Singh (1992) developed a wheeled hoe weeder with ergonomic considerations to improve its design and for commercialisation through small scale manufacturers. It required 60-110 man-hr/ha for weeding in black heavy soil and 25 man-hr/ha in light soil.

The usage of the wheel long handle weeding hoe has not been widely accepted by majority of the local farmers. In some countries e.g. Nigeria, most people often seen standing while working in the farm are

ISSN 0511-5728 The West Indian Journal of Engineering

Vol.37, No.2, January 2015, pp.37-44

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 40: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

S.O. Nkakini and A. Husseni. Development and Evaluation of Wheeled Long-Handle Weeder 38

regarded as been lazy. As a result of this misconception, the use of the wheeled long-handle weeding hoe has not been fully accepted.

Rangasamy et al. (1993) and Kamal and Babatunde (1999) reported that when a manually operated weeder, having a field capacity greater than the traditional hoe, was tested on a small plot of land with crops planted in rows, the machine had a low output of about 0.02ha/hr because of a lot of rigour involved in its usage. Farmers also experience a lot of rigour when using short handle hoes such as backaches or strains and spine problems at old age. Improving effectiveness in agricultural mechanization and ease in drudgeries associated with the use of short handle hoe, has initiated the design of wheeled long-handle hoe.

The designing and development of a wheeled long handle weeding hoe (WLHWH) was conceived by the desire to eliminate the drudgeries and remedy the difficulties associated with short handle hoe and save the peasant farmers the stress of bending while working, for effectiveness in agricultural mechanization. Thus, the objectives of this study are to design and fabricate a wheeled long-handle weeding hoe, test and analyse its performance on field capacity and efficiency, and compare its performance with that of short handle weeding hoe.

3. Material and Methods The weeder was designed based on the principle of weed stem failure due to soil shearing, impact and abrasion. The material selection was considered in terms of cost, availability, durability, overall weight and affordability. The design parameters considered were the ease of operation, average walking speed of the operator (0.8m/s), energy requirement of the weeder, and types of weeds to be operated upon. The material used for the shaft was mild steel. The shaft was designed based on strength, rigidity and stiffness. The shear stress, bending moment and deflections were also considered.

The push-type long handle weeder is shown in isometric view, plan view and rear view as in Figures 1, 2 and 3, respectively.

Figure 1. Isometric view of the weeder

Figure 2. Plan view of the weeder

Figure 3. Rear view of the weeder.

The main components of the machine and their functions are as follows:

1) The Handle - This was constructed with two galvanized pipes of lengths 900mm and 471mm respectively, making a total length of 1371mm. The galvanized pipes were welded across the mainframe handle to form the hand grip which has a length of 140mm. The handle enables the operator to push or pull and direct the machine during operation within the crop rows. It also enables the operator to raise the cutting blade a little bit high, should stone and stumps be encountered during operation. The handle is made adjustable to create comfort to the operator irrespective of the operator’s height. The essence of the long handle is to enable an upright posture while on weeding operation.

2) Weeding blade - The weeding blade was made from 51mm × 210mm mild steel having a thickness of

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 41: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

S.O. Nkakini and A. Husseni. Development and Evaluation of Wheeled Long-Handle Weeder 39

14mm. The blade at the lower end was sharpened and slanted to an angle of 15° to the horizontal. It is attached to a headpiece by means of bolt for easy replacement due to wear and tear. The blade has a maximum cutting depth of 0.6m with design width of cut of 0.2m.

3) Ground wheel - The ground wheel has a diameter of 300mm and a hub of 25mm made from mild steel. The hub was attached to the centre of the wheel with the aid of spokes. The essence of the wheel is to enable easy movement while the implement is in use.

4) U-channel - The U-channel is made of a steel plate of 1.5mm thick with dimension 124mm × 120mm × 51mm. The U-channel creates a fulcrum base for the ground wheel, blade, handle and the connecting flat bar linking the hub to the U-channel.

5) Blade connecting bar - This is made of steel flat bar 250mm×51mm. It acts as linkage from the blade head piece to the U-channel with the aid of bolt and nut.

6) Ground wheel connecting bar - This is made of a flat bar 295mm × 24mm. It connects the ground wheel from the hub to the U-channel. The ground wheel connecting bar is a two-piece flat bar.

7) Blade headpiece - This unit consists of mild steel with dimensions 170mm x 125mm. The blade is connected to the blade headpiece by means of bolt and nut. It is curved to almost a semi U-shape. 4. Design Theory 4.1 Shaft Design The design of the shaft for the rigidity was based on the shaft diameter, the maximum impact force by the operator and soil resistance force. The detailed calculation for obtaining maximum load that the shaft would be subjected to during operation is presented in equation 1 (Khurmi and Gupta, 2005).

3max32

πτBMd = . ............................................... (1)

where d = diameter of the shaft (m) BM = bending moment (Nm) τ = the allowable shear stress, 99999N/m2,

4.2. Finite Element Method The first step in the Galerkin Finite Element Method is the discretization of the domain. Here, the domain of the problem (length of the beam) is divided into a finite set of line elements, each of which has at least two end nodes. The second step is to obtain the weak form of the differential equation. Therefore, the corresponding system can be represented as equation 2 (Rao et al., 2012).

=

−−−

L

LLP

UUU

U

LLLLLLLLLLLL

LEL

616

1

246266126122646612612

4

3

2

1

2

22

3

.........(2)

[Stiffness matrix][Displacement matrix] = force vector.

The system of equations is solved using MATLAB. Results were found for various numbers of elements under different loads. The computer program used for calculating the shear force, bending moment, slope and deflection diagrams for the design of the push-type adjustable long-handle weeder is presented in Appendix I. 4.3 Design for U-Base Channel Figure 4 shows the U-channel base. The U-base channel is carrying load along section (x-x), therefore equation 3 (Khurmi and Gupta, 2005) shows the section modulus.

( )btaH

btaHIZ XXXX +

+≥ 2

2 ................................................(3)

BahthbBhI XX

353 )( +−−= ..............................................(4)

Equation 4 is the moment of inertia at (x-x). where

a = area of the section H = height of the U- channel h = height of the inner U-channel b = width of the inner U- channel t = thickness of the U- channel B = width of the outer U – channel

Figure 4. U- channel base

4.4 Design of Cutting Blade Figure 5 shows the cutting blade. The draft force (D) is the resultant of the normal loading of the soil on the metal and component parallel to the blade.

222 WVD +=

)(,22 NWVD += ………………………….(5)

Figure 5. Cutting blade

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 42: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

S.O. Nkakini and A. Husseni. Development and Evaluation of Wheeled Long-Handle Weeder 40

When soil-acting mechanical weeder-control implements are used, the soil is subjected to cutting or shear force which cause it to fail. Also the movement of the soil acting elements of a weeder through the soil is affected by: adhesion of the soil /material, friction between soil and metal (cutting blade) described by the angle of soil/metal friction.

The relation between the resistance to soil sliding over the metal (cutting blade) surface is given by equation 6 (Hendrick and Bailey,1982) as

Hmax = CA + W tan …………….......................(6) where

Hmax = the maximum soil/metal sliding force A=Area of metal in contact with the soil. W = normal loading of the soil on the metal Tan = adhesion of the soil material interface

Angle of inclination of the blade (angle of attack) The blade has to be inclined at a suitable angle so as

to allow easy penetration into the soil and to avoid excessive tilling of the soil. Angle of attack φ of approximately 150 is ideal to lift and separate the weeds from the soil. The draught force of weeder can be calculated from equation 7 (Yadav and Pund 2007)

SRxdwxWD = ...................................................(7) where

D = Draft force of the weeder (N) dw = Depth of cut (m) W = Width of cut (m) Rs = Soil resistance (N/m2)

4.5 The Power Requirement Power required to move weeder and weed the grasses is calculated from equation 8 (Yadav and Pund 2007)

( )hpSxDP75

= .......................................................(8)

where D = Draft force of the weeder, N S= Travelling speed (m/s)

4.6 Design of the Handle Figure 6 shows the handle. The handle is made of a cylindrical steel pipe of diameter 20mm.

213 FFF +=

abFF ×

= 21

22

3 Fa

bFF +×

= ……………………………….(9)

It is subjected to bending forces and is determined by equation 10 (Hall et al., 1980).

w

hhh R

Ym=δ ............................................................(10)

where hδ = bending stress on handle,

mh = moment in handle, Yh = distance from neutral axis,

Rw = second moment of area

Figure 6. The handle

5. Experimental field layout and Performance Test A field test was conducted to evaluate the performance of the developed weeder in terms of field capacity, weeding index, plant damage, performance index, field efficiency and effective field capacity and the weeding efficiency for comparison with the short handle hoe and others.

The test was conducted in a field of Okra plant having an inter-row spacing of 800mm with an area of 6m by 2m. The field area of 12m2 was divided into four plots of 2m by 1m wide with a space of 0.5m between each of the plots. Each of the plots was further subdivided into four blocks of 2m by 800mm. The blocks were denoted as block 1, block 2, block 3 and block 4, respectively. The weeding operations in four of the blocks were randomly carried out. These were conducted in four replicates and average readings were taken. Before and after the weeding operations, numbers of grasses with average plant height of 610mm and grasses which varied between 70mm-200mm were randomly counted and recorded. With the aid of a stop watch, times taken for weeding each of the blocks for instance, block 1, block 2, etc. were recorded, excluding the turning times of the weeder. Altogether there were sixteen (16) treatments. The forward speeds of the operator were computed respectively. The depths and widths of cut were measured using a steel rule. The number of weeds per area before and after weeding operations were counted and recorded. The number of the damaged plants were counted and recorded and the required draught force was determined.

In the test, anthropometric and ergonomic data were collected from 10 randomly selected farmers, males and females of age group between 20 to 50 years. The weeding operations were carried out by each of the participating farmers. The sequence of weeding operations was done until the area of the field was completed. The ergonomic parameters analysed were based on human body measurement which includes body weight, standing height, arm reach, palm length and functional leg height, which were all measured. At the end of the weeding operations, each of the operators gave information concerning ergonomic rating within

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 43: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

S.O. Nkakini and A. Husseni. Development and Evaluation of Wheeled Long-Handle Weeder 41

the range 0 to 10 i.e. uncomfortable to very comfortable rating, after which statistical modelling was employed.

Proper ergonomic design is very necessary for the construction of this implement for ultimate comfortability, convenience and safety of the operators. Soil texture of the experimental field was determined using sieve analysis. The soil was found to be sandy-loam, with average moisture content of 12%. 5.1 Weeding Index Weeding index can be calculated using the following equation 11 (Yadav and Pund, 2007).

100(%)1

21 xW

WWindexWeeding −= ........................(11)

where W1 = number of weeds per area before weeding W2 = number of weeds per area after weeding

5.2 Plant Damage Plant damage percentage is measured using the following equation 12 (Yadav and Pund, 2007).

1001 xpqQ

−= ..........................................(12)

where Q = plant damage q = number of plant in a 6m row length after weeding P = number of plant in a 6m row length before weeding

Thus, DP = P – Q the number of plant damaged (DP) 5.3 Performance Index The weeder performance was accessed through performance index (PI) by using equation13 (Yadav and Pund, 2007)

FaqePI = .............................................................(13)

where a = field capacity of weeder (ha/hr) q = plant damage (%) e = weeding index (%) F = required draught force

5.4 Field Capacity and Efficiency Theoretical field capacity

(FCt) = S x W, ha / hr ........................................(14) where

S = forward speed of weeder, m/s2 W = width of the implement, m

Effective field capacity (FCe) =

hrha

timeland

Field efficiency = %100xFCFC

t

e ...............................(15)

where FCe = Effective field capacity FCt = Theoretical capacity

6. Results and Discussions The performance data of the weeder is presented in Table 1, while in Table 2, showed the comparison of field performance of the wheeled long and short handle hoes. Table 1. The performance parameters of the wheeled long- handle

hoe S/no Description Parameters 1 Width of cut 0.45m 2 Depth of cut 0.4m 3 Weeding index 86.5% 4 Field capacity 0.050ha/hr 5 Weeding efficiency 91.7% 6 Weight 5.3 kg 7 Performance index 1108.48 8 Plant damage percentage 8%

Table 2. Comparison of field performance of the wheeled long-

handle and Short handle hoes Implements Moisture

Content (%)

Actual Field

Capacity (ha/hr )

Weeding Index ( % )

Plant Damage

( % )

Performance Index

Wheeled long-

handle hoe

13.09 10.05

0.050 0.045

86.6 84.4

8 10

1108.48 1899.0

Short handle hoe

13.09 10.05

0.028 0.014

83.3 84.4

7 12.5

901.2 1342.7

It is indicated in Table 2 that as moisture content decreased, there was also a decrease in actual field capacity. For example, at the moisture content of 13.09%, the actual field capacities were 0.050 and 0.08ha/hr for long-handle hoe and short handle hoe while at moisture content of 10.05%, the actual field capacities were 0.045 and 0,014 ha/hr respectively. This might be due to stickiness of soil which causes clogging of soil weed mass in weeding element. The weeding index was found to be in the range of 83.30 to 86.6% at different moisture contents. The developed wheeled long hoe has a maximum weeding index of 86.6% and a minimum value of 84.4% while short handle hoe had 84.4% and 83.3%. This could be as a result of differences in the soil moisture contents.

The highest plant damages of 10% and 12.5% at 10.05 moisture content were obtained from wheeled and short handle hoes respectively. There is every tendency for an increase in the number of plant damaged at moisture content below 10.05%. This is so because with decrease in moisture content soil hardness is increased, hence causing difficulties in penetration of weeding blade to desired depth, and sometimes skid over and strike the plant. A higher percentage of plant damage at 13.09% moisture content of the short handle hoe can be attributed to carelessness of the farmer. The highest performance index of 1899.0 and 1342.7 were obtained respectively at 10.05% moisture content, while the lowest value of 1108.48 and 901.2 were obtained at 13.09% moisture content. The decrease in the

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 44: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

S.O. Nkakini and A. Husseni. Development and Evaluation of Wheeled Long-Handle Weeder 42

performance index may be due to lower field capacity and higher plant damage.

It is shown in Table 3 that the existing developed machines by some researchers and their field performance. These weeders were engine powered. For instance, the engine-powered, manually operated roto- weeder employing the principle of a rotary tiller was powered by a 1.45hp petrol engine (Nkakini et al., 2009). Engine-powered rotary weeder for wet land paddy was developed by Viren and Ajav (2003), but had difficulties of manoeuvrability and was not easily affordable by peasant farmers. The Push-type Oscillatory power weeder was also engine powered (Kamal and Babatunde, 1999). They all ended up in introducing gasses such as carbon monoxide into the environment, because they were engine powered. The wheeled long-handle weeder is friendly to the environment, since it does not emit carbon monoxide into the environment. It is also simple and affordable by peasant farmers for small-scale farm mechanization. Table 3. Field performance Comparison of other similar existing

weeders Implements (Weeders)

Field capacity (ha/hr)

Weeding efficiency

(%)

Width of cut(m)

Depth of cut(m)

Wheeled long-handle weeder

0.050 ha/hr

91.7% 0.45m 0.4m

Traditional Short handle hoe

0.028 ha/hr

87.5% 0.35m 0.6m

Engine-powered, manually operated roto-weeder

0.037 ha/hr

90% 0.35m 0.3m

Push type Oscillatory power weeder

0.036 ha/hr

81.34% 0.36m 0.4m

Wheeled hoe based on ergonomic factors

0.048 ha/hr

92.5% 0.45m 0.5m

The weeder is used for weeding operation along the inter-row spaces of crops and weeding operations are done in an upright position, resulting in the reduction of backache, pains in the spine and waist pain by the operator. Thus, it is different from the traditional short handle hoe (Singh, 1992). Furthermore, this weeder also has other advantages such as reduction of labour, time and drudgery when compared with traditional weeders. This weeder performs well in flat dried soil. Table 4 presents material bill for construction of wheeled long-handle weeder.

Table 4. Material bill S/No. Descriptions Materials Quantity 1 Round pipe Galvanised pipe 2.44m long 2 U- Channel Steel plate of mild steel iron 1 3 Bolts and nuts Mild steel 18 4 Flat bar Mild steel iron 0.609m 5 Shaft Round iron of mild steel 1 6 Ground wheel Rubber and mild steel iron 1 7 Tyre Rubber 1 8 Hub Mild steel 1

The effect of width of cut on efficiency of the machine is shown in Figure 5. The result indicates that from 0.2m to 0.3m width of cut, the efficiency increased to 93.5%. The highest efficiency of 95% was attained at 0.4m width of cut.

Figure 5. Effect of width of cut on efficiency

The effect of speed on efficiency of the weeder is

shown in Figure 6. The weeding efficiency linearly increased from 90% to 91.7% within the speeds of 0.02m/s to 0.04m/s. The weeding speed of 0.04m/s recorded the maximum efficiency of 91.7% which sharply dropped to efficiency of 90% at 0.06m/s. It then remained constant from 0.06m/s to 0.8m/s at weeding efficiency of 90%. This indicates that the best weeding speed for push-type wheeled long-handle weeder is 0.04m/s.

Figure 6. Effect of speed on efficiency

Figure 7 shows the effect of angle of cut on

efficiency. The trend shows a non-linear relationship between the angle of cut and efficiency. It is obvious that at 15° angle of cut, the highest efficiency of 90% was obtained, while the lowest efficiency of 60% was obtained at 20° angle of cut. The best angle of cut is 15°

and followed by 30°.

7. Conclusion A wheeled push-type long- handle weeder was designed, fabricated and tested. The weeder performed

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 45: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

S.O. Nkakini and A. Husseni. Development and Evaluation of Wheeled Long-Handle Weeder 43

Figure 7. Effect of angle of cut on efficiency

satisfactorily with a weeding efficiency of 91.7%, weeding index of 86.5% and performance index of 1108.48. A field capacity of 0.050 ha/hr at width of cut of 0.45m and depth of cut of 0.4m was obtained. The comparison of field performance of the weeders showed that the performance index of short handle is a bit higher than wheeled long- handle weeder at 13.09% moisture content. The wheeled push-type long-handle weeder is user friendly and easy to maintain. It is however not common in the commercial market because of lack of awareness and the fact that it is more expensive than the short handle hoe.

On the whole, it is a better option because of the standing position of the operator which eliminates backache, pains at the spine and lower waist region of the operator, reduction in time spent in operation and the energy/force applied. This wheel-long handle hoe consists of a wheel, a weeding blade, U-channel and an adjustable long handle to enable the operator to use it even at an erect/standing position giving it a better edge over the short handle hoe.

Appendix I:

% A COMPUTER PROGRAM TO SOLVE AND PLOT THE SHEAR FORCE AND... %BENDING MOMENT DIAGRAMS FOR THE DESIGN OF A MANUALLY OPERATED WEEDER Wth=1.%Width of Cut dth=.4%Depth of Cut Rs=80e3%Soil Resistance Force Df=Wth*dth*Rs%Draft Force F3=Df F1=-F3/2 F2=-F3/2 Ft=[F1 F1 F3 F3 F3 F2 F2]'%Shear Force x=[0 1 1 2 3 3 4]'%Length of Shaft BM=Ft.*x%Bending Moment subplot(2,1,1) plot(x,Ft,'--*') xlabel('Length of Shaft(m)') ylabel('Shear Force(N)') title('Shear Force Diagram') grid on subplot(2,1,2)

plot(x,BM,'--*') xlabel('Length of Shaft(m)') ylabel('Bending Moment(Nm)') title('Bending Moment Diagram') grid on gtext('BMmax') x(cm) SF(N) BM(Ncm) 0 -16000 0 1 -16000 -16000 1 32000 32000 2 32000 64000 3 32000 96000 3 -16000 -48000 4 -16000 -64000 References: Hall, A.S., Holowenko, M.S and Loughlin, H.G (1980), Theory

and Problem of Machine Design, McGraw Hill, New York. Hendrick, J.G. and Bailey, A.C. (1982), “Determining components

of Soil- metal sliding resistance”, Transaction of the ASABE.25 (4) 0845-0849(doi:10.13031/2013.33625 @1982.

Jadhav, R.D. and Tubatmath, P.A. (1991), “Development and performance evaluation of bullock drawn multipurpose hoe”. Agricultural Mechanization in Latin America, Africa and Asia, Vol. 22, No. 2 pp. 9-14.

Kamal, A.R. and Babatunde, O.O. (1999), “Development of push-type oscillatory power weeder”, Journal of Agricultural Engineering and Technology, Vol. 7, pp. 23-28.

Khurmi, R.S. and Gupta, J.K (2005), A Textbook of Machine Design, Eurasia Publishing House (PVT), Ram Nagar, New Delhi-110-055.

Nwuba, E.I.U. (1981), Human Energy Demand of Selected Agricultural Hand Tools, M. Eng. Thesis (Unpublished), Department of Agricultural Engineering, Ahmadu Bello University, Zaria, Nigeria.

Nag, P.K and Dutta, P. (1979), “Effectiveness of some simple agricultural weeders with reference to physiological responses”, Journal of Human Ergonomics, Vol. 8, pp.11-21

Nkakini, S.O., Akor, A.J., Efenudu, O.E. and Ayotamuno, M.J. (2009), “Design and fabrication of an engine powered roto-weeder”, Journal of Agricultural Engineering and Technology (JAET), Vol.17, No.1, pp.6-14.

Odigboh, E.I.U. and Ahmed, S.F. (1980), Development of a ridge profile weeder. NIJOTECH Vol.4. No.1, pp 1-8.

Rangasamy, K., Balasubramanium, M. and Swaminathan, K.R. (1993), “Evaluation of power weeder performance”, Agricultural Mechanization in Asia, African and Latin America, Vol. 24, No. 4, pp.16-18.

Rao, G.S., Comissiong, D.M.G., Jordan, K. and Alana, S. (2012), “A finite element solution of the beam equation via MATLAB”, International Journal of Applied Science and Technology, Vol.2 No. 8, October, pp.80-85

Singh, G. (1992), “Ergonomic considerations in development and fabrication of manual wheel hoe weeder”, Indian Journal of Agricultural Engineering, Vol 2, No.4, pp.234-243.

Singh, G. (1988), “Development and Fabrication Techniques of Improved grubber”, Agricultural Mechanization in Asia, Africa and Latin America, Vol.19, No.2, pp.42-46.

Viren, M.V. and Ajav Verns, A. (2003), “Design and development of power operated rotary weeder for wet land paddy”,Agricultural Mechnization in Asia, Africa and Latin America. Vol.3, No.4, pp.27-32.

Yadav, R and Pund, S. (2007), “Development and ergonomic evaluation of manual weeder”, Agricultural Engineering, CIGR E-Journal, Vol.9, pp.5-7.

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 46: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

S.O. Nkakini and A. Husseni. Development and Evaluation of Wheeled Long-Handle Weeder 44

Authors’ Biographical Notes

Silas O. Nkakini is a lecturer in Department of Agricultural and Environmental Engineering, Rivers State University of Science and Technology Nkpolu-Oroworukwu, Port-Harcourt, Nigeria. He is a holder of MSc degree in Agricultural Machinery Engineering, and has authored many articles in reputed journals. His major research interest is design of simple agricultural machines and also in soil tillage operations. Mr. Nkakini has designed some simple agricultural implements such as Cassava lifter, Maize sheller and Engine powered weeder. He is currently researching on tractive force effect on agricultural soil during tillage

operations. Abu Husseni is a holder of B. Tech. Degree in Agricultural and Environmental Engineering, Rivers State University of Science and Technology Nkpolu-Oroworukwu, Port-Harcourt Nigeria. He is a research student in area of design of agricultural machines. ■

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 47: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

S.A.Oke et al.: A Robust Regression Approach for Excess/Shortage Spare Parts Cost Estimation 45

A Robust Regression Approach for Excess/Shortage Spare Parts Cost Estimation

Sunday A. Oke a,Ψ, , Desmond E. Ighravweb, and Gholahan Shyllon c

Department of Mechanical Engineering, University of Lagos, Lagos, Nigeria

a,c E-mail: [email protected] bE-mail: [email protected]

Ψ - Corresponding Author (Received 24 June 2014; Revised 31 October 2014; Accepted 30 January 2015)

Abstract: Spare parts management has been an area of increasing interest to service engineers in the current decade due to its potentials of improving business performance in terms of profit improvement, downtime minimisation and improved service deliveries through direct and indirect means. The present investigation deals with the development of a predictive model for estimating the amounts of spare parts holding and the cost effects of poor spare parts holding in a system. The model effectively uses an integrated methodology of the penalty cost and the wear technique for the unconstrained optimisation of the excessive spares using big-bang big-crunch (BB-BC) algorithm and the data of a petroleum products off-loading service company from shipping vessels. The model is validated by comparing the results obtained using the in-sample analysis with an out-sample approach. The results obtained show that the proposed spare parts monitoring model has the potential of predicting with high accuracy when used for in-sample and out-sample predictions. The developed excess spare parts model predicts that spare parts exhibit non-linearity under interest and inflation considerations. It shows that it is infeasible to track spares under changes in monetary polices of a country and in evaluating the cost of excess/shortage of stockings of multi-items in spare parts inventories.

Keywords: Wear rate, maximum wear, robust regression, big-bang big-crunch, spare parts

1. Introduction Just-in-time (JIT) spare parts supplies describe an arrangement in which spares are requested shortly before the time of need. Supplies are met without delay by the suppliers. In developing countries such as Nigeria, there is great difficulty in meeting up with the challenges and requirements of JIT systems due to the limiting factors such as erratic electricity supply, traffic congestion and poor delivery networks that militate against efficient JIT delivery systems. Consequently, the JIT system does not work properly; spare parts must be requested several days before needs to avoid interruptions in production or service activities due to spare shortages. Excess spares are therefore kept. Consequently, stocks are difficult to manage without keeping excess for the future periods. There are often problems of supply shortages, which lead to loss of production or service since the required spares for repairs of broken-down machineries or equipment/facilities may not be available when needed by the production plant or service.

Therefore there is a general understanding that keeping excessive spare stocks would stabilize production or service activities. However, it is widely known that excess/shortage stock handling is a cost to the organisation, which indirectly reduces the organisation’s profit margin and may significantly increase maintenance expenditure. Consequently, it

reduces spendings in other vital aspects of maintenance activities. Pressure is usually exerted on maintenance management to hold minimum amounts of excess stocks from purchases. The intention to minimise excessive quantities held is to keep the cost implication of maintenance stock to the minimum.

In this paper, the development of an analytical framework to evaluate excessive stocks kept by the organisation is pursued. The model’s application will lead to the reduction in maintenance cost since approximate amounts of spare parts needed could be estimated with minimal errors. It reduces the amount of money spent on spare parts. This model will help in reducing the amount of money paid as interest since less amounts of spare parts will be tied down in inventory (spares). Also, in organisations that have loose controls, where pilferages and obsolescence of stored spares are common, the model will help in the proper tracking of spare parts as approximate amounts will be planned for. This will reduce or eliminate unaccounted cost in production/manufacturing as well as service costs. Given these motivations for the current work, the main objective of the current paper was to develop a mathematical model that addresses the gap of paucity of mathematical models and/or quantitative measures for estimating the cost of poor spare part stockings.

The structure of the paper is organised as follows:

ISSN 0511-5728 The West Indian Journal of Engineering

Vol.37, No.2, January 2015, pp. 45-53

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 48: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

S.A.Oke et al.: A Robust Regression Approach for Excess/Shortage Spare Parts Cost Estimation 46

Section two presents the literature review, and Section three describes the research methodology. A numerical example and discussion of results are contained in Section four. Section five presents the conclusions of this study. 2. Literature Review In trying to manage spare part problems, different authors have presented a wide variety of spare part models that are problem-dependent. One technique, which has dominated spare parts literature is the use of optimisation models. Krenenbury and van Houtum (2009) developed an optimisation model that minimises multi-item, single-stage spare parts provisioning costs by considering the effects of commonality of spare parts. This was achieved using Dantzig-Wolfe decomposition, which aids the work in obtaining lower and upper bounds and heuristic solution for optimal costs.

In another contribution, Monte Carlo simulation was utilised by Lanza et al. (2009) to develop a stochastic optimisation algorithm that minimises the long-term cost of preventive maintenance for machine tools. The objective was achieved by solving the problem of fatigue failure of spare parts resulting from varying loads. Similarly, combined probabilistic simulation and genetic algorithm were utilised by Marseguerra et al. (2005) to develop a multi-objective Pareto optimisation model that maximises system revenue while minimising total spare parts volume. The use of probabilistic simulation aids the study in modelling system failure, repair and replacement while genetic algorithm was used as a solution method in solving the multi-objective problem.

Krenenbury and van Houtum (2009), Lanza et al. (2009) and Marseguerra et al. (2005) have all addressed critical issues on spare parts management. The question of how to evaluate the cost of excessive spare parts at various operational times in an inventory system was not adequately addressed in their studies. Solving this problem will help in tracking the frequency of disbursing funds for spare parts acquisition, thus, leading to proper management of organisational resources. This problem is considered in the current study. Also, a multi-objective model for machine and equipment that are repaired only by exchanging bad parts with good ones was proposed by Nosohi et al. (2011). In their work, a non-dominated solution was obtained for the model that minimised purchasing cost of spare parts, cost of preventive maintenance and production, residual lifetime of spare parts and corrective failure using minmax and e-constraint method.

From critique perspective, the issue of cost management of spare parts inventory as it affects business operations, using various optimisation concepts, was not addressed by Nosohi et al. (2011). The interest of business managers is what will be the monetary value for implementing a plan (i.e. how much will the company save on the long run?). An evaluation

of long-run savings requires the use of simulation models that has the capacity of evaluating various instances of stock control impacts on business profits. Stock control with emphasis on excessive stocks at various times was not directly investigated in the aforementioned studies, which opens the way for investigation on excessive spare parts control. Thus, the current paper examines how much stock will be left out at the various instances in an inventory system.

Furthermore, Kumar and Knezevic (1997) and Wang (2012) have also proposed the use of optimisation technique in solving spare parts-related problem. In the work of Kumar and Knezevic (1997), the problem of maximising availability and spares requirements was solved using Excel solver and branch-and-bound procedure. They considered series-parallel structures for spare part management. However, incorporating excessive stock value in planning was not considered in the studies by Kumar and Knezevic (1997), although this would have made it more robust.

Wang (2012) presented an optimisation algorithm for solving spare parts inventory and preventive maintenance problem jointly. Optimal order quantities of spare parts and preventive maintenance intervals were determined using enumerative and stochastic dynamic programming algorithms. The use of stochastic dynamic programming algorithm aided modelling of the expected costs. Jilka et al. (2011) analyse logistic network optimisation problem for aircraft spare parts using prognostic health management concepts while Regaltieri et al. (2005) investigated the accuracy of twenty different forecasting techniques on lumpy demand for aircraft spare parts. They report that weighted moving averages, the Croston method and exponential weighted moving average models are the best for forecasting this kind of demand. Since spare parts failure rate is usually stochastic, maximising their availability may result in excessive stocking at an acceptable level. The estimation of such excess stocks needs to be considered in order to properly utilise business resources.

Several authors who apply similar and other approaches in solving this problem include the following. Lau et al. (2006) develop a mathematical model of multi-echelon systems for carrying out corrective maintenance under passivation. The model estimates time-varying expected back-order and operational availability correctly. When compared with the results from Monte Carlo simulation, the same outcome was obtained. Cobbaert and Van Oudheusden (1996) extend economic order quantity (EOQ) inventory model by considering the variation and constant obsolescence of spare parts under conditions where storages are allowed or not. This was achieved using probability concepts for both cost and risk that are associated with spare parts obsolescence. van Jaarsveld and Dekker (2011) develop an optimisation algorithm that minimises shortages using reliability-centered maintenance data. The model has the tendency of

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 49: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

S.A.Oke et al.: A Robust Regression Approach for Excess/Shortage Spare Parts Cost Estimation 47

enhancing stocking decisions. Avoiding shortage of spare parts is one phase to spare parts stocking problem. The other phase which involves excessive stocking cost estimation was not addressed in their work.

de Smidt-Destrombes et al. (2006) estimate the expected values for maintenance time, operational time and lead time update for maintenance in developing two models that analyse the relation among spare parts inventory levels, repair capacity, maintenance policy and system availability for a k-out-of-N system. In order to demonstrate the superiority of partial pooling of spare parts over total pooling of spare parts, a multi-item, multi-location, single-echelon inventory system with lateral transshipment was developed. Kalchischmidt et al. (2003) present a framework for managing supply chains with different numbers of echelons, multi-model and extremely variable demand. They observe that improvements in supply chain performance can be achieved through proper collation of purchasing plans information. Haffer (1995) models the economic order quantity, economic order point, lag period and variations in monthly demand resulting from seasonal agriculture operations in order to aid the management of spare parts inventories of agriculture machineries dealership. The model was converted into software to ease usage.

Dohi et al. (1995) present two mathematical models for dealing with stochastic lead time and expected ordering options for spares inventory. The work analysed optimal ordering policy that minimises long-run average cost. Ghobbar and Friend (2003) develop a forecasting model for dealing with random nature of aircraft maintenance repair parts demand. The model enhanced the selection of appropriate forecasting method for critical demand of spare parts. The problem of reorder point for spare parts inventory was solved by Chang et al. (2005) while Chang and Tsao (2010) work on the determination of the quantities of spare parts for rolling components using analytical network process (ANP) in determining the ratio of preventive to corrective maintenance. We focus the current work on excessive spare parts stocking cost which the above studies lack. Another difference between these studies and the current paper is that we provide a structure that can be used in simulating spare parts usage with spreadsheets while others did not provide relevant information in this direction. Peres and Noyes (2006) investigate the use of maintenance support system for on-the-spot-demand and manufacturing. Kargari et al. (2012) consider Euclidean distance function, order concurrency and lot size in developing three different clustering algorithms for automobile spare parts. The work of Ke and Wang (2007) applies direct search algorithm in minimising total cost function for two kinds of spare parts. From the reviews, no documented work was found on modelling the cost of excessive stocking of spare parts. As the overwhelming importance of understanding this aspect of research has been stated from the results of

the above review, there is an urgent need to investigate the quantitative implications of keeping excessive stocks. This is particularly important in environments where business turbulence has led to closure of factories while many more are on the verge of collapse. The case in the manufacturing sector of the Nigerian economy is an example, which demands urgent attention to this gap. This work is motivated by the need to bridge this knowledge gap in order to save the failing industries from collapse and enrich the spare parts literature. The current paper considers wear rate of machine parts in proposing a mathematical model for determining level of spare parts in an inventory system and none of the previous studies has considered this approach. The introduction of this approach serves as a contribution to spare parts inventory literature. 3. Research Methodology In this section, brief explanations of the assumptions used in facilitating the development of the proposed predictive model and the various mathematical expressions used in modelling cost of poor spare parts management are presented. 3.1 Model Assumptions The assumptions used in the development of the proposed model are as follows: 1) The factory (system) has enough space to stock

excess spare parts. This creates the opportunity for purchasing excess spares when there is a known reduction in the unit cost of spares. This helps in reducing the production cost. However, associated with this benefit is the problem of what level of excess spares should be allowed in a system? The proposed model provides insights on how to manage this problem.

2) The problem of spare parts obsolescence is insignificant. By considering this assumption, the problem of excess stocking of spares may exist in an organisation.

3) Production time or usage of equipment can stretch beyond the timeframe specified for them. This implies that the total volume of spares wear-out at the end of a planning period may vary from the expected to another value.

4) Spare parts are believed to have the same initial level of wear before usage. This assumption is applicable to spares which can be reconditioned.

3.2 Model Development From Figure 1, the area of the rectangle ABCD gives the total cost of excess stocks of spare parts. By using accounting and maintenance records, the values of spare parts ordering cost (P1), holding cost (P2) and purchasing cost (P3) are used to determine the expected excess spare parts (Es) for a system. Traditionally, it is difficult to estimate due to its stochastic nature. However, to obtain

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 50: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

S.A.Oke et al.: A Robust Regression Approach for Excess/Shortage Spare Parts Cost Estimation 48

the value of ENs, mathematical model may be required. This study proposes a mathematical model for computing ENs in manufacturing industries. To know the amount of excess spare parts for machinery and equipment that undergo wear and tear, the volume of wear and tear in a unit period need to be estimated. Equation 1 gives the derivation for the amount of wear per hour (m3/h). This equation was obtained using dimensional analysis.

HtTDw = …. Eq.1

where T is Torque of an equipment (Nm), H is hardness of material that undergoes wear (N/m2) and t is the observation time (h).

Figure 1. Evaluating excess stocks of spare parts

With the knowledge of Equation 1, the amount of spare parts that will be required at different production or service periods can be estimated. The excess wear computations could be made using the maximum allowable wear (Wmax), the number of time corrective maintenance is carried out, testing time for restored equipment, the service time for unit output or the production time unit output as well as the estimated annual demand for organisation service. Since the unit time required by a team to produce a batch of product or render service to prospective customer varies from one team to another, to estimate the total time that the equipment is put into use, the total productive and unproductive times could be estimated using a stochastic expression.

Equation 2 is used in estimating the expected productive time for equipment. In this study, we present Equation 3 as a means for estimating the expected unproductive time of equipment (Ut) using the average number of times that corrective maintenance is carried out in a month (η) and the average time required to test and restore equipment back to service state during corrective maintenance (w). Thus, the total time that the equipment (At) is used in a particular period can be

represented as Equation 4.

∫∞

∞−

= dxxfPt )(µd .…. Eq.2

∫∞

∞−

= dvvwfU t )(η .…. Eq.3

ttt PUA += .…. Eq.4 where, δ is the periodic demand for a service, f(x) is the probability density function of time that a machine is used for productivity tasks and f(v) is the probability density function of time a machine is run during maintenance activities.

By combining Equations 1 and 4, the total volume of materials (Vi) that will wear and tear if the operating conditions (wear coefficient, contact area) are constant and the materials used for the spare manufacture is the same, is given as Equation 5.

twi ADV = .…. Eq.5 Using the maximum wear (wmax) allowable (i.e. the extent to which a part will wear up before it will be changed), the number of spare parts (Ni) that will be required in a given period can be represented as Equation 6.

maxwV

N ti = .…. Eq.6

The problem of spare part sometimes involves wastages during replacement of parts in maintenance system. To address this loss in the proposed model, we consider fixed quantity spare parts loss (Ψ) as a proportion of the quantity be required in a particular period. Thus, Equation 7 is a modified version of Equation 6 that considers what-if condition when dealing with spare parts losses in a maintenance system.

Ψ+

=Otherwise

,considered is loss If )1(

max

max

wV

wV

Nt

t

i

.…. Eq.7

At the beginning of an operation year, the amount of spare parts in a system (Si) may be expressed as Equation 8. For other periods, the amount of inventory in a system will be based on the decision on whether or not to order for spare parts and what quantity of spare parts in a system currently exists. This decision is often based on the reordering point (RP) for inventory in a particular system. The mathematical expression for this decision is presented as Equation 9.

iii BS φ/= .…. Eq.8

<+

>=

Otherwise,

if

RPSBQ

RPSQS

ii

i

i

i

φ

.…. Eq.9

where, Bi is the amount budgeted for a particular spare part, Q is the quantity of inventory in a system before

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 51: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

S.A.Oke et al.: A Robust Regression Approach for Excess/Shortage Spare Parts Cost Estimation 49

reordering and iφ is the forecast unit cost price for spare part i. The three costs associated with inventory management are ordering cost (P1i), holding cost (P2i) and purchasing cost (P3i). These costs are used in evaluating the cost of excess stocking of spare parts in an organisation. A linear interrelation is assumed among these costs, and this relationship is represented as Equation 10. The computation of quantities of excess spare parts (Ei) i in a system, which is a function of the differences between Si and Ni is given as Equation 11. The cost of stocking excess spare parts for a single item in a system is presented as Equation 12.

iiiit PPPP 321 ++= .…. Eq.10 The penalty cost (PCit) and the unit cost of keeping excess spare parts in a system can be used to jointly analyse the spare parts management of a system. This thought is presented as Equation 12.

ititit NSE −= .…. Eq.11

=tEPC

tEPCSP

itit

ititit periodin exist parts sparesin shortage Otherwise,

periodin exist parts spares excess If

.…. Eq.12 By considering the amount of money paid as interests for loans used in acquiring and keeping excess spare parts, this study modifies Equation 10 to incorporate the prevalent interest rate (τ) and the inflation rate (f). According to Ardalan (2000), the interrelationship among τ, f and i is given as Equation 13. By the combination of Equations 12 and 13, the actual cost for stocking excess spare parts (CEi) in a single item is obtained as Equation 14.

ffi

−−

=1τ

.…. Eq.13

1

21

−−+

=

i ti t

i ti t

i t

EP C

ffEP

C S P

τ

If excess spares exist in period t Otherwise, shortages in spare parts exist in period t. …. Eq.14

.…. Eq.14

With Equation 14, the cost of excess spare parts for multiple spares in a system can be estimated. We considered Equation 15 in evaluating the cost of excessive multiple spare parts for a system. The cost for poor spare part management, which incorporates excess spare parts and penalty costs for spare parts shortages is given as Equation 16.

∑∑= =

−−+

=N

iitit

T

tS f

fEPMCE1 1 1

21 τ .…. Eq.15

∑∑= =

=N

i

T

titCSPCSP

1 1

.…. Eq.16

The inclusion of τ and f into Equation 14 introduces

non-linearity into this equation. This non-linearity becomes more obvious in a system where spare parts are obtained from different environments. The changes in the monetary policies of a country may result in a drastic rise or fall in the cost of excess spare parts obtained from overseas. Thus, it may be deduced that spare parts cost may exhibit non-linearity tendency under interest and inflation considerations. With significant variations in the amounts of spare parts used in a particular period, there may be the tendency for the cost of excess spares to follow as a fluctuating trend. To the control the number of spare parts that will be changed, there is the need to optimise the level to which a part must wear before being changed during maintenance activities. This problem motivates the optimisation of wmax using the Equation 17. This equation is considered as an unconstrained optimisation model, where bounds will be set for the various wmax in this equation. We select the big-bang big-crunch (BB-BC) algorithm as a solution method in this study due to its low computational time and high quality solution abilities. Since the value for Vi will vary from one period to another, the current study sets the bound for Vi in order to obtain realistic values for wmax.

Min Z =

−−+∑= ))(1(

))(21(

max

max

1 i

iiin

ii wf

VBwfP

φφτ .…. Eq.17

BB-BC algorithm is a meta-heuristic which employs the principle of population-based stochastic search in obtaining optimal values for decision variables based on assigned solution space assigned to decision variables in a combinatorial problem. This is achieved by computing the center-of-mass for each decision variable (big-bang) and updating the value of decision variables (big-crunch). From the BB-BC literature (Rao and Yesuratnam, 2012; Sakthivel et al. 2013), the centre-of-mass for a decision variable can be estimated as Equation 18 or simply taken as the global solution.

=−

=−−

= P

igi

P

igigij

cjg

f

fww

1)1(

1)1()1(

1 .…. Eq.18

To generate a new particle using centre-of-mass of decision variables in BB-BC algorithm, the current study utilises Sakthivel et al.’s (2013) approach using Equation 19.

gwwr

ww jjcjgjg

)( minmax

1

−+=+

a .…. Eq.19

where, g is known as step size, r is a normal distribution which lies between -1 and 1, and α is a parameter used to control the search space of decision variables. After generating the cost of poor spare parts management in a system for different planning periods, the information obtained is used in developing a predictive model that can be easily applied by decision

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 52: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

S.A.Oke et al.: A Robust Regression Approach for Excess/Shortage Spare Parts Cost Estimation 50

makers in a system. The proposed predictive model is controlled mainly by the amount of spare parts in a system as a dependent variable and the expected volume of wear of a spare part during a planning period as an independent variable. This study then applies robust regression models using STAT software. The flowchart for the proposed predictive model used in this study is presented as Figure 2, while a summarised implementation procedure for the model is presented as follows: Step 0: Determine the optimal value of wear that spare

parts will experience before being repaired or replaced using the BB-BC algorithm.

Step 1: Determine the number of working hours per day, the number of working days per year and the total overtime periods in a production/ service year.

Step 2: Determine the initial wear of spare parts, the maximum wear allowable and the wear rate per day of spare parts.

Step 3: Determine the quantity of spare parts stocked at the start of a production/service period and unit cost of spares.

Step 4: Determine the quantities the spare parts used in a production/service period.

Step 5: Determine the quantities of excess spare parts, if none, go to step 6.

Step 6: Determine the cost of excess spare part. Repeat steps 2 to 6 for multi-item spare parts.

Step 7: Sum the cost of poor spare parts management in the system.

Step 8: Develop a predictive model using the cost obtained in step 1 as dependent variables, spare parts level and the amount of wear as independent variables.

Step 9: Validate the developed predictive model using out-of-sample approach.

Figure 2: Flowchart computing cost of excess spare parts

4. Numerical Example and Discussion of Results To validate the proposed model in this study, a case study of a haulage company spare parts inventory system is considered. This study focuses on marine unloading arm of mechanical and hydraulic spare parts. The company studied unloads vessels from ships that arrive from other locations. The unloading arm is engaged during this operation and the mechanism of operation of this unloading arm is electrically-, mechanically-, and hydraulically controlled. From our observations, the most frequently used spare parts for marine unloading arm are mechanical and hydraulic spare parts majorly D1¼” balls and V-seal VA-0040, respectively. The part numbers for these spare parts are 70V400080534 and 246153040894 for D1 ¼” balls and V-seal VA- 0040, respectively. Due to the difficulty in obtaining the ordering costs and holding costs for the two spare parts, these costs were taken as 5% of the unit cost for each of the items based on one of the authors’ industrial experiences. At the time of conducting this study, one United States dollar is approximately N165. We compliment the practical datasets obtained with laboratory simulation. This aids the applicability of the proposed model.

The determination of the range for wear volume for the two spare parts considered are 0.01 to 0.04 and 0.01 to 0.03 mm3 for D1¼” balls and V-seal VA-0040, respectively. The total wear volume expected at the end of a planning period for D1¼” balls and V-seal VA-0040 are taken as 6 and 0.5 mm3, respectively. Using the maximum epoch of (100) as the stoppage criterion for the BB-BC algorithm, the optimisation of Equation 16 gives the results of the optimal wear volume for D1¼” balls and V-seal VA-0040 as 0.046 and 0.034 mm3, respectively.

With the knowledge of the optimal total wear volume for D1¼” balls and V-seal VA-0040 spare parts and the amount of wear that a spare part could experience before it is replaced, the amount of spare parts to order for D1¼” balls and V-seal VA-0040, based on Equation 6 are 118 and 20 units, respectively. From the above discussion, the proposed model has the potential for determining the minimum amount of spare parts inventory for a system given that actual parameter values in Equation 16 are available to decision makers.

Since the amount of service rendered to prospective clients in a shipping business does not always follow a linear distribution, the current study simulates the amount of total wear to expect at the end of each planning period (month) using Monte Carlo simulation. The application of classical inventory models like economic order quantity model can be used to determine the reorder point for each spare part in a system. The information used in estimating the cost of excess/shortage in spare parts for the case study is presented in Table 1.

Output solution

Initialisation: nx, xi, D, n

Determine: Sp, Po, PCit

Determine Esp: Are all items treated?

Stop

No

Yes

Yes

Determine cost of Esp

Is nx = nmax?

Sum each cost of CSP

No

Start

No

Yes

Create initial values for decision variables for each particle

Evaluate the quality of each particle

Calculate centre of mass

Create new values for each particle using center of mass

Evaluate the quality of the new particle

Check stoppage criterion

Develop a predictive model for excess and shortage spare parts cost

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 53: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

S.A.Oke et al.: A Robust Regression Approach for Excess/Shortage Spare Parts Cost Estimation 51

Table 1. Spare parts information D1¼” balls V-seal VA-0040 Unit cost N1694.92 N450 Unit excess spare cost N1844.92 N4650 Unit shortage cost N1800 N4600 Inflation 11 % 11 % Interest rate 16 % 16 % Re-order point 60 7

This information is combined with the results

obtained in the previous paragraphs in this section. The

cost of excess/shortage in spare parts for the mechanical spares (i.e. D1¼” balls) is first considered. The results for the cost of excess/shortage in D1¼” balls spare part when simulated are presented in Table 2.

A negative value for the cost of excess/shortage in spares for the system as shown in Table 3, implies that the system is experiencing shortages in spare parts during that particular period while a positive value for the cost of excess/shortage in spare for a system indicates excess spare parts in the system.

Table 2. Simulation results for D1¼” balls spare parts Period Budget

(N) Quantity purchase

Quantity in the system

Wear (mm3)

Quantity used

Spare part in the system

Excess/shortage cost (N)

1 200000 118 118 1.22 26 92 169732.20 2 0 0 92 5.01 109 -109 -196200.00 3 200000 118 9 2.09 45 73 134678.81 4 0 0 73 2.32 50 -50 -90000.00 5 200000 118 68 5.61 122 -4 -7200.00 6 200000 118 114 3.64 79 39 71951.69 7 200000 118 157 5.16 112 6 11069.49 8 200000 118 124 2.97 64 54 99625.42 9 200000 118 172 4.56 99 19 35053.39

10 200000 118 137 4.31 94 24 43200.00 11 200000 118 142 1.36 30 88 162352.54 12 0 0 88 1.26 27 -27 -48600.00 13 200000 118 91 4.9 106 12 22138.98 14 200000 118 130 2.73 59 59 106200.00 15 200000 118 177 3.57 78 40 73796.61 16 200000 118 158 2.92 63 55 101470.34 17 200000 118 173 1.39 30 88 162352.54 18 0 0 88 1.86 40 -40 -72000.00 19 200000 118 78 2.51 55 63 116229.66 20 0 0 63 2.53 55 -55 -99000.00 21 200000 118 63 1.26 27 91 167887.29 22 0 0 91 2.04 44 -44 -79200.00 23 200000 118 74 2.02 44 74 136523.73 24 0 0 74 1.61 35 -35 -63000.00

Table 3. Simulation results for V-seal VA-0040 spare parts Period Budget

(N) Quantity purchase

Quantity in the system

Wear (mm3)

Quantity used

Spare part in the system

Excess/shortage cost (N)

1 90000 20 20 0.08 2 18 83700.00 2 0 0 18 0.19 6 12 55800.00 3 0 0 12 0.47 14 -2 -9200.00 4 90000 20 18 0.42 12 6 27900.00 5 90000 20 26 0.52 15 11 50600.00 6 0 0 11 0.32 9 1 4600.00 7 90000 20 21 0.20 6 15 69750.00 8 0 0 15 0.26 8 8 37200.00 9 0 0 8 0.26 8 0 0.00

10 90000 20 20 0.26 8 13 59800.00 11 0 0 13 0.24 7 6 27900.00 12 90000 20 26 0.45 13 12 55800.00 13 0 0 12 0.43 13 0 0.00 14 90000 20 20 0.09 3 17 78200.00 15 0 0 17 0.07 2 15 69750.00 16 0 0 15 0.24 7 8 37200.00 17 0 0 8 0.06 2 6 27900.00 18 90000 20 26 0.53 16 10 46500.00 19 0 0 10 0.13 4 6 27900.00 20 90000 20 26 0.38 11 15 69750.00 21 0 0 15 0.43 13 3 13950.00 22 90000 20 23 0.22 7 16 74400.00 23 0 0 16 0.30 9 7 32550.00 24 0 0 7 0.19 6 2 9300.00

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 54: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

S.A.Oke et al.: A Robust Regression Approach for Excess/Shortage Spare Parts Cost Estimation 52

To evaluate the performance of the proposed model for hydraulic system spare parts, which undergoes wear, V-seal VA-0040 spare parts information is used and the results obtained is depicted in Table 4. The ability of the proposed model to detect when shortages in spare parts occur has also been verified from the results in Table 4. Thus, it can be inferred that the proposed model has the capacity to be used as a simulation model when studying spare parts management for both mechanical and hydraulic spare parts.

The applicability of robust regression model in modelling the cost of excess/shortage in spare part for marine unloading arm is carried out. The robust

regression model used is implemented using STATA /STA 11.0 software. The quantity of spare parts in the system, the amount of wear at the end of a planning period and the volume of spare parts at the planning period are taken as explanatory variables in determining the cost of excess/shortage in spare part for the system. Table 4 shows the results for in-sample datasets for the first 18 periods.

In order to validate the performance of the robust regression model, information in Tables 2 and 3 (periods 19 to 24) was used in carrying an out-sample datasets validation of the robust regression model. The results obtained are presented in Table 5.

Table 4. In-sample comparison of robust regression model results with simulated results Periods (Month)

D1¼” balls V-seal VA-0040 Actual value Forecasted value Actual value Forecasted value

1 169732 169732 83700 83700 2 -196200 -199299 55800 55800 3 134679 134679 -9200 -9300 4 -90000 -90449 27900 27900 5 -7200 -7379.7 50600 51150 6 71951.7 71951.7 4600 4650 7 11069.5 11069.5 69750 69750 8 99625.4 99625.4 37200 37200 9 35053.4 35053.4 0 1.14E-10

10 43200 44278 59800 60450 11 162353 162353 27900 27900 12 -48600 -48016 55800 55800 13 22139 22139 0 -2.15E-10 14 106200 108850 78200 79050 15 73796.6 73796.6 69750 69750 16 101470 101470 37200 37200 17 162353 162353 27900 27900 18 -72000 -72000 46500 46500

Table 5. Out-sample comparison of robust regression model results with simulated results Periods (Month)

D1¼” balls V-seal VA-0040 Actual value Forecasted value Actual value Forecasted value

1 116229.7 116224.4 27900 27900 2 -99000 -99673.7 69750 69750 3 167887.3 167887.3 13950 13950 4 -79200 -79379.7 74400 74400 5 136523.7 136523.7 32550 32550 6 -63000 -62775.4 9300 9300

The prediction results for D1¼” balls spare part show that the robust regression model has the capacity to predict the cost of spare part shortages as indicated with the negative costs in Table 5. The mean absolute percentage error (MAPE) of the robust model when used for predicting in-sample datasets for D1¼” balls and V-seal VA-0040 are 0.6 and 0.06 %, respectively. These results show that the robust regression model has high predictive accuracy. To further test the suitability of the robust regression model, the MAPE for out-sample dataset obtained for D1¼” balls prediction is 0.21 % while it is 0 % for V-seal VA-0040, respectively. Thus, it can be deduced that robust regression model is well suited as a predictive tool for the case study under

concurrent occurrence of excesses and shortages in spare parts in a system. The application of this predictive model will reduce the burden of service engineers in evaluating the different equations in order to obtain excess/surplus spare part cost. 4. Conclusions In this study, a mathematical model that can be used in estimating the cost of excess/shortage of spare parts, the level of spare parts in a system and the quantities of spare part to order has been successfully developed. Also, a predictive model for the cost of poor spare parts management was presented. The quest to determine the optimal level of spare parts wear before it can be

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 55: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

S.A.Oke et al.: A Robust Regression Approach for Excess/Shortage Spare Parts Cost Estimation 53

changed was successfully carried out using the proposed model as an unconstrained optimisation model. This affords the opportunity to experiment the potentials of BB-BC algorithm as a solution technique for the unconstrained model and satisfactory results was obtained. A combined real and laboratory simulated datasets are used in evaluating the performance of the model and we observed that the proposed model is effective and efficient in estimating the cost, number of excess stocking of spares. We conclude that the proposed model can be used in evaluating the cost of excess/shortage of stockings of multi-items in spare parts inventories.

The proposed model can be used as a simulation tool in observing variations in spares usage for production/service that are either repairable or changed once failures occur. This work deepens our understanding of the negative and positive influences of excessive stock in the economy of organisations. Thus, excess/shortage costs could retard the economic progress of organisation by causing more charges on usage. From the review of literature, the cost of excess spare parts using conventional algorithm and computational intelligence does not exist. Thus, future research can be conducted on these since spares are used in maintenance activities, joint optimisation of maintenance activities and spare usage can be modelled in future investigations.

References: Ardalan, A. (2000), Economic and Financial Analysis for

Engineering and Project Management, Technomic Publishing Company, Inc. Pennsylvania, USA.

Cobbaert, K. and van Oudheusden, D. (1996), “Inventory models for fast-moving spare parts subject to ‘sudden death’ obsolescence”, International Journal Production Economics, Vol. 44, pp. 239-248.

de Smidt-Destombes, K.S., Vander Heiden, M.C. and van Harten, A. (2006), “On the interaction between maintenance part inventories and repair capacity for a k-out-of-N system with wear-out”, European Journal of Operational Research, Vol. 174, pp. 182-200.

Dohi, S.T. and Osaki, S. (1997), “Models for 1-out-of Q systems with stochastic inventory”, European Journal of Operational Research, Vol.103, pp. 255-272.

Haffer I. (1995), “SPAM: A computer model for management of spare parts inventories in agricultural machinery dealership”, Computers and Electronics in Agriculture, Vol.12, pp. 323-332.

Jilka, N., Thiruravukkarasu, A., Lendermann P., Gan B.P., Schirrmann A., Formm, H. and Wang, E. (2011), “Making use of prognostics health management information for aerospace components logistics network optimization”, Computers in Industry, Vol. 62, pp. 613-622.

Kalchischmidt, M., Zotteri, G. and Verganti, R. (2003), “Inventory management in a multi-echelon spare parts supply chain”, International Journal Production Economics, Vol.81-82, pp. 397-413.

Kargari, M. and Sepehri, M.M. (2010), “Stores clustering using a data mining approach for distributing automotive spare parts to reduce transportation costs”, Expert Systems with Applications, Vol. 39, pp.4740-4748.

Ke, J-C. and Wang, K-H. (2007), “Vocation policies for machine

repair problem with two-type spares”, Applied Mathematical Modelling, Vol. 31, pp. 880-894.

Kumar, U.D. and Knerzevic, J. (1997), “Spare optimisation models for series and parallel structure”, Journal of Quality in Maintenance Engineering, Vol. 3, pp. 117-188.

Krenenbury, A.A. and van Houtun, G.J. (2009), “A new partial pooling structure for spare /parts networks”, European Journal of Operational Research, Vol. 99, pp. 908-921.

Lanza, G., Niggerschmidt, S. and Werner, P. (2009), “Optimisation of preventive maintenance and spare part provision for machine tools based on variable operational conditions”, CIRP Annals Manufacturing Technology, Vol.58, pp. 429-432.

Lau, H.C., Song, H., See, C.T. and Cheng, S.Y. (2006), “Evaluation of time – varying availability in multi-echelon spare parts systems with passivation”, European Journal of Operational Research, Vol. 170, pp. 91-105.

Marseguerra, M., Zio, E. and Podofillini, L. (2005), “Multi-objective spare part allocation by means of geometric algorithms and Monte Carlo simulation”, Reliability Engineering and System Safety, Vol.87, pp. 325-335.

Nosohi, I. and Hejezi, S.R. (2011), “A multi-objective approach to simultaneous determination of spare part numbers and preventive replacement times”, Applied Mathematical Modelling, Vol.35, pp.1157-1166.

Peres, F. and Noyes, D. (2006), “Envisioning e-logistics developments: Making spare parts in-situ and on demand, state-of-the-art and guidelines for future development”, Computers and Industry, Vol. 57, pp. 490-503.

Rao, C.V.G.K. and Yesuratnam, G. (2012), “Big-bang and big-crunch (BB-BC) and firefly optimisation (FFO): Application and comparison to optimal power flow with continuous and discrete control variables”, International Journal on Electrical Engineering and Informatics, Vol.4, No. 4, pp. 575-583.

Regaltieri, A., Gamberi, M., Gamberini, R. and Manzini, R. (2006), “Managing lumpy demand for aircraft spare parts”, Journal of Air Transport Management, Vol. 4, pp. 426-431.

Sakthivel, S., Arun, P.S., Marikani, S. and Selvi, K.S. (2013), “Application of big-bang big-crunch algorithm for optimal power flow problems”, The International Journal of Engineering and Science, Vol. 2, No. 4, pp. 41-47.

van Jaarsveld, W. and Dekker, R. (2011), “Spare parts stock control for redundant systems using reliability centered maintenance data”, Reliability Engineering and System Safety, Vol. 96, pp. 1576-1586.

Wang, W. (2012), “A structure model for joint spare parts inventory and planned maintenance optimization”, European Journal of Operational Research, Vol. 216, pp. 127-139.

Authors’ Biographical Notes: Sunday Ayoola Oke received Ph.D. from the University of Ibadan in 2008. He lectures in the Department of Mechanical Engineering, University of Lagos, Lagos, Nigeria. His research interests include maintenance, production, composites and optimisation.

Desmond Eseoghene Ighravwe is a Ph.D. student in the Department of Mechanical Engineering, Ladoke Akintola University of Technology, Ogbomoso. His current area of research includes maintenance, production and optimisation.

Gholahan Shyllon is a graduate of the Department of Mechanical Engineering, University of Lagos, Lagos, Nigeria. ■

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 56: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

K. Ramesar, C. Maharaj and U. Persad: A Mechanism for Cutting Coconut Husks 54

A Mechanism for Cutting Coconut Husks

Kishan Ramesara, Chris Maharajb,Ψ, and Umesh Persad c

a,b Department of Mechanical and Manufacturing Engineering, The University of the West Indies, St. Augustine Campus, St.

Augustine, Trinidad and Tobago, West Indies. E-mails: [email protected]; [email protected]

c Centre for Production Systems, The University of Trinidad and Tobago, O'Meara Campus 78-94 O'Meara Industrial Park, Arima, Trinidad and Tobago, West Indies; E-mail: [email protected]

Ψ - Corresponding Author (Received 29 August 2014; Revised 19 November 2014; Accepted 30 January 2015)

Abstract: This paper details the conceptual design of a machine for cutting coconut husk halves into pieces for activated carbon production. Alternative interlocking and welded blade arrangements are presented with the potential for scaling up the processing of coconut husks into smaller pieces. Virtual simulations and the experimental testing of a functional prototype are used to validate the conceptual design. The design is shown to be functionally acceptable, and directions for further improvements and development are outlined.

Keywords: Machine design, conceptual design, coconut shell processing

1. Introduction Activated carbon has the strongest physical adsorption forces of any material known to mankind and it is applied in numerous industries including the Semi-Conductor, Petrochemical and Gold Recovery Industries (Cheremisinoff, 2002a). It is the most commonly used product for the adsorption of volatile organic compounds from vapour and gas phases (Cheremisinoff, 2002b; Wypych, 2001) and is typically applied in water treatments ranging from municipal water treatments to even treatment of landfill leachate water (Grand View Research, 2014). This USD$2 billion product industry is traditionally made from bamboo, coconut shells, wood, ignite and coke (Grand View Research, 2014; Rodriguez-Reinoso, 2001). Currently, there is a need for finding inexpensive and effective alternatives to the existing commercially available activated carbon because the environmental sustainability and benefits for future commercial applications could be improved (AlOthman et al., 2013).

Recent studies have shown that activated carbon made from coconut husks can adsorb impurities ranging from dyes (Purkait et al., 2005) to heavy metals such as arsenic (Manju et al., 1998). In addition to this, it has been suggested that activated carbon made from coconut husks can be used as a replacement for the existing commercially available activated carbon (Gupta et al., 2010). However, in utilising coconut husks for activated carbon production, there is a need to resize the coconut husk pieces to improve the efficiency of the conversion process. In anticipation of industrial scale devices that are capable of cutting the coconut husks into required

sizes, this paper presents a conceptual design for a cutting mechanism that addresses this problem. 2. Background to the Study There are two main methods of manufacturing activated carbon: physical activation and chemical activation. However, many variations of these processes have been employed and in some instances both of these processes have even been combined (Gupta et al., 2010; Low & Lee 1990; Manju et al. 1998; Vargas et al. 2011). Figure 1 highlights the main steps in activated carbon manufacturing.

Figure 1. The activated carbon manufacturing process

ISSN 0511-5728 The West Indian Journal of Engineering

Vol.37, No.2, January 2015, pp.54-62

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 57: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

K. Ramesar, C. Maharaj and U. Persad: A Mechanism for Cutting Coconut Husks 55

In using coconut husks for producing activated carbon, the coconut husks are initially cut into smaller pieces as the first preparatory step. This is done to increase the efficiency of the carbonisation process where a size of approximately 50 mm cubed was recommended (Dyall, 2012).

The production of activated carbon from coconut husks has not been attempted on an industrial scale.

Therefore, in anticipation of this requirement to effectively and efficiently size coconut husk pieces for activated carbon production, cutting mechanisms embedded in devices such as wood chippers, guillotines, paper/plastic shredders and fruit/vegetable processing equipment are potentially applicable. Three of these prototypical mechanisms are shown in Figure 2.

Figure 2. Cutting mechanism configurations from filed patents

The majority of cutting mechanisms could be

regarded as multi-blade linear food cutters due to the plurality of blades used and relative motion between the food item and the device’s cutting blades. In some devices, as shown by Wheeler (1908) in Fig 2(c), the food item is stationary and the cutting blades descend with force through the food item thereby cutting it. The opposite is observed in some devices (Bulette, 1896; Farabough, 1935) as shown in Fig 2(a) and 2(b), where the food item is forced through stationary blades.

Further to this, the latter mentioned devices use plungers to force the food item through the device. In one configuration, the plunger is designed to accommodate movement beyond the surface of the blades to ensure the entire food item is cut (Farabough, 1935). In another configuration, the plunger uses small spikes to penetrate the top of the food item in order to restrict undesirable rotation and translation of the food item before or while it is being forced through the blades (Bulette, 1896).

Although all cutting mechanisms use a plurality of blades, there are differences in their orientations and connections. On one hand, in the design presented by Wheeler (1908), the blades are angled at 60˚ to each other and bolted together. In other designs, the blades are positioned perpendicular to each other, with either a welded configuration or an interlocking configuration (Farabough, 1935).

The means of actuation of the cutting process is also diverse as some designs utilise mechanical advantage via a lever system to amplify the cutting force applied

(Farabough, 1935; Wheeler, 1908) while others are actuated by the linear translation of a plunger (Bulette, 1896). In addition to this, the device designed by Gangi (1993) incorporates a hydraulic motor in conjunction with a pulley configuration to actuate the system.

The cutting devices are capable of cutting various food items into specific pieces, yet they are insufficiently equipped to cut coconut husks due to the size of the husk and the magnitude of the cutting force required. Other devices such as shredders were not considered as possible solutions because they are unable to cut the coconut husk into specific sizes. Therefore, there remains a need for an effective method of cutting coconut husks into pieces that could be scaled up for the future industrial production of activated carbon. 3. Methodology In order to address the need for an effective coconut husk cutting method, a survey of patents and existing cutting machines (wood chippers, guillotines, paper and plastic shredders and fruit and vegetable processing equipment) was undertaken to gain insight into current solutions. Additionally, in order to identify the main design requirement, an experiment was conducted to determine the magnitude of the force required to cut half the coconut husk at various orientations and ages. 10 machinists were also interviewed to determine their requirements for operating a machine to cut coconut husk halves into pieces, and their views were taken into consideration at the early conceptual design stages.

Varying alternative conceptual designs were

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 58: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

K. Ramesar, C. Maharaj and U. Persad: A Mechanism for Cutting Coconut Husks 56

generated to fulfil the main function of cutting the coconut husks halves into pieces. These were iteratively developed and evaluated against each other according to their ability to fulfil functional requirements. A final conceptual design was selected and refined to better satisfy ergonomic and safety requirements of the device to produce a final working concept (Ullman, 2009).

To validate the new design concept for cutting the coconut husk, a simulation study and an experimental study were performed. The Finite Element Analysis (FEA) simulation study was executed on CAD models using SolidWorks 2013. A physical prototype was constructed and an experimental study was also performed to gain insight into the performance of the cutting mechanism. The new design concept and the design validation studies would be detailed in the following sections. 3.1 Cutting Force Requirements A Tinius Olsen Benchtop Tester was fashioned with a specially fabricated blade and used to cut the coconut husk half sections at different orientations. This was done to investigate the force required to cut the coconut husk when the blade was angled to approximately 45˚ to the grains of the husk, parallel to the grains of the husk, and perpendicular to the grains of the husk. This was repeated for the same species of coconut husks that were less than 3 days old after being cut by the coconut vendor. A summary of the results is presented in Table 1.

Table 1. Results from Cutting Force Experiment Test Orientation

Force Parallel to grain (N)

Force Perpendicular to grain (N)

Force Angled at 45 degrees (N)

Test 1 819 2329 2264 Test 2 1163 1881 2138 Test 3 1222 2700 2107

Average 1068 2304 2169 The recorded maximum force required to cut the husk perpendicular to the grains (2700N) was more than double that which was required when the blade was oriented parallel to the grains (1222 N). This highlighted that there was a significant relationship between the cutting force required and the orientation of the coconut husk, given the fibrous nature of the coconut husk. It was observed that if the cutting force was applied parallel to the grains, the husk split along the grains, but if the cutting force was applied at an angle to the grains, the husk deformed before the grains were cut.

For the purposes of this study, the maximum force of 2700N was used as the maximum cutting force required by the cutting mechanism. A more extensive study to determine the statistical distribution of cutting force by type and orientation of coconut husk is required to fully characterise the design requirements. However,

for the purposes of developing the conceptual design presented in this paper, the maximum cutting force found experimentally coupled with an appropriate safety factor was deemed sufficient to serve as the maximum cutting force requirement. Pending such a study, the conceptual design presented could be easily scaled to accommodate a change in the maximum required cutting force. 3.2 Machine Conceptual Design A conceptual design for a coconut husk cutting machine was developed and this is shown in Fig. 3. The sequence of operation is as follows. The operator will place the coconut husk on the ‘Loading Area (4)’, and then the ‘Pusher (5)’ will be used to slide the coconut on top of the ‘Cutting Blades (10)’ in the ‘Cutting Chamber (3)’ of the device. As the ‘Pusher (5)’ advances towards the ‘Cutting Chamber (3)’, a pulley system, which connects the ‘Pusher (5)’ to the ‘Sliding Shield (2)’, raises the ‘Sliding Shield (2)’ to allow the coconut husk to enter the ‘Cutting Chamber (3)’. Further to this, when the ‘Pusher (5)’ is retracted to its initial position, so too does the ‘Sliding Shield (2)’. Figure 3. Isometric view of the cutting machine conceptual design

At this juncture, the ‘Controls Panel (6)’ will be

used to activate the system. The ‘Controls Panel (6)’ contains a Programmable Logic Computer (PLC) to control the hydraulic cylinder extension and retraction. The PLC is coded so that the two buttons, on either side of the ‘Controls Panel (6)’, must be activated for the cutting process to begin. This was done to ensure that both of the operator’s hands were away from the ‘Cutting Chamber (3)’ when the cutting process starts. Further to this, a limit switch is placed inside of the cutting chamber, so that after the plunger descends and

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 59: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

K. Ramesar, C. Maharaj and U. Persad: A Mechanism for Cutting Coconut Husks 57

cuts the coconut husk, a signal is sent to the PLC to retract the cylinder. This automatic step reduces the amount of human input required by the machine and also decreases operational time. In addition to this, the button at the centre of the ‘Controls Panel’ is the emergency stop button which will stop the process at any point once activated.

After the hydraulic system is actuated, the ‘Hydraulic Cylinder (1)’, at the top of the machine, will force the ‘Plunger Plate (12)’ to descend towards the ‘Cutting Blades (10)’. The plungers, which are attached to the bottom of the ‘Plunger Plate (12)’, will force the coconut husk through the arrangement of the ‘Cutting Blades (10)’, thereby, cutting them. The cut coconut husks will then fall onto the ‘Ejection Sheet (9)’ and slide out of the device. All of these components can be seen in Figure 3.

The proposed design solves the dilemma of the current state of the art, i.e. the insufficient cutting force, by utilising a hydraulic actuator to apply the cutting force to the system. The use of hydraulic equipment (specifically hydraulic presses) in the extraction of oils and fruit juices is fairly common (Fellows, 2009). However, this technology has not been applied in the fruit cutting industry as it has been in the current design.

As a consequence of the magnitude of the applied force and the toughness of the coconut husk, significant stresses will develop in the cutting blades. To counteract this, supports for the cutting blades in the central portion of the cutting area were incorporated into the design of the system. This feature was also absent in the devices reviewed. In addition to this, the cutting blades used in the proposed device are basically an evenly spaced grid of three rows of blades in one direction connected to three perpendicular rows of blades. However, the device is capable of using blades with different connections between them. On one hand, the blades can be welded together, as seen in Figure 4, akin to that highlighted by Bulette (1896).

Figure 4. Isometric view of the welded blade configuration

Alternatively, the blades can be a series of specially

machined interlocking blades, as shown in Figure 5, similar to that presented by Farabough (1935).

Figure 5. Isometric view of the interlocking blade configuration

Figure 6 shows the cutting mechanism assembly.

The ends of the ‘Cutting Blades (10)’ are wedged into groves cut into the ‘Blade Holder Strips (14)’. This provides a tight fit for the blades. In turn, the ‘Blade Holder Strips (14)’ are bolted onto the ‘Locator Plate’ and the combination of the two restricts motion of the blades. Furthermore, ‘Guide Pins (16)’ are bolted onto the ‘Locator Plate’ through the ‘Blade Holder Strips (14)’ on one end, and bolted onto the top of the machine at the other end. These ‘Guide Pins (16)’ are used to align the ‘Plunger Plate (12)’ over the arrangement of blades.

Figure 6. Isometric view of cutting mechanism assembly

The cutting process begins when the hydraulic

piston is actuated. This causes the ‘Plunger Plate (12)’ to descend towards the coconut husk placed on the blades. Attached beneath the ‘Plunger Plate (12)’ are individual ‘Plungers (13)’, as shown in Figure 6, which are specifically sized to fit between the blades and are carefully positioned to ensure that no collision takes place during the cutting operation. These ‘Plungers (13)’ apply the cutting force to the coconut husk.

When the ‘Plungers (13)’ make contact with the coconut husk, the initial force is transferred through the coconut husk onto its support, the blades, and its reaction is what starts cutting the husk. As the ‘Plunger Plate (12)’ descends further, the ‘Plungers (13)’ force the coconut husk between the blades. This continues

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 60: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

K. Ramesar, C. Maharaj and U. Persad: A Mechanism for Cutting Coconut Husks 58

until the ‘Plungers (13)’ advance beyond the surface of the blades thus cutting the coconut husk. The pieces of the husk then fall through the blades and the ‘Plunger Plate (12)’ is retracted to end the cutting process. 4. Manufacturing and Assembly Process of Cutting

Mechanism 4.1 Cutting Blades Two lengths of 800 mm x 1.5 mm x 25.4 mm of Steel Rule were used in the manufacturing of the cutting blades (10) (see Figure 6). They were cut into lengths of 230.2 mm via a Wire Electrical Discharge Machine (EDM). Subsequent to being cut to length, the Wire-EDM was again used to cut three 12.7 mm slots (one at the blade mid-point and two others spaced 50.8 mm from the mid-point). The slots were cut on the top of the blades that pointed in the x-direction (x-blades) and on the bottom of the blades that pointed in the z-direction (z-blades) when assembled, as seen in Figure 5. 4.2 Blade Holder Strips An 850 mm x 50.8 mm x 30 mm mild steel block was used to manufacture the 204.8 mm x 50.8 mm x 25.4 mm blade holder strips (14) as shown in Figure 6. The block was first milled to the required dimensions. Afterwards, three 25.4 mm tall by 12.7 mm deep grooves were cut into the 25.4 mm face of the strips via Wire-EDM. The slots were cut so that the cutting blades would be wedged into them as to restrict motion in all directions.

Following this, a pair of counter bore holes, to accommodate M8 Socket Head Cap Screws, was drilled on the 50.8 mm face of each of the strips. These holes were used to attach the strips onto a Locator Plate. Two additional 8 mm diameter through holes were drilled, reamed and counter bore with a 19.1 mm diameter to a depth of 3mm in one pair of strips, which were placed on either side of the cutting chamber. This was done to allow the Guide Pins (16), as shown in Figure 6, to be attached to the Locator Plate beneath the Blade Holder Strips. Finally, one pair of threaded 6.4mm holes was drilled on the strip with which the Loading Area (4), as shown in Figure 3, was bolted onto. 4.3 Plunger Plate A 350 mm x 350 mm x 20 mm mild steel plate was used to manufacture the Plunger Plate (12) in Figure 6. Firstly, the plate was machined via milling to a flat 306 mm x 306 mm plate, and then four 50.8 mm x 50.8 mm squares were cut out from the corners of the plate. Proceeding from there, four rows of four evenly spaced 38.1mm x 38.1mm x 6.35 mm slots were milled in the central portion of the plate and 8mm diameter through holes were drilled through their centres. This was done so the plungers (13) could be fitted into and bolted onto the Plunger Plate.

5. Design Simulation Study The most critical components in the cutting process are the cutting blades. A CAD model of the blade arrangement and geometry was used to investigate the responses of the cutting blades to various loading scenarios. The worst case scenario was modelled as a statically applied load acting on the central portion of one blade as shown in Figure 7. Von Mises Stress, resulting displacement and factor of safety studies were executed for both the interlocking and welded blade configurations. The cutting force modelled included a 2.4 factor of safety of the maximum recorded force (2700N) required to cut the coconut husk. This was due to the assumptions that the machine will be operated at room temperature and that the model accurately represented the system.

Figure 7. Interlocking blade configuration Von Mises stress analysis results

In the case of the interlocking blades configuration,

the Von Mises Stress Analysis illustrated that the most stressed portions of the blades were localised at the areas where the perpendicular blades connected with each other. This was expected because the area surrounding geometric discontinuities would have a stress concentration. In addition to this, it is expected that the majority of the deformation will be in the y-direction. The maximum stress was calculated to be 140 MPa which is less than the rated 210 MPa Yield Strength of the low-carbon steel material.

The resulting displacement study showed that the central portion of the blade deflected the most. This also was anticipated because that segment of the blade was only supported at its ends. More so, the defection is expected mainly in the y-direction. The displacement of 0.054mm as seen in Figure 8 is acceptable. The factor of safety study revealed that the weakest point in the blades was at the junction of the intersecting blades. The minimum factor of safety registered at 1.3 as shown in Figure 9.

In the case of the welded blades configuration, the Von Mises Stress analysis indicated that the most stressed area was the central portion of the interlocking blade assembly (see Figure 10). This was reckoned to be because the ends of the blades are supported and the

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 61: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

K. Ramesar, C. Maharaj and U. Persad: A Mechanism for Cutting Coconut Husks 59

blade has a constant cross-section. The maximum stress was calculated to be 42 MPa which is much less than the rated 210 MPa yield strength of the material.

Figure 8. Interlocking blade configuration displacement study results

Figure 9. Interlocking blade configuration factor of safety study results

Figure 10. Welded blade configuration Von Mises stress analysis

results

The resulting displacement study showed that the

upper portion of the majority of the blade deflected the most. This result was attributed to the constant cross-section of the blade. More so, the deflection is expected mainly in the z-direction. The displacement of 0.005mm as seen in Figure 11 is acceptable.

The factor of safety study resulted in a minimum factor of safety of 4.9 at the upper portion of the blade (see Figure 12). However, the majority of the blade appears to have a similar value. The constant cross-

section of the blade was again credited for this. Figure 11. Welded blade configuration displacement study results

Figure 12. Welded blade configuration factor of safety study results

Though these values can be deemed safe, it is in

excess of what is required which implies that the blade is over designed. A review of the summary of results from the simulation tests (shown in Table 2) shows that both types of blade configurations are capable of cutting the coconut husk with the welded blade design appearing superior.

Table 2. Summary of results from the Simulation Study Study Type Interlocking Blade

Design Welded Blade

Design Maximum Von Mises Stress

140 MPa 42MPa

Resulting Displacement

0.054 mm 0.005 mm

Factor of Safety 1.3 4.9

However, the ease of manufacture and availability of the particular blade designs need to be considered. The blades used in the interlocking configuration are readily available and only require the cutting of grooves on individual blades to complete its manufacture. This can be easily done via electrical discharge machining. Alternatively, the welded blade configuration must be fabricated in its entirety.

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 62: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

K. Ramesar, C. Maharaj and U. Persad: A Mechanism for Cutting Coconut Husks 60

The general maintenance or repair of the blades is also affected by the design. The blades in the interlocking configuration are capable of being separated from each other easily whereas the welded blade configuration is not capable of being disassembled. In the event that a particular blade is chipped or even broken, the said blade can be simply replaced in the interlocking configuration but the entire blade unit in the welded blade configuration will have to be replaced. Given these considerations, the interlocking blade configuration was selected as being better suited to the overall design requirements and it was used to fabricate a working prototype for further testing. 6. Experimental Testing of Prototype Four main experiments were performed with the prototype to investigate the functional acceptability of the design and whether variations in the position and orientation of the coconut husk, with respect to the blades, had any effect on the final product.

In the first two experiments, the coconut husk was positioned with the flat side faced downwards on the blades, and in the other test, the coconut husks were placed with the flat side faced upwards. Further to this, the coconut husks were oriented, in such a manner, so that the grains of the husk were parallel to a blade. These experiments were repeated, but instead of the grains of the husk being oriented parallel to the blade, they were oriented at an angle to a blade. More so, all of the main experiments were repeated for coconuts that were aged up to three days after being cut by the coconut vendor.

The prototype successfully cut the coconut husks into the required size, regardless of the position and orientation of the husks. However, more useable pieces were obtained from the ‘parallel to grain’ experiments. Figure 13 shows a ‘0 days old’ coconut that was cut by the machine which was positioned with its flat face on top of the blades and oriented with its grain parallel to a blade. Figure 14 shows a ‘3 days old’ coconut that was cut by the machine which was positioned with its flat face on top of the blades and oriented with its grain angled at 45˚ to a blade. Figure 13. Final product from cutting machine (0 days old, faced

downward, cut parallel to blade)

Figure 14. Final product from cutting machine (3 days old, faced

downward, cut angled to blade)

Further to this, it was observed that the coconut

husks exhibited behaviour of compression, deflection and fracturing under the applied load before they were fully cut which is an acceptable and typical behaviour for this type of solid food (Fellows, 2009). Also the lack of support to the cavity in the concaved coconut husk has been observed as the key reason for this behaviour.

In the experiment with the flat side of coconut facing upwards towards the plungers, when the plunger descended upon the husk, the portion of the husk that was in contact with the blades was briefly indented. At that point, the area of contact between the flat side of the husk and the plungers was larger than the area of contact between the coconut husk and the blades. Therefore, the force received by flat face of the husk was large enough for the reaction of the blades to indent the skin of the coconut.

As a result of the indenting/cutting, a larger area of the coconut was in contact with the cutting blades which also exposed grains that were perpendicular to the blades. As a consequence to that, the energy used to produce the cutting force was absorbed by the coconut and it compressed slightly. Then, as the plungers descended further because of the increasing cutting force, the deflection phase began.

In this phase, the flat face of the husk began to spread across the face of the plungers. This showed that the coconut husk had a strong ability to deform without sustaining permanent damage i.e. a relatively small modulus of elasticity. This, in conjunction with the cavity of the coconut husk, was attributed as an explanation for the behaviour of the coconut husk in this phase. The deformation continued until the coconut husk was almost flattened and then it fractured along its centre-most portions. At this juncture, the husk was unable to be compressed further and the plungers forced the husk through the blades thereby cutting it.

In the experiment with the flat side of coconut faced downwards, the results of this series of experiments were similar to when the husks were faced flat side up, in that the coconut husk exhibited similar behaviours of

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 63: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

K. Ramesar, C. Maharaj and U. Persad: A Mechanism for Cutting Coconut Husks 61

compressing, deforming and fracturing before being cut. This can be seen in Fig 15. However, in this experiment, the deflection phase came to a premature end when the coconut husk fractured before any significant deflection occurred. Afterwards, the plungers then proceeded to compress the husk until it was almost flat before finally being forced through the blades and ending the cutting process.

Figure 15. Fracture of coconut husk in cutting machine

7. Discussion A design process was executed to develop a mechanism to cut half of a coconut husk into smaller pieces ideal for conversion into activated carbon. After validating the overall conceptual design via virtual simulation, a prototype of the design was fabricated and it successfully passed performance testing by safely and effectively cutting the coconut husk into specific sizes regardless of the husk’s age, position and orientation.

Activated carbon produced from coconut husks has not been attempted on an industrial scale. As a result of this, customers (from whom the customer needs were to be derived) and existing equipment (from which system requirements were to be obtained) were not present. To overcome these issues, machinists were interviewed (due to their familiarity with operating equipment) and similar equipment was investigated. However, some parameters in the equipment examined were unavailable so their values were assumed during conceptual design. The conceptual design must now be subjected to more rigorous user and functional testing so that usability and acceptability requirements could be met.

Limitations in the research also included the number and type of coconuts used to investigate the cutting force required. Small samples of coconuts were used to investigate the force required to cut the coconut husk at the specified positions and orientations and, in addition to this, most were of the same species and collected from the same tree. Further to this, the cutting speed when investigating the cutting force required was not varied to determine whether there was a significant relationship between the cutting speed and the cutting force. Future

work can address these issues in addition to improving the overall operation of the cutting mechanism.

One such improvement can be for the machine to better accommodate the coconut husk. During the cutting operation, it was observed that the lack of support, for the concaved shaped coconut husk, caused it to be compressed until it was flat before it was cut. Although the effect of this is negligible, it is still undesirable. Assuming that the half coconut husks were positioned with their flat sides facing upwards, if the plungers that were in-line with the cavity of the husk were longer, they would make contact with the centre of the cavity first and support it. This will eliminate the compression of the husk before it was cut and would improve the quality of the cut pieces. However, the husk would have to be precisely placed on the cutting area with its cavity directly beneath the plungers.

Also, further research would have to be executed to determine the extension of the centre plungers required, in relation to the others, and the width of the plungers so that they fit inside of the husk. Another potential solution to the identified problem would be to curve the blades to accommodate the curved face of the husk assuming the same positioning. Also, the blades could be made to accommodate the cavity of the husk assuming positioning with the flat side facing downwards. Furthermore, both the plungers and blades can be fabricated to accommodate a particular positioning. Research into the curvature of the coconut husk and its cavity will have to be extensively performed to facilitate coconut husks of various ages and sizes.

Another proposal for future work is to automate the entire system. The reduction of human input can increase the efficiency of the system and reduce the time required to load, actuate and cut the coconuts. A constant supply of half-sectioned coconut husks can be provided to a cam actuated cutting mechanism via a conveyor. The only human input required would be to start the system and place the husks on the conveyor. However, this will need financial justification.

A final recommendation for future work is to surpass the limitation of cutting half coconut husks and to design a machine to process a whole coconut husk. The device can be made so that the husk is strategically cut in halves which then fall into a loading mechanism that delivers them to a pair of machines similar to the one designed in this project. This will expand the availability of raw material for this machine. More so, the device can also be designed to cut a whole coconut (one that still contains water) into the required pieces and also harvest the water inside which can be used for commercial benefit. 8. Conclusion A conceptual design for a mechanism to cut coconut halves has been presented. The results from a virtual simulation study and the functional testing of a

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 64: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

K. Ramesar, C. Maharaj and U. Persad: A Mechanism for Cutting Coconut Husks 62

fabricated prototype demonstrate that the proposed design is capable of cutting coconut husk halves regardless of position and orientation. Further work involves detailed performance testing, design optimisation and user acceptability testing in order to produce a working product for use in industry. It is hoped that the solution offered in this paper would contribute to the ease of production of activated carbon from coconut husks.

References: AlOthman, Z.A., Habila, M.A., Ali, R., Ghafar, A.A., El-din

Hassouna, M.S. (2013), “Valorization of two waste streams into activated carbon and studying its adsorption kinetics equilibrium isotherms andthermodynamics for methylene blue removal”, Arabian Journal of Chemistry, available at http://ac.els-cdn.com/S1878535213001275/1-s2.0-S1878535213001275-main.pdf?_tid=b7599dec-2fae-11e4-b0f9-00000aab0f6c&acdnat=1409338997_a2c24dee08b0a23f07a0a68781f4703e (Dated: 18th February 2014)

Bulette, O.A. (1896). Potato Cutter, US Patent 563652 Cheremisinoff, N.P. (2002a), “Ion Exchange and Carbon

Adsorption”, Handbook of Water and Wastewater Treatment Technologies, Elsevier, USA, pp. 404-440

Cheremisinoff, N.P. (2002b), “Prevention and Control Hardware”, Handbook of Air Pollution and Control, Elsevier, USA, pp. 446-448; 467

Dyall, K. (2012), “An investigation into the creation of charcoal and activated carbon from coconut by-products for use in metallurgy and filtration applications”, MSc Thesis in Industrial Innovation, Entrepreneurship and Management, The University of Trinidad and Tobago.

Farabough, G.M. (1935), Potato Cutter, US Patent 2004858. Fellows, P.J. (2009), Food Processing Technology - Principles

and Practice, Woodhead Publishing, England Gangi, J.C. (1993), Machine for Cutting Fruit into Sections, US

Patent 5241902 Grand View Research (2014), Activated Carbon Market Analysis

And Segment Forecasts To 2020, Grand View Research, California, USA

Gupta, V.K., Jain, R. and Shrivastava, M. (2010), “Adsorptive removal of Cyanosine from wastewater using coconut husks”, Journal of Colloid and Interface Science, Vol. 347, No.2, pp. 309-314

Low, K.S. and Lee, C.K. (1990), “The Removal of Cationic Dyes Using Coconut Husk as an Adsorbent”, Pertanika, Vol.13, No.2, pp. 221-228

Manju G. N., Raji C. and Anirudhan T. S. (1998), “Evaluation of Coconut Husk Carbon for the removal of arsenic from water”, Pergamonl, Vol. 32, No. 10, pp. 3062-3070.

Purkait, M. K., Das Gupta, S., and De, S. (2005), “Adsorption of eosin dye on activated carbon and its surfactant based

desorption”, Journal of Environmental Management, Vol.76, No.2, pp. 135-142.

Rodriguez-Reinoso, F. (2001), “Activated Carbon and Adsorption”, Buschow, J.K.H., Cahn, R.W., Flemings, M.C., Ilschner, B., Kramer, E.J., Mahajan, S. (eds), Encyclopedia of Materials - Science and Technology, Volumes 1-11, Elsevier, USA, pp. 22-35

Ullman, D.G. (2009), The Mechanical Design Process, McGraw-Hill, New York, USA.

Vargas, A.M.M., Cazetta, A.L., Garcia, C.A., Moraes, J.C.G., Nogami, E.M., Lenzi, E., Costa, W.F. and Almeida, V.C. (2011), “Preparation and characterization of activated carbon from a new raw lignocellulosic material: Flamboyant (Delonixregia) pods”, Journal of Environmental Management, Vol. 92, No.1, pp. 178-184.

Wheeler, M.M. (1908), Potato Cutter, US Patent 880057. Wypych, G. (2001), “Solvent Recycling, Removal, and

Degradation”, Handbook of Solvents, ChemTec Publishing, Canada, pp.1513-1514

Authors’ Biographical Notes Kishan Ramesar graduated from the Faculty of Engineering at The University of the West Indies in 2013 majoring in Engineering Design. He is currently working as a Teacher’s Assistant in the Faculty of Engineering in addition to pursuing a Master’s of Science Degree in Petroleum Engineering at The University of the West Indies.

Chris Maharaj is a Lecturer in the Mechanical and Manufacturing Engineering Department of The University of the West Indies (UWI). He holds BSc and MSc qualifications in Mechanical Engineering and Engineering Management respectively from UWI. He started his career as a Mechanical Engineer in Condition Monitoring and Inspection and worked in the industry for five years. He later went on to pursue his PhD at Imperial College London in Mechanical Engineering. His present teaching and research interests are in alternative use of waste materials, mechanical design optimisation, failure analysis, component life assessment, asset management, and innovation management.

Umesh Persad is an Assistant Professor in Design and Manufacturing at The University of Trinidad and Tobago. He obtained his BSc. in Mechanical Engineering (First Class) from The University of the West Indies, and his Ph.D. from The University of Cambridge in the area of Engineering Design, with a special focus on Inclusive Design and Healthcare Design. He is a member of ASME, ACM, and The Design Society. His research interests include the computational design of medical products and general product design. ■

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 65: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

I.O Oladele: Mechanical Properties of Steel-making Slag Reinforced Polyester Composites 63

Mechanical Properties of Steel-making Slag Reinforced Polyester Composites

Isiaka Oluwole Oladele

Department of Metallurgical and Materials Engineering, Federal University of Technology, Akure, Ondo State. Nigeria.

African Materials Science and Engineering Network (AMSEN): A Carnegie-IAS (RISE) Network E-mail: [email protected]

(Received 11 September 2014; Revised 27 November 2014; Accepted 30 January 2015) Abstract: In order to assess the viability of utilising steelmaking slag for reinforcing polyester matrix to form composites with improved mechanical properties, slag was obtained from an indigenous steel production plant and prepared by crushing and pulverizing. This was followed by sieving into 75, 106 and 300 µm sizes and, varied masses of the particles were used to develop the composites by reinforcing the unsaturated polyester resin with the steelmaking slag particles. The homogeneous mixtures were poured into the flexural and tensile tests moulds and allowed to cure before being stripped from the moulds. The samples were further allowed to cure for 30 days before carrying out the mechanical tests. The results showed that the composites produced have indeed gained increment in these properties compared to the unreinforced polyester material. The optimum results were obtained from the use 106 µm and 2 wt% slag particles.

Keywords: Steelmaking slag, reinforcement, composites, mechanical properties.

1. Introduction To cope with the obvious limitations of polymers, for example, low stiffness and low strength, and to expand their applications in different sectors, inorganic particulate fillers, such as micro-/nano-SiO2, glass, Al2O3, Mg (OH)2 and CaCO3 particles, carbon nanotubes and layered silicates, are often added to process polymer composites, which normally combine the advantages of their constituent phases. Particulate fillers modify the physical and mechanical properties of polymers in many ways (Fu et al., 2008).

It has been shown that dramatic improvements in mechanical properties can be achieved by incorporation of a few weight percentages of inorganic exfoliated clay minerals consisting of layered silicates in polymer matrices (Kim et al., 2004). Commonly used layered silicates have a thickness of ≈1 nm and lateral dimensions of ≈30 nm to several microns or larger. The large aspect ratios of layered silicates are thought to be mainly responsible for the enhanced mechanical properties of particulate-polymer nano-composites.

Many studies have been conducted on the mechanical properties of these particulate-filled polymer composites. Stiffness or Young’s modulus can be readily improved by adding either micro- or nano-particles since rigid inorganic particles generally have a much higher stiffness than polymer matrices (Zhu et al., 1999; Fu and Lauke, 1997; Wang et al., 1998). However, strength strongly depends on the stress transfer between the particles and the matrix. For well-bonded particles, the applied stress can be effectively transferred to the

particles from the matrix; this clearly improves the strength (Reynaud et al., 2001; Ou et al., 1998). However, for poorly bonded micro-particles, strength reductions occur by adding particles (Liang et al., 1997; Tjong and Xu, 2001). The drawback of thermosetting resins is their poor resistance to crack growth. But inorganic particles have been found to be effective tougheners for thermosetting resins. In contrast, most studies on thermo plastics filled with rigid particulates reported a significant decrease of fracture toughness compared to the neat polymers (Jancar and Dibenedetto, 1995; Fu and Lauke, 1998). There are, however, several studies that show toughness to have increased with the introduction of rigid particles in polypropylene and polyethylene, respectively (Pukanszky, 1995). Impressively, enhanced impact toughness has been reported for polyethylene filled with calcium carbonate particles. Enhancement of impact properties of some pseudo-ductile polymers by the introduction of inorganic particles has also been achieved (Fu and Wang, 1993; Bartczak et al., 1999).

Slag is the by-product of iron and steel production processes and, has been used in civil engineering for more than 100 years. Rapidly water-cooled Electric Arc Furnace Slag (EAFS), due to its relative high amorphous silica content which has pozzolanic activities, is to be employed in the production of blended cement. Even there are some research works about the properties of concretes, in which air-cooled and ground granulated EAFS were used as aggregates. The conclusions of these studies indicate that there is a great likelihood to use

ISSN 0511-5728 The West Indian Journal of Engineering

Vol.37, No.2, January 2015, pp.63-67

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 66: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

I.O Oladele: Mechanical Properties of Steel-making Slag Reinforced Polyester Composites 64

EAFS instead of natural aggregate in concrete. It has been widely employed as aggregate, mainly in base, sub-base and bituminous pavement for road construction, in which steel slag provides many advantages in comparison with natural aggregates (Chaurand et al., 2007).

Although many studies have been conducted on the evaluation of steel slag to be used in road construction, there are rare researches regarding the utilisation of steel slag in concrete. ASTM C336 gives specifications for the use of blast furnace slag as aggregates in concrete, while there is not such a standard for steel slag (Frias and Sanchez, 2004).

Chemical composition and temperature during manufacturing determines the structure of the slag which further influences other properties. Granulated slag aggregate is currently utilised mainly in earthworks of linear transport communications, as backfill in retaining structures, backfill of roads and pipeline structures, in motorway rugs and groundwork layers, for terrain modification works, during manufacturing of burnt brick products, and also for revitalisation of depleted mines. Beside granulated slag aggregate, which is not yet so widely used, slag is generally utilised in form of fine grounded blast furnace slag which is used as active and filler admixture to concrete.

Utilisation of finely ground slag is an established method of reusing of this metallurgical by-product, mainly in concrete and cement industries. Special attention has been directed at investigating the possibilities of it being used as substitute for natural mineral aggregates when producing asphalt mixtures. Results of analyses usually conducted when testing physical and chemical characteristics of natural mineral aggregates intended for the same purpose have been demonstrated (Motz and Geiseler, 2001).

Having noticed from previous works that particulate fibres are good source of reinforcement materials to fill and enhanced the properties of matrixes. And considering the wide application of steelmaking slag in ceramic matrix, this work was carried out to investigate the influence of steelmaking slag on unsaturated polyester material. This was done with the aim of turning waste to wealth by using slag for engineering applications.

2. Materials and Methods This research was carried out with the following materials; Unsaturated polyester resin, Ethyl Ethyl Ketone Peroxide (MEKP), Cobalt 2% in solution, polyvinyl acetate and ethanol. The steelmaking slag that was a waste product of electric arc furnace was sourced from universal steel Ikeja Lagos Nigeria. XRD was carried out to examine the composition of the slag, and it was as shown in Table 1. 2.1 Material Preparation

The steelmaking slag that was obtained in lumps form was crushed with hammer and finally pulverized using Denver laboratory ball mill. The particle from the process was sieved with sieve shaker 16155 Model into 75, 106 and 300 µm sieve sizes. XRD was carried out with Fluxana-HD Elektonik Vulcan Fusion Technology.

Table 1. XRD analysis for the steelmaking slag SiO2 Al2O3 Fe2O3 CaO MgO SO3 K2O 36.79 10.55 21.05 5.79 2.58 0.06 0.4

NaO LOI LSF SR AR Mn Ti 1.39 0.08 5.73 1.01 0.4 11.01 2.26

2.1 Material Preparation The steelmaking slag that was obtained in lumps form was crushed with hammer and finally pulverized using Denver laboratory ball mill. The particle from the process was sieved with sieve shaker 16155 Model into 75, 106 and 300 µm sieve sizes. XRD was carried out with Fluxana-HD Elektonik Vulcan Fusion Technology. 2.2 Mould Production Tensile mould of gauge length 90 x 5 x 3 mm of a dumb-bell shape and flexural mould of 150 x 50 x 3 mm were used for the production of tensile and flexural samples respectively from where the hardness samples were obtained. 2.3. Production of Composites To develop the composites, 1.5 g each of catalyst and accelerator was added to 120 g of the unsaturated polyester resin while the steelmaking slag particulate was varied in a predetermined proportion of: 2, 4, 6, and 8 wt%. After proper stirring, the homogenous slurry is poured into the mould and allowed to cure inside the mould at room temperature before it is removed. The cured sample is left for 3 weeks before the mechanical tests were carried out. 2.4.Mechanical Property Tests Cured composite samples were prepared for tensile and flexural tests after which Scanning Electron Microscope (SEM) was used to investigate the miscibility between the filler and matrix at the fractured surfaces. These tests were carried out as follows; (a) Determination of the tensile properties of the

materials - In the present study, tensile tests were performed on INSTRON 1195 at a fixed Crosshead speed of 10 mm min-1. Samples were prepared according to ASTM D412 and tensile strength of the standard and conditioned samples were calculated.

(b) Determination of the flexural property of the materials - Flexural test was carried out by using Testometric Universal Testing Machine in accordance with ASTM D790. To carry out the test,

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 67: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

I.O Oladele: Mechanical Properties of Steel-making Slag Reinforced Polyester Composites 65

the grip for the test was fixed on the machine and the sample that has been cut into the test piece dimension of 150 x 50x 3 mm, was hooked on the grip and the test commenced. As the specimen is stretched the computer generates the required data and graphs. The Flexural Test was performed at the speed of 100 mm/min.

(c) SEM Observation - SEM of the composites was observed using Zeiss SEM: Zeiss Ultra Plus 55 FECSEM, Zeiss, Oberkochen Germany. Before the examination, the samples were prepared by cutting with bench vice and hacksaw followed by gluing on sample holder and finally coated with carbon using Carbon Coater: EMITECH K950X, EM Technologies, Kent, England.

3. Results and Discussions The mechanical properties of particulate–polymer composites depend strongly on the particle size, particle–matrix interface adhesion and particle loading. Particle size has an obvious effect on these mechanical properties. Various trends of the effect of particle loading on composite strength have been observed due to the interplay between these three factors, which cannot always be separated.

Figures 1-3 show the results of the tensile properties for the various samples. It was noticed that the reinforcements bring about improvement in all the tensile properties compared to the unreinforced polyester. The results revealed that steelmaking slag particulates would actually be utilised to improve the tensile properties of unsaturated polyester.

From Figure 1, it was observed that 8 wt% of 300 µm slag particle addition gave the best ultimate tensile strength result with a value of 61.03 MPa compared to the unreinforced polyester material (control) with a value of 50.76 MPa. This implies that the property has been enhanced by 20 %.

Figure 1. Variation of ultimate tensile strength with filler content

for the varied steelmaking slag particles

Figure 2 shows the ultimate tensile strain results for the samples. From the results, it was observed that, 6 wt% from 106 and 300 µm particle sizes gave the best values of 0.028 and 0.027 mm/mm, respectively compared to the unreinforced polyester with a value of 0.016 mm/mm. These culminate to 75 % and 69 % enhancements, respectively.

Figure 2. Variation of ultimate tensile strain with filler content for

the varied steelmaking slag particles

The results of the Young’s modulus in Figure 3

show that 2 wt% of 300 µm particle size reinforced sample gave the best result with a value of 4084.26 MPa compared to the unreinforced polyester matrix which has a value of 3966.15 MPa. This amounts to 2 % increment in this property. The results revealed that 300 µm particle size followed by 106 µm particle size would be the best particle size for tensile property enhancement.

Figure 3. Variation of tensile modulus with filler content for the varied steelmaking slag particles

Figures 4 and 5 show the flexural properties results

from where it was observed that, the reinforcement also brings about enhancement in flexural properties of the composites compared to the unreinforced polyester material. From Figure 4, it was observed that 2 wt% from 106 µm particle size gave the best flexural strength

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 68: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

I.O Oladele: Mechanical Properties of Steel-making Slag Reinforced Polyester Composites 66

at peak with a value of 69.91 MPa compared to the unreinforced polyester sample with a value of 43.25 MPa. This amounts to 62 % enhancement compared to the neat polyester.

Figure 4. Variation of flexural strength at peak with filler content

for the varied steelmaking slag particles

Also from Figure 5, the flexural modulus results

showed that 2 wt% from 106 µm particle size gave the best flexural modulus with a value of 11957 MPa compared to the unreinforced polyester sample with a value of 7452MPa. This amounts to 60 % enhancement compared to the neat polyester. These results show that flexural properties were highly improved with 106 µm particle sizes addition. This was in agreement with the tensile test results where it was noted that 106 µm particle sizes gave the best enhancement in tensile strain property. Tensile strain property is a measure of similar mechanical property in terms of flexural strength properties which is the ability of the materials to be stress for long time before fracture. They both relate to the ductility nature of a material. However, these properties tend to decreases as the filler content increases from 2-8 wt% considering the particle size (106 µm) that gave the best results in flexural property investigation. Figure 5. Variation of flexural modulus with filler content for the

varied steelmaking slag particles

From Figures 6-8, it was observed that as the particle size increases, the number of particles that are present decreases due to weight increase. Figure 6. SEM of Fractured surface of 8 wt % from 75 µm particle

size steelmaking slag filled polyester composites

Figure 7. SEM of Fractured surface of 8 wt % from 106 µm particle size steelmaking slag filled polyester composites

Figure 8. SEM of Fractured surface of 8 wt % from 300 µm particle size steelmaking slag filled polyester composites

Steelmaking Slag particle

Polyester Matrix

Polyester Matrix

Steelmaking Slag Particle

Steelmaking Slag Particle

Polyester Matrix

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 69: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

I.O Oladele: Mechanical Properties of Steel-making Slag Reinforced Polyester Composites 67

The images revealed that there is proper bonding between the steelmaking slag particles (white part) and the unsaturated polyester matrix (dark part) which was responsible for the good mechanical properties that was obtained from the mechanical tests results. As a result of the good interfacial bonding strength between the slag and the polyester matrix, as well as adequate particle dispersal in the polyester matrix, better enhancement of properties was obtained for the composites for both tensile and flexural properties compared to the neat polyester matrix. Despite these good observations with respect to the interaction between the steelmaking slag particles and the unsaturated polyester, 75 µm particle size performed less compared to 106 and 300 µm particle sizes. This may be due to low strength from the fine particles compared to others. 4. Conclusion The desire to ensure that there is proper utilisation of every materials and the drive for zero waste have motivated the use of steelmaking slag as a reinforcement in unsaturated polyester matrix in order to develop polymer based composites for different applications in areas like automobile and structural industries. As a result of the capability of particulate materials to reinforce and bring enhancement in polymer properties, steelmaking slag has also confirmed the suitability of its use as a filler and reinforcement in unsaturated polyester material.

From this work, it was observed that; (a) The mechanical properties; flexural and tensile,

were highly enhanced by the addition of 106 µm particle sizes of the steelmaking slag. While the addition of 2 wt% of this particle sizes gave the best flexural property results, the addition of 6 wt% gave the best tensile strain result.

(b) The use of 300 µm particle sizes of the steelmaking slag led to the enhancement of the tensile properties of unsaturated polyester matrix where it was observed that 2 wt %and 8 wt % gave the best results for modulus and ultimate tensile strength respectively.

(c) The work shows that the addition of these steelmaking slag particles in small quantity, that is, low fibre content is the best for optimum results. This was the case since both properties were highly enhanced by the addition of 2 wt% fibre content compared to others.

References: Bartczak, Z., Argon, A.S., Cohen, R.E, and Weinberg, M. (1999)

“Toughness mechanism in semi-crystalline polymer blends: II. High-density polyethylene toughened with calcium carbonate filler particles”, Polymer, Vol.40, pp.2347-65.

Chaurand, P., Rose, J., Briois, V., Olivi, L., Hazemann, J.L., Proux, O., Domas, J. and Bottero, Y. (2007), “Environmental impacts of steel slag reused in road construction”, Journal of

Hazardous Materials, Vol.B139, pp.537-542 Frias, M. R. and Sanchez, de R. (2004), “Chemical assessment of

EAF slag as construction material”, Cement and Concrete Research, vol 38, issue 10, pp.1881-1888

Fu, S-Y., Feng, X-Q., Lauke, B. and Mai, Y-W. (2008), “Effects of particle size, particle/matrix interface adhesion and particle loading on mechanical properties of particulate–polymer composites”, Composites Part B, Vol.39, pp.933-961.

Fu, S.Y. and Lauke, B. (1997), “Analysis of mechanical properties of injection molded short glassfibre (SGF)/calcite/ABS composites”, Journal of Material Science Technology, Vol.13, pp.389-396.

Fu, S.Y. and Lauke, B. (1998), “Fracture resistance of unfilled and calcite particle filled ABS composites reinforced by short glass fibers (SGF) under impact load”, Composite Part A; Vol.29A, pp.631-641.

Fu, Q. and Wang, G. (1993), “Effect of morphology on brittle–ductile transition of HDPE/CaCO3 blends”, Journal of Applied Polymer Science, Vol.49, ppp.1985-88.

Jancar, J. and Dibenedetto, A.T. (1995), “Failure mechanics in ternary composites of polypropylene with inorganic fillers and elastomer inclusions”, Journal of Materials Science, Vol.30, pp. 2438-45.

Kim, M.H., Park, C.I., Choi, W.M., Lee, J.W., Lim, J.G., and Park, O.O. (2004), “Synthesis and material properties of syndiotactic polystyrene/organophilic clay nanocomposites”, Journal of Applied Polymer Science, Vol.92, pp.2144-50.

Liang, J.Z., Li, R.K.Y and Tjong, S.C. (1997), “Tensile fracture behaviour and morphological analysis of glass bead filled low density polyethylene composites”, Plastics and Rubber Processing and Applications, Vol. 26, pp.278-282.

Motz, H. and Geiseler, J. (2001), “Products of steel slag: An opportunity to save natural resources”, Waste Management Vol.21, pp285-293.

Ou, Y., Yang, F. and Yu, Z.Z. (1998), “A new conception on the toughness of nylon 6/silica nanocomposite prepared via in situ polymerization”, Journal of Polymer Science Part B: Polymer Physics, Vol.36, pp.789-795.

Pukanszky, B. (1995), “Composites”, In: Karger-Kocsis, J., (Ed), Polypropylene: Structure, Blends and Composites, Vol.3. London: Chapman & Hall; p.1-70.

Reynaud, E., Jouen, T., Gauthier, C., Vigier, G. and Varlet, J. (2001), “Nanofillers in polymeric matrix: a study on silica reinforced PA6”, Polymer, Vol.42. No.87, pp.59-68.

Tjong, S.C. and Xu, S.A. (2001), “Ternary polymer composites: PA6,6/maleated SEBS/glass beads”, Journal of Applied Polymer Science, Vol.81, pp.3231-37.

Wang, M., Berry, C., Braden, M. and Bonfield, W. (1998), “Young’s and shear moduli of ceramic particle filled polyethylene”, Journal of Materials Science Materials in Medicine, Vol.9, pp.621-624.

Zhu, Z.K, Yang, Y., Yin, J. and Qi, Z.N. (1999), “Preparation and properties of organosoluble polyimide/silica hybrid materials by sol–gel process”, Journal of Polymer Science, Vol.73, pp.2977-84.

Author’s Biographical Notes Isiaka Oluwole Oladele obtained Masters and PhD in the area of natural fibre reinforced polymer composites in the Department of Metallurgical and Materials Engineering, Federal University of Technology, Akure, Ondo State, Nigeria. Dr. Oladele has supervised many undergraduates and postgraduates research and has published in both local and international journals and conference proceedings in this area. ■

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 70: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

R. Maharaj et al.: Rheological Study of Cement Modified with a Lignin Based Admixture 68

Rheological Study of Cement Modified with a Lignin Based Admixture

Rean Maharaja, Lebert H. Griersonb, Chris Maharajc, Ψ and Vitra Ramjattan-Harryd

a University of Trinidad and Tobago, O’Meara Industrial Estate, O’Meara, Arima, Trinidad and Tobago, West Indies

E-mail: [email protected] b Department of Chemistry, University of the West Indies, St. Augustine, Trinidad and Tobago, West Indies;

E-mail: [email protected] c Department of Mechanical & Manufacturing Engineering, The University of the West Indies, St. Augustine, Trinidad and Tobago,

West Indies; E-mail: [email protected] d University of Trinidad and Tobago, Point Lisas Industrial Estate, Point Lisas, Trinidad and Tobago, West Indies

E-mail: [email protected] Ψ - Corresponding Author

(Received 5 October 2014; Revised 31 December 2014; Accepted 30 January 2015) Abstract: Experiments involving the use of the dynamic shear rheology technique utilising a parallel plate configuration were conducted to investigate changes in the rheological properties of Trinidad Portland cement paste blended with Lignosulfonic acid, acetate sodium salt additive. The rheological properties of plastic viscosity (PV) and yield stress (YS) of the cement blend as defined by Bingham were calculated. Water/cement ratios of 0.40, 0.45 and 0.50 were used with a 0 - 0.50% additive at room temperature and the samples were tested at intervals of 10, 30, 60 and 90 minutes after mixing. The results showed maximum values of the PV between 0.05% and 0.10% admixture concentrations for the various water/cement ratios and time measurements. PV values were generally lower as water/cement ratios increased demonstrating improvements in the rheology. A PV value of 0.7 centipoise obtained with the control sample can be reproduced with the addition of approximately 0.05% admixture using 20% less water. Maximum values of YS generally occur between 0.05% and 0.10% admixture concentrations as a more compact, homogeneous paste system develops. Consistent with previous studies utilising this technique, YS data was generally sporadic. The ability to synthetically alter the rheological properties of Trinidad Portland cement adding a lignin based admixture can serve to optimise the strength, workability and shrinkage due to the reduced water-cement ratio.

Keywords: Portland cement; Rheology, Admixtures, Plastic Viscosity, Yield Stress, Lignin

1. Introduction In recent times there has been a marked increase in the use of mineral admixtures to improve the durability of concrete. Economic considerations mainly due to the necessity for lower cement requirement as well as environmental considerations have also contributed towards the increase in the usage of admixtures (Ferraris et al., 2001). Chemical admixtures are known to improve specific physical and mechanical properties of cement/concrete blends by modifying pore/void structures and enhancing interfacial cohesion between aggregate and cement paste and have been used to improve the quality of the concrete during mixing, transporting, placement and curing (Grierson et al., 2010; Faddi, 2011). However, the influence of admixtures on cement and concrete workability exhibits wide variability attributed to the origin of the raw material used in the cement manufacture, the place of manufacture, the cement composition and the type and concentration of the admixture (Collerpardi et al., 1983; Ben-Dor et al., 1985; Billingham and Coveney, 1993).

2. Literature Review Concrete workability as defined by the American Concrete Institute (ACI) is the ease of placement of concrete, and practically is the ability of a fresh concrete mix to fill the form or mold without reducing the quality of the concrete. Workability is usually measured by the concrete slump test, following the ASTM C 143 or EN 12350-2 test standards. These tests for measuring workability have proven to require a large amount of material and labour and are very expensive (Ferraris et al., 2001). Researchers have successfully employed rheology tests based on the Bingham equation to describe the flow of concrete (Ferraris, 1999; Banfill, 2006; Mitsoulis, 2007; Mukhopadhyay and Jang, 2009). There are two basic parameters that characterise the rheology of the cement and concrete pastes: yield stress and plastic viscosity (Ferraris et al., 2001). As described by Ferraris et al. (2001), the yield stress is related to slump and the plastic viscosity is related to performance attributes such as stickiness, placeability, pumpability and finishability.

ISSN 0511-5728 The West Indian Journal of Engineering

Vol.37, No.2, January 2015, pp.68-73

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 71: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

R. Maharaj et al.: Rheological Study of Cement Modified with a Lignin Based Admixture

69

Lignosulfonates is a common plasticizer used in concrete manufacture where they reduce the amount of water required to make the concrete (giving stronger concrete) while still maintaining the ability of the concrete to flow and be pumped. Lignosulfonates are water-soluble anionic polyelectrolyte polymers obtained as a by-product in the production of wood pulp using sulfite pulping. The Howard process is a commonly used industrial method, in which 90–95% yields calcium lignosulfonates that are produced after precipitation with excess calcium hydroxide. They can also be separated from spent pulping liquids by Ultrafiltration and ion-exchange techniques. Lignosulfonates are also used as grinding aids in the cement mill and as a slurry deflocculant. The characteristic of lignosulfonates to reduce the viscosity of mineral slurries is utilised in oil drilling mud applications (Lebo et al., 2001).

In Trinidad and Tobago, research conducted by Grierson et al. (2004) investigated the influence of the admixture lignosulfonic acid, acetate sodium salt, on the chemical hydration of Trinidad Portland cement. They found that the lignin based admixture functioned as an inhibitor during hydration and as a dispersant, the ultimate effects manifested only during the latter stages of hydration, thereby improving the hardening properties of hydrated cements. Further work conducted by Grierson et al. (2010) measuring the influence of this admixture on the physical and mechanical properties of Trinidad Portland cement found that it altered the properties of apparent porosity (and water absorption), bulk density, compressive strength and setting times at different dosages and hydration times.

Although chemical admixtures are known to improve specific physical and mechanical properties of cement/concrete blends (Ferraris et al., 2001), the ability to formulate dosages of admixtures to improve the workability and ultimate performance qualities of Trinidad Portland cement has not been possible due to the unavailability of comprehensive research information on the low shear rate rheological (flow) properties of the cement-admixture blend. The effect of the admixture on the rheological properties of yield stress and plastic viscosity on Trinidad Portland cement remains unclear and additional information in the literature has proven limited. The ability to synthetically alter the flow properties of Trinidad Portland cement using a lignin based admixture can be very profitable and more sustainable approach since this base material is readily available as a by-product derived from plants.

Therefore, this paper seeks to investigate the rheological properties and in particular, the yield stress (YS) and plastic viscosity (PV) of Trinidad Portland cement blended with varying concentrations of lignosulfonic acid, acetate sodium salt, at different water/cement ratios.

3. Experimental procedures 3.1 The raw materials The cement used was commercial grade ordinary Portland cement manufactured by Trinidad Cement Ltd. The lignosulfonic acid, acetate sodium salt, was purchased from Sigma-Aldrich Company, (catalogue number 370983-10G, lot number 00515CUV). 3.2 Sample Preparation The initial pastes used to prepare samples for the rheological studies were of water/cement ratio of 0.40 by weight and were carried out at room temperature while the additive/cement percentage covered the range 0 - 0.50 %. The mass of the lignosulfonic acid, acetate sodium salt, admixture required was weighed using a Mettler Toledo AL204 analytical balance and dissolved in an amount of water appropriate to maintain the water/cement ratio constant at 0.40. The water containing the admixture and the cement were placed in a 25 ml beaker and mixed for two minutes using a digital IKA (Model RW20D) Overhead Stirrer, at approximately 450 rpm. This procedure was repeated for water cement ratios 0.45 and 0.50. The cement in the stated water/cement ratio is a combination of the mass of the cement and admixture. Table 1 shows the proportion of the constituents for the production of the different cement-additive blends for the different water/cement ratios.

Table 1. The proportion of the constituents for the production of the different cement-additive blends for the different water/cement ratios

% additive

Mass of sample 20 g Water/Cement

ratio 0.40 Water/Cement

ratio 0.45 Water/Cement

ratio 0.50 Mass

of water

(g)

Mass of additive

(g)

Mass of

water (g)

Mass of additive

(g)

Mass of

water (g)

Mass of additive

(g)

0.00 8.00 0.00 9.00 0.00 10.00 0.00 0.05 8.00 0.01 9.00 0.01 10.00 0.01 0.10 8.00 0.02 9.00 0.02 10.00 0.02 0.30 8.00 0.06 9.00 0.06 10.00 0.06 0.50 8.00 0.10 9.00 0.10 10.00 0.10

3.3 Sample Characterisation The rheological properties of plastic viscosity and yield stress of the cement blends as defined by Bingham were determined using an ATS RheoSystems Dynamic Shear Rheometer (Viscoanalyzer DSR) as outlined by Banfill (2006) and Mukhopadhyay and Jang (2009). The DSR test geometry used was the plate–plate configuration (diameters 25mm) with a 1 mm gap (sample thickness) and the measurements were conducted at 25°C. The cement paste to be analysed was placed on the bottom plate using a syringe. The gap was set to 1mm and the excess cement paste removed using a spatula. The shear rate was slowly increased from 0 to 200/s in ten steps (representing the up curve of the hysteresis loop) and

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 72: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

R. Maharaj et al.: Rheological Study of Cement Modified with a Lignin Based Admixture

70

then from 200 to 0/s in 10 steps (representing the down curve of the hysteresis loop) and the corresponding value of the shear stress recorded. The recorded value for each step in the up and down curves represented an average of 5 values measured by the instrument. The various blends containing 0%, 0.05%, 0.10%, 0.30% and 0.50% of the admixture by weight at water/cement ratios 0.40, 0.45 and 0.50 (a total of 15 different blends) were subjected to this analysis. Each of these blends was tested after intervals of 10, 30, 60 and 90 minutes after mixing.

The results obtained were analysed using the Viscoanalyzer software. Consistent with the methodology outlined by Banfill (2006) and Mukhopadhyay and Jang (2009), the values of the rheological parameters associated with the plastic viscosity and yield stress for each sample were determined from the plot of shear rate versus shear stress whereby the plastic viscosity was calculated from the slope of the linear region of the down curve (decreasing shear rate) and the yield stress was calculated from the intercept of the down curve and the y-axis. 4. Results and Discussion Figure 1 shows the hysteresis loop for the blend at 0.40 water cement (w/c) ratio containing 0.10% admixture 10 minutes after mixing and the curve was typical of those obtained for all the cement-lignosulfonic acid, acetate sodium salt, admixture blends analysed in this study. This blend was randomly chosen to highlight consistency with previous work conducted by Banfill (2006) and Mukhopadhyay and Jang (2009). Figure 1. Graph shows the typical up and down curves for the sample

at 0.4 w/c ratio containing 0.1% admixture after 10 minute

As seen in this example and from the equation of

the linearized down line of the hysteresis loop (y = 1.18x + 4.40), the intercept of the down curve with the y-axis was 4.40 Pa and represents the yield stress (YS) of the blend whereas the gradient of the line was calculated to be 1.18 centipoise and represents the plastic viscosity (PV) of the blend.

The rheological parameter of PV as a function of admixture concentration for water/cement ratios 0.40, 0.45 and 0.50 for a time of 10 minutes after mixing is shown in Figure 2.

Figure 2. Plastic Viscosity vs. Admixture Concentration for various water/cement ratios after a mixing time of 10 minutes

The results demonstrate that after 10 minutes a

maximum value of the PV was observed between 0.05% and 0.10% admixture concentrations. This indicates that the cement pastes containing these levels of admixtures demonstrate a relatively higher resistance to flow as it is relatively viscous due to the existence of greater amounts of colloidal solids (Alp et al., 1986). This observation correlates with results from a previous study conducted by Grierson et al. (2010) as they observed that the apparent porosity and the related parameter of water absorption of Trinidad Portland cement containing lignosulfonic acid, acetate sodium salt, at various aging times were generally lower for blends containing 0.05% plasticiser concentration. Grierson et al. (2010) associated these observed reductions at the 0.05% level of admixture with an increase in the inter-particle attraction between the cement particles, thus increasing agglomeration and reducing dispersion or deflocculation giving rise to a more homogeneous, compact paste system of lower porosity. However, at concentrations > 0.10%, they found that the lignin based surfactant sufficiently lowers the surface energy of the water, and thus the surface tension, allowing air entrainment to occur, resulting in higher porosity values.

The observations represented in Figure 2 also show that PV values were generally lower as water/cement ratios increased. Previous work conducted by Ferraris et al. (2001), obtained similar trends when investigating the influence of different water/cement ratios on the rheological properties of cement blends containing ultrafine fly ash. The purpose of their study was primarily to determine the level of water reduction accomplished by using the ultrafine fly ash admixture while maintaining the same YS and/or PV. In other words, the objective was to quantify the reduction of the water requirement due to the addition of ultrafine flyash or in our case lignosulfonic acid, acetate sodium salt, to

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 73: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

R. Maharaj et al.: Rheological Study of Cement Modified with a Lignin Based Admixture

71

the cement paste while maintaining the same YS and PV as the control. As can be seen from Figure 2, a PV value of 0.7 centipoise obtained with the control sample (no admixture added) at a water cement ratio of 0.50 can be obtained with the addition of approximately 0.05% lignosulfonic acid, acetate sodium salt, admixture with a 20% reduction in water requirement (water/cement ratio 0.4). This specific observation demonstrates a distinct advantage (reduction in water requirement) obtained in the rheological properties of the Trinidad cement paste blended with the lignin based additive and offered supporting evidence for previous work documented by Lebo et al., (2001).

The variation of PV with time after mixing for the different additive concentrations at a water/cement ratio of 0.50 is shown in Figure 3. What is most evident from the relationships obtained is that the PV value was highest for the paste containing 0.10% admixture for all the setting times measured. The blend containing 0.05% additive demonstrated a marginal increase in the PV compared to the control sample. As previously explained, this occurrence can be attributed to an increase in the inter-particle attraction and agglomeration thus reducing dispersion or deflocculation which gives rise to a more homogeneous, compact paste system of lower porosity and higher PV.

Figure 3. Plastic Viscosity vs. Time for the various Admixture Concentration at a water/cement ratios of 0.5

The unique rheological behaviour of the blends

containing 0.05% and 0.10% additive was further exemplified in Figure 3 as these two blends demonstrated an initial decrease in their PV after 30 minutes after which the values gradually rise as the cement hydration processes continue and the pastes harden. The pastes containing 0.30% and 0.50% showed an opposite trend as the PV of the pastes maximised after 30 minutes. Although the PV of the paste containing 0.50% admixture subsequently increased after 90 minutes, the paste containing 0.30% admixture demonstrated a continued downward trend in the value of PV. The unique behaviour of the cement paste containing 0.30% admixture was previously observed by

Grierson et al. (2010) where attempted measurements of physical parameters of porosity and bulk density at higher concentrations (>0.1%) of lignosulfonic acid, sodium salt, acetate in Trinidad Portland cement were not possible even after 2 days as the samples were too liquefied in nature.

As explained by Grierson et al. (2010), cement pastes containing high levels of this admixture (>0.1%) result in significant retardation of the hydration process of the cement particles which results in a highly dispersed, low viscosity substance. Studies conducted by Maximilien et al., (1997) and Grierson et al. (2004) associated changes in the hydration times with Ca2 + ion release. Grierson et al. (2004) found that the magnitude of the conductivity change of cement pastes containing 0% and 0.05% plasticiser was similar however the change was greater for the sample with 0.50% plasticizer. They suggested that while low concentrations of the plasticiser have little or no control over the release Ca2+ ions into the solution, blends containing 0.50% admixture or more act to inhibit Ca2

+release. Their observations showed that there is competition between the principal hydration reaction which causes the paste to harden and a plasticiser-induced inhibition of Ca2+ release, with the inhibition process dominating at 0.5% plasticiser concentration. At these high admixture levels, poisoning of the process of crystalline ettringite formation probably occurs which reduces initial C3S hydration resulting in the highly dispersed and liquefied sample with a low PV as observed in this study.

Figure 4 shows the variation of Yield Stress (YS) as a function of Admixture concentration for water/cement ratios 0.40, 0.45 and 0.50 for times after mixing of 10 minutes. The maximum value of YS occurred at 0.05% concentration.

Figure 4. Yield Stress vs. Admixture Concentration for various water/cement ratios after a mixing time of 10 minutes

Figure 5 also shows the variation of Yield Stress

(YS) as a function of Admixture concentration but for times after mixing of 30 minutes. The maximum value of YS occurred at 0.10% concentration. The maximum values of YS occurred within the concentration range of

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 74: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

R. Maharaj et al.: Rheological Study of Cement Modified with a Lignin Based Admixture

72

0.05% to 0.10% for times after mixing of 60 and 90 minutes as well.

Figure 5. Yield Stress vs. Admixture Concentration for various water/cement ratios after a mixing time of 30 minutes

As explained by Grierson et al. (2010), yield stress

is associated with slump and again the higher values observed between 0.05% and 0.10% can be associated with a reduction in the inter-particle attraction, and agglomeration of the cement particles and effecting dispersion or deflocculation. At these low concentrations of admixture, a more compact, homogeneous paste system with higher slump develops. It must be noted however that presentation of additional Yield Stress data proved difficult as these values were generally sporadic. The sporadic nature of the calculated YS data is not unique to this study and has also been reported by other researchers (Nedhi and Rahman, 2004), who associated errors in reliability when estimating yield stress values. 5. Conclusions The addition of lignosulfonic acid, acetate sodium salt, to Trinidad Portland cement resulted in changes in the rheological properties of the blends as demonstrated by changes of both the PV and YS of the modified pastes. The results showed the following: 1) Maximum values of the PV were observed between

0.05 and 0.10% admixture concentrations for the various water/cement ratios and time measurements.

2) PV values were generally lower as water/cement ratios increased demonstrating improvements of the Trinidad cement paste blended with the additive. Addition of lignosulfonic acid, acetate sodium salt, can be used to obtain the same yield stress and PV as the control using significantly less water.

3) Maximum values of YS (directly associated with slump) also occurred between 0.05% and 0.10% addition of lignosulfonic acid, acetate sodium salt, for all water/cement ratios as a more compact, homogeneous paste system develops. Yield Stress data were generally sporadic associated with the reliability of estimating yield stress values.

Water-cement ratios beyond a minimum, generally 0.4, are required for suitable workability. However high water-cement ratios cause decrease in compressive strength and increase in shrinkage of concrete during curing and hence increased cracking. Therefore a compromise between strength, workability and shrinkage is necessary. Organic plasticisers allow better maintenance of workability of cement pastes at lower water-cement ratios with their expected improved hydrated cement properties. There is a drive to develop new improved plasticisers/superplasticisers with even lower workable water-cement ratios, especially those based on quasi-natural materials such as lignosulphonate (a derivative of wood pulp). The workability, i.e. plastic viscosity (PV) and yield stress (YS), results shown for (0.05 – 0.50%) lignosulphonate, sodium salt, acetate cement admixture compositions in this work (and from our previous work) make it a good candidate plasticiser for commercial use in the cement and concrete industry.

Acknowledgements: The authors acknowledge research students Mr. Vivek Rampersad and Ms. Cheryon Morin of the University of the West Indies who assisted with the rheology experiments.

References: Alp, E., Chilingarian, G.V., Caenn,R., Mouhammed Al-Salem,

Uslu, S., Gonzales, S., Dorovi, R.J., Mathur, R.M., and Yen, T.F. (1986), “Drilling Fluid Evaluation Using Yield Point-Plastic Viscosity Correlation”, Energy Sources, Vol. 8 No. 2-3, pp. 233-244

Banfill, P.F.G. (2006), “Rheology of fresh cement and concrete”, Rheology reviews, pp. 61-130

Ben-Dor, L., Wirguin H. C., and Diab, H. (1985), “The effect of ionic polymers on the hydration of C3S”, Cement and Concrete Research, Vol.15, pp.681-686

Billingham, J. and Coveney, P.V. (1993), “Simple chemical clock reactions: Application to cement hydration”, Journal of the Chemical Society, Faraday Transactions, Vol.89, pp.3021-3028.

Collepardi, M., Monosi, S., Moricini, G. and Pauri, M. (1983), “Influence of lignosulphonate, glucose and gluconate on the C3A hydration”, Cement and Concrete Research, Vol.13, pp.568-574.

Faddi, Norbert (2011), Concrete Quality Improvement, VIA University College, p.1-50.

Ferraris, C.F. (1999), “Measurement of the rheological properties of cement paste: A new approach”, RILEM International Symposium on The Role of Admixtures in High Performance Concrete, 21st-26th March, Monterrey pp.333-342.

Ferraris, C.F., Obla, K.H. and Hill, R. (2001). “The influence of mineral admixtures on the rheology of cement paste and cement”, Cement and Concrete Research, Vol.31, No.2, pp.245-255.

Grierson, L.H., Knight, J.C. and Maharaj, R. (2004), “The role of calcium ions and lignosulphonate plasticiser in the hydration of cement”, Cement and Concrete Research, Vol.35, pp.631-636.

Grierson, L., Maharaj, R. and Knight, J.C. (2010), “The effect of lignosulphonic acid, sodium salt, acetate on the physical and mechanical properties of Trinidad Portland cement”, International Journal of Material Science, Vol.5, No.1, pp.75-82.

Lebo, Stuart E. Jr.; Gargulak, Jerry D. and McNally, Timothy J.

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 75: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

R. Maharaj et al.: Rheological Study of Cement Modified with a Lignin Based Admixture

73

(2001), “Lignin”, In: Kirk-Othmer Encyclopedia of Chemical Technology, John Wiley & Sons, Inc.

Maximilien, S., Pera, J. and Chabannet, M. (1997), “Study of the reactivity of clinker by means of the conductometric test”, Cement and Concrete Research, Vol.27, No.1, pp.63-73.

Mathur, R.M. and Yen, T.F. (1986), “Drilling fluid evaluation using yield point-plastic viscosity correlation”, Energy Sources, Vol.8, No.2-3, pp. 233-244

Mitsoulis, Evan (2007), “Flows of viscoplastic materials: Models and Computations”, Rheology Reviews, pp.135-178.

Mukhopadhyay, A.K. and Jang, S. (2009), Using Cement Paste Rheology to Predict Concrete Mix Design Problems, Accessed from http://tti.tamu.edu/documents/0-5820.pdf

Nehdi, M. and Rahman, M.-A. (2004), Effect of Geometry and Surface Friction of Test Accessory on Oscillatory Rheological Properties of Cement Pastes, ACI Materials, No. 101-M47.

Authors’ Biographical Notes: Rean Maharaj is holder of a Bachelors (B.Sc.) Degree in Chemistry, specialising in Analytical Chemistry at The University of the West Indies (UWI), St. Augustine, a Masters of Philosophy (M.Phil.) degree in the field of Applied Physical Chemistry (UWI) and a PhD. Degree in Process and Utilities Engineering from The University of Trinidad and Tobago. Dr. Maharaj spent several years as a Forensic Analyst at the Trinidad and Tobago Forensic Science. He was extensively trained in Forensic Chemistry under a United Nations Fellowship drug training programme and by the US Drug Enforcement Agency (DEA). He is currently Assistant Professor in Process Engineering of The University of Trinidad and Tobago (UTT). His research interests are mainly based in the applied chemistry/materials science area including cement and asphalt technology.

Lebert H. Grierson, former Head of Chemistry Department, is a Lecturer in Physical Chemistry in the Chemistry Department of The University of the West Indies. He holds BSc and PhD from The

University of London. He did postdoctoral work at The Chemistry Department, Rijkuniversitat Groningen, The Netherlands and then at Max Planck Institute for Strahlenchemie (Radiation Chemistry) in Muelheim a.d. Ruhr, Germany between 1988-1991. His present research interest focuses on applied physical chemistry including topic such as waste polymer cracking using fluidised-bed technology, asphalt and cement research and application of microcalorimetry techniques to biological systems.

Chris Maharaj is a Lecturer in the Mechanical and Manufacturing Engineering Department of The University of the West Indies (UWI). He holds BSc and MSc qualifications in Mechanical Engineering and Engineering Management respectively from UWI. He started his career as a Mechanical Engineer in Condition Monitoring and Inspection and worked in the industry for five years. He later went on to pursue his PhD at Imperial College London in Mechanical Engineering. His present teaching and research interests are in alternative use of waste materials, mechanical design optimisation, failure analysis, component life assessment, asset management, and innovation management.

Vitra Ramjattan-Harry is a Laboratory Technician in the Process Engineering Department at The University of Trinidad and Tobago (UTT). She has a B. Eng. in Process and Utilities Engineering and has work experience in both the Food and Chemical Industries where she was involved with product and process optimisation, quality control and ISO certification. At UTT, she assists research, postgraduates and final year students by providing technical support for the analytical equipment in the Material Characterisation laboratory. ■

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 76: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

E.I. Ekwue et al.: Effect of Dynamic and Static Methods of Compaction on Soil Strength 74

Effect of Dynamic and Static Methods of Compaction on Soil Strength

Edwin I. Ekwue a,Ψ, Robert Birchb, and Jared Chewitt c

Department of Mechanical and Manufacturing Engineering, The University of the West Indies, St. Augustine, Trinidad and Tobago,

West Indies a,E-mail: [email protected] bE-mail: [email protected] a,E-mail: [email protected]

Ψ - Corresponding Author (Received 24 September 2014; Revised 7 January 2015; Accepted 30 January 2015)

Abstract: The effect of static (hydraulic press) and dynamic (Proctor) methods of compaction on the strength of soils was investigated in the laboratory. Soil samples of different densities were obtained by incorporating peat into three agricultural soils at 0%, 4%, 8% and 12%, air-dry mass basis. The soils were dynamically compacted using 5, 15 and 25 blows of the Proctor hammer at moisture contents which varied from 5% to 55%, after which bulk density and penetration resistance were measured. The soil was then loosened and repacked to the same bulk densities using static compaction imposed via a hydraulic press and penetration resistance was again measured. Peak strengths of soils achieved from the two compaction methods were compared and the two sets of values were highly correlated (P = 0.001). Results indicate that as long as the same soils are compacted statically or dynamically at the similar moisture contents to same bulk densities, similar strength values are expected. The effect of method of soil compaction on soil strength is not important.

Keywords: Static, dynamic, soil, compaction, strength, impact

1. Introduction Soil compaction refers to the method of mechanically increasing the density of a soil by reducing the volume of air. Soil compaction has direct effect on soil physical properties such as bulk density, strength and porosity; therefore these parameters are normally measured and used to quantify soil compactness (Ohu, 1985). The commonly used laboratory methods for measuring compaction include the impact, static, kneading and vibratory ones (Seed, 1954). The impact test commonly adopted is the Proctor test (Lambe, 1951) while the static laboratory compaction tests involve the use of hydraulic press or pump (Amani et al., 2011; Garcia et al., 2012) to press soils to given bulk densities.

Some researchers (Seed, 1954; Asmani et al., 2011; Crispim et al., 2011) compared the effects of static and impact methods on soil compaction and noted different results. Seed (1954) observed that at equal densities and water contents, soil samples compacted by static pressures exhibited higher stabilities (strengths) than those compacted by Proctor impact methods. Crispim et al. (2011) noted similar results for a silty sandy clay soil they tested. They, however, noted a reverse result for the clayey-silty sand soil in that the strength of the dynamically compacted soil exceeded that of the statically compacted one. Asmani et al. (2011) found that statically compacted soils achieved greater bulk densities and strengths than those compacted with the dynamic Proctor impact test although the Proctor test

impacted greater energies in compacting the soils. Asmani et al. (2001) did not, however, compare the strength of their soils at the same water content and bulk density. It is possible that the soil strength achieved will depend on whether static or dynamic methods are utilised in compacting the soil. These authors only measured few soils with few moisture contents and densities. There is need to perform more comprehensive number of experiments to investigate this hypothesis further.

This paper compares the effect of static (hydraulic press) and dynamic (impact) methods of compaction on the strength of soils compacted at the same moisture content to the same bulk densities. Several soil samples with different bulk densities were obtained using soils amended with organic matter in form of peat. Results will determine whether methods of soil compaction have a significant effect on soil strength and will generally increase the understanding of how the common laboratory methods utilised for compacting soils affect soil strength.

2. Materials and Methods Three soils: Piarco sandy loam, Maracas clay loam and Talparo clay (see Table 1) were selected and used to represent some of the major agricultural soils in Trinidad. They were collected from the 0 to 20 cm depth of the soil profile, air-dried and ground to pass a 5 mm sieve. Particle size distribution was performed

ISSN 0511-5728 The West Indian Journal of Engineering

Vol.37, No.2, January 2015, pp.74-78

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 77: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

E.I. Ekwue et al.: Effect of Dynamic and Static Methods of Compaction on Soil Strength 75

using the hydrometer method (Lambe, 1951). Organic matter content in the samples was measured using the Walkley-Black (1934) method. Organic matter content in the samples was increased by adding air-dry sphagnum peat moss (0.08 Mg m-3 air-dry density) at 4%, 8%, and 12%, air-dry mass basis.

Table 1. Classification, organic matter, and the particle size distribution (%) of the soil

Soil Series Classification*

Organic Matter

Content (%)

Sand (0.06-0.002)

mm

Silt (0.06-0.002)mm

Clay (<0.002)mm

Piarco

Aquoxic Tropudults **

1.7

64.9

17.0

18.1

Maracas

Orthoxic Tropudults

4.7

44.7

24.7

30.6

Talparo

Aquentic Chromuderts

2.7

25.4

28.3

46.3

* - Classification according to the Soil Taxonomy System (soil survey Staff, 1999).

** - All values are means of three replicates

To determine the bulk density and strength of the soils after compaction, two replicate soil samples were compacted using the Proctor method (Lambe, 1951). Compaction was carried out at different moisture contents (varying from 5 to 55%) using 5, 15 and 25 Proctor hammer blows applied in three layers on soils put in cylindrical moulds of 10.2 cm diameter and 11.8 cm height. The moisture contents for compacting the soils were chosen according to the soil consistency limits, which were determined from an earlier experiment by Ekwue and Stone (1995).

After compaction at given moisture content, the mould with the compacted soil was weighed to determine the bulk density. Soil strength was measured on the samples using penetration tests conducted using a hand-pushed spring-type Proctor penetrometer (ASTM, 1985). The soil was then removed from the mould, loosened and then repacked into the same Proctor mould using a hydraulic press configured to facilitate the compression of soil. A flat circular metal plate with 2 mm clearance which allowed it to fit into the 10.2 cm diameter Proctor mould, was used as the interface between the piston arm of the hydraulic press and the soil. The compression process was carried out until the same bulk density of the Proctor test was gained. This was achieved by gauging the depth of the piston as it descended into the mould.

The same penetrometer was again used to measure soil strength. The dynamic (impact) Proctor and the static (hydraulic) compaction methods were then continued for the other moisture contents following the same procedure each time. The optimum moisture contents and the maximum densities from the Proctor test for each soil, as well as the peak resistance from both methods of compaction were noted for each soil.

3. Results and Discussion Figure 1 shows the bulk density-moisture content plots for the three soils each with four peat contents and compacted with 25 Proctor blows. The nature of the graphs follows typical soil behaviour (De Kimpe et al.,1982; Ohu et al., 1985) The plots for the 5 and 15 Proctor blows were similar except that the values of the maximum bulk density and the optimum moisture contents were different as shown in Table 2. Figure 1. Bulk density and moisture content for the three soils at

the compaction level of 25 Proctor blows

As was expected, the bulk densities declined, while

the optimum moisture contents increased with increasing peat contents from 0% to 12%. This was attributed to the lower bulk density of the peat incorporated into the

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 78: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

E.I. Ekwue et al.: Effect of Dynamic and Static Methods of Compaction on Soil Strength 76

soils. This occurred for all the three soils and all the three compaction levels. Maximum bulk densities increased while the optimum moisture contents declined with increasing levels of soil compaction.

Figure 2 shows the penetration resistance-moisture content plots of the three soils using the 25 blows of the Proctor method compared with the same plots obtained using the hydraulic press to compact the soils at the same moisture contents to the same bulk densities

achieved during Proctor test. The nature of the graphs for the two compaction methods followed typical soil behaviour fully described by Ekwue and Stone (1995) and were similar in values and shape. Generally as was obtained in previous studies by Ohu (1985) and Ekwue and Stone (1995), for each soil and compaction method, peak penetration resistance occurred at lower moisture contents when compared with the optimum moisture contents for maximum compaction.

Table 2. Values of maximum bulk density, peak penetration resistance, Tmax (MPa) and the moisture contents at which they occurred for the soils with peat content and compacted using Proctor blows and hydraulic press

Soil type

Proctor

compaction blows

Peat content (%) 0 4 8 12

ρmax

Tmax (Proctor)

Tmax (Hydraulic

press) ρmax

Tmax

(Proctor)

Tmax (Hydraulic

press) ρmax

Tmax

(Proctor)

Tmax (Hydraulic

press) ρmax

Tmax

(Proctor)

Tmax (Hydraulic

press) Piarco 5 1.6/20a 3.0/13 4.0/12 1.3/31 2.9/20 3.6/18 1.1/44 2.7/30 2.8/30 1.0/34 2.1/31 2.7/34 sandy loam 15 1.8/16 7.5/11 8.3/11 1.4/27 3.5/16 4.4/16 1.3/37 3.6/26 3.9/27 1.1/32 2.8/30 3.0/30

25 1.8/14 9.0/10 10.0/10 1.5/25 8.3/15 7.2/15 1.3/30 6.0/25 5.0/25 1.1/30 5.7/28 4.8/30 Maracas 5 1.6/32 8.0/21 8.8/21 1.3/37 2.1/29 2.1/27 1.2/45 2.2/30 2.1/29 1.1/58 1.9/32 1.9/32

clay loam 15 1.7/26 8.5/20 9.3/20 1.4/30 5.9/26 5.0/25 1.2/39 4.6/28 4.8/28 1.1/42 3.6/31 1.2/31

25 1.6/25 12.0/20 11.0/20 1.4/29 6.0/25 7.0/25 1.3/36 5.6/26 5.0/26 1.2/40 4.2/30 4.9/30 Talparo 5 1.2/30 5.1/23 4.9/25 1.2/32 4.0/28 4.2/30 1.1/38 3.4/30 4.0/31 1.0/52 1.4/37 1.4/35

clay 15 1.4/27 6.3/21 6.0/21 1.4/30 4.9/27 5.0/27 1.2/36 4.2/28 4.4/30 1.1/43 3.3/36 3.3/33 25 1.5/26 7.4/20 7.2/20 1.3/28 6.7/25 6.8/25 1.2/35 5.5/27 6.0/30 1.2/41 4.2/35 4.6/32

a Maximum values and the moisture content at which they occurred. ρmax is maximum bulk density (Mg m-3)

Figure 2. Penetration resistance and moisture contents of the soils compacted (a) with 25 blows with the Proctor Method and (b) at the equivalent bulk density using the hydraulic press

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 79: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

E.I. Ekwue et al.: Effect of Dynamic and Static Methods of Compaction on Soil Strength 77

Table 2 details the values of peak penetration resistance and the moisture contents at which they occurred for the three soils at three compaction levels using the Proctor and the hydraulic press compaction methods. The trend in peak penetration resistance values was the same as was observed for bulk density in that penetration resistance declined with increasing peat content but increased with increasing compaction effort. Values of peak penetration resistance from the two compaction methods were close to each other (see Table 2) and were highly correlated, P = 0.001 (see Figure 3). Figure 3. Comparison of peak penetration resistance (MPa) using

the two compaction methods

The correlation coefficient (0.97) is close to 1.00

representing an almost perfect correlation. The value of the slope of regression line (0.94) and the intercept (0.39) were significantly close to 1.00 and 0.00 respectively, which showed that there was little or no bias in the prediction by the equation. As stated in the materials and methods section, the soils were compacted at the same water content to the same bulk densities as was obtained from the Proctor compaction test. Results show that once the soils were compacted at the same water content and bulk density, values of peak soil strength were almost the same irrespective of the method of compaction. This disagrees with the results of Seed (1954) which showed that statically compacted soils are more easily compacted to higher strength values than dynamically compacted ones.

Crispin et al. (2011) also observed the same trend for the silty sandy clay soil they studied. They observed that interparticle forces are destroyed by dynamic compaction producing structures with lower strength. Asmani et al. (2011) attributed this soil behaviour to the non-uniformity of the dynamically compacted soils using the Proctor test. They stated that since soils are compacted in three layers using the Proctor test, the bottom layer will normally have greater densities than the middle and the upper soil layers in the mould.

Crispin et al. (2011), however, observed that dynamic compaction produced greater strength than static compaction for the clayey silty sand soil they studied. These authors used separate soil samples to perform the static and dynamic compaction of their soils and this may have contributed to the results they obtained.

In the present study, the same soils were compacted in the Proctor mold using the two methods of compaction. Moreover, these previous authors did limited soil tests. For instance, Crispin et al. (2011) only tested two soils each at less than 3% optimum, optimum, and 2% above optimum moisture contents. A more comprehensive soil test programme using three soils, four peat contents, three compaction efforts and several water contents were adopted in the present study.

4. Conclusion This paper has demonstrated that the method of soil compaction utilised to achieve soil compaction does not affect soil strength as long as samples are compacted at the same moisture contents to same bulk densities.

References: Asmani, D., Hafez, M.A. and Nurbaya, S. (2011), “Static

laboratory compaction method”, Electronic Journal of Geotechnical Engineering, Vol.16, pp.1653-1663.

ASTM (1985), “Standard test method for moisture content penetration resistance relationships of fine grained soils”, Annual Book of American Society for Testing and Materials, 04.08, 289-292.

Crispin, F.A., de Lima, D.C., Schaefer, C.E., Silva, C.H., de Carvalho, C.A., Barbosa, P.S. and Brandao, E.H. (2011), “The influence of laboratory compaction methods on soil structure: Mechanical and micromorphological analyses”, Soils and Rocks, Vol.34, pp.91-98.

De Kimpe C R; Bernier-Cardou M; Jolicoeur P (1982). Compaction and settling of Quebec soils in relation to their soil-water properties. Canadian Journal of Soil Science, 62, 165 – 175.

Ekwue, E.I and Stone, R.J. (1995), “Organic matter effects on the strength of compacted agricultural soils”, Transactions of the ASAE, Vol.38, pp.357-365.

Garcia, C., Alemany, E. and Bautista, I. (2013), “Relationship among compaction, moisture condition and penetration resistance in horticultural soil”, International Conference of Agricultural Engineering, CIGR-Ageng, Valencia, Spain. Accessed 31 December 2013, from http://cigr.ageng2012.org/images/fotosg/tabla_137_C0913.pdf.

Lambe, T.W. (1951), Soil Testing for Engineers, John Wiley, New York

Ohu, J.O. (1985), Peatmoss Influence on Strength, Hydraulic Characteristics and Crop Production of Compacted Soils, Ph.D Thesis, McGill University, Montreal, Canada.

Ohu, J.O., Raghavan, G.S.V. and Mckyes, E. (1985), “Peatmoss effect on the physical and hydraulic characteristics of compacted soils”, Transactions of the ASAE, Vol.28, pp.420-424.

Seed, H.B. (1954), “Stability and swell pressure characteristics of compacted clays”, Clays and Clay Minerals, Vol.3, pp.483-501.

Soil Survey Staff (1999), Soil Taxonomy: A Basic System for Making and Interpreting Soil Surveys, In: Agriculture Handbook 436, 2nd Edition, US Government Printing Office, USDA, Washington DC, p.128-129.

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 80: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

E.I. Ekwue et al.: Effect of Dynamic and Static Methods of Compaction on Soil Strength 78

Walkley, A. and Black, I.A. (1934), “An examination of the effect of Degtjareff method for determining soil organic matter and a proposed modification of the chromic acid titration method”, Soil Science, Vol.37, pp.29-38.

Authors’ Biographical Notes: Edwin I. Ekwue is the Head of the Department of Mechanical and Manufacturing Engineering and Professor in charge of the Biosystems Engineering program at The University of the West Indies, St Augustine, Trinidad and Tobago. He is a member of the Editorial Board of the West Indian Journal of Engineering. His specialties are in Water Resources, Hydrology, Soil and Water Conservation and Irrigation. His subsidiary areas of specialisation are Structures and Environment, Solid and Soil Mechanics, where he has teaching capabilities. Professor Ekwue has published widely. He had served as the Deputy Dean (Undergraduate Student Affairs), the Deputy Dean (Post-graduate Affairs and Outreach), the Chairman of Continuing Education Committee, and the Manager of the Engineering Institute in the

Faculty of Engineering at UWI.

Robert Birch is an Instructor in the Department of Mechanical and Manufacturing Engineering at The University of the West Indies, St Augustine, Trinidad and Tobago. He is a registered Professional Engineer (R.Eng) and Project Management Professional (PMP) with over sixteen years of industrial and teaching experience. He has a BSc. (Eng) and MPhil in Agricultural Engineering from The University of the West Indies and is presently pursuing a PhD in Mechanical Engineering. Mr. Birch is a member of the Institution of Agricultural Engineers (UK). His interests are in Field Machinery and Heavy Equipment Design, Fluid Power Technology and Soil-Machine interaction.

Jared Chewitt is a graduate of the Faculty of Engineering, The University of the West Indies in 2013 majoring in Mechanical Engineering with Biosystems. ■

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 81: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

KF. Pun: A Strategic Initiative on Enhancing Postgraduate Throughputs at The UWI St. Augustine Campus 79

A Strategic Initiative on Enhancing Postgraduate Throughputs at The UWI St. Augustine Campus

Kit Fai Pun

Department of Mechanical and Manufacturing Engineering, The University of the West Indies, St. Augustine, Trinidad and Tobago,

West Indies; E-mail: [email protected]

(Received 31 December 2014; Revised 20 January 2015; Accepted 30 January 2015) Abstract: In the past five years, The University of the West Indies (UWI) has been sustaining a strong growth in postgraduate (PG) enrolments in taught and research programmes. However, the throughput rate of PG could not match with the increase in PG enrolment in both relative and absolute terms. The UWI’s Strategic Plan 2012-2017 centres around six (6) integrated Perspectives. One of the three strategic goals under the Research and Innovation (R&I) perspective is to ‘enhance graduate studies and increase postgraduate research output’. In such context, this paper reviews the key areas of priority for the Campus for the period 2014-2017, and informs the strategic initiative with a proposed Throughput Enhancement Project (TEP) at The UWI St Augustine Campus. It then presents the structure of TEP and a schedule of its implementation. For facilitating the TEP initiative, project leaders and process owners are identified, and resource requirements versus savings are explored. The paper concludes by discussing the evaluation of the TEP efficacy in relation to achieving the R&I strategic goals of The UWI.

Keywords: Graduate intakes, throughputs, enrollment, postgraduate programmes, university 1. Introduction The University of the West Indies (UWI) was established in 1948 as a University College of the University of London becoming an independent university in 1962. UWI has four Campuses: Mona (Jamaica), St Augustine (T&T), Cave Hill (Barbados) and Open Campus, and has been supported financially by 16 countries in the Caribbean. It has an enrolment of some 50,000 students in 2013/2014, and graduates some 10,000 per annum. The UWI Mission Statement is “to advance education and create knowledge through excellence in teaching, research, innovation, public service, intellectual leadership and outreach in order to support the inclusive (social, economic, political, cultural, environmental) development of the Caribbean region and beyond” (UWI, 2014).

At The UWI St Augustine Campus, research and teaching influence each other in a symbiotic way. Outlining a robust research mandate, providing a supportive environment for creativity and innovation, and facilitating effective knowledge transfer can only serve to complement and enrich the teaching and learning experience for students both at the undergraduate (UG) and postgraduate (PG) levels. In the past five years, The UWI St Augustine Campus has been sustaining a strong growth in postgraduate (PG) enrolments in taught and research programmes. However, the throughput rate of PG could not match with the increase in PG enrolment in both relative and absolute terms (Pun, 2014a, 2014b).

Improving throughput means different things to different people. For instance, there is a view that

improving throughput can be adequately achieved by measures that do not require changes in current teaching-and learning practices, such as reformulating admissions policy to exclude disadvantaged students, or various manipulations of performance data. The cynical view is sometimes expressed that higher throughput targets can be met simply by lowering academic standards (UCT, 2012; Pun 2014c). This paper discusses the challenges of and explores the need on improving the PG throughputs. An initiative on enhancing PG throughputs is proposed in relation to achieving the strategic goals of research and innovation (R&I) of The UWI. 2. The UWI’s Strategic Goals and SGSR The University’s Strategic Plan 2012-2017 centres around six (6) integrated Perspectives. One of them is ‘Research and Innovation’ that stresses three (3) strategic goals: 1) to create an enabling environment to support, foster and increase the output of high quality research and innovation, 2) to enhance graduate studies and increase postgraduate research output, and 3) to increase funding and strengthen research partnership (UWI, 2014). Table 1 depicts the associated sub-objectives of respective goals under the ‘Research and Innovation’ perspective. These strategic goals and associated sub-objectives are in nature interrelated.

At The UWI St Augustine Campus, The School for Graduate Studies and Research (SGSR) is an operational entity within the university context and has been providing administrative services to various faculties, departments and teaching units (such as Schools, Institutes, Centres, and the like) on all kinds of matters

ISSN 0511-5728 The West Indian Journal of Engineering

Vol.37, No.2, January 2015, pp.79-84

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 82: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

K.F. Pun: A Strategic Initiative on Enhancing Postgraduate Throughputs at The UWI St. Augustine Campus

80

Table 1. Associated Sub-objectives of University Strategic Goals Strategic goals Strategic objectives Faculty-led research and innovation

1) Develop and implement supportive policies, processes and incentives for research; 2) Promote research accomplishments locally and internationally; 3) Increase the number of peer-reviewed publications and citations, and 4) Develop market products based upon cutting-edge research.

Graduate Studies and Research

1) Increase enrolment of full-time graduate research students; 2) Improve throughput of research students; 3) Strengthen supervision and other support systems and policies, and 4) Establish and implement mechanisms for measuring output and disseminating student research.

Increasing funding and strengthening research partnerships

1) Rationalise and enhance support for the development of research proposals, implementation and management of research grants;

2) Explore and increase donor funding for research and innovation, and 3) Expand the range of strategic private and public sector partnerships, locally, regionally and internationally.

Source: UWI (2014)

related to graduate studies and research. It has been working with the Office of the Graduate Studies and Research (OGSR) towards aligning the core activities with the university’s strategic initiatives, as related specifically to the responsibilities of the unit, with recruitment, enrolment and throughput being prime targets (Pun, 2013).

Under the auspices of the Campus Principal and the Campus Registrar as well as the support from both the Campus Committee for Graduate Studies and Research (CCGSR) and the Campus Research and Publication Fund (CR&P) Committee, the SGSR and OGSR have been in 2012-2014:

1) Providing administrative support to the increasing enrolments of graduate students;

2) Facilitating the review and approval of new programmes proposed; and

3) Facilitating the approval and provision of financial support to staff members and students engaging in research.

The SGSR had identified several key priorities while working with academic faculties and units across campus. Among others, two priorities are to expand the enrolment of full-time graduate research students, and to improve throughputs of research students. 3. Mismatch of PG Enrolments versus Throughputs

The UWI St Augustine campus has been sustaining a strong growth in PG enrolments with good joint efforts from faculties and supporting units in designing and promoting suitable programmes that meet the study needs of current and prospective students. There has been a continuous trend of increase in the graduate applications for admissions, and enrolments in both taught and research graduate students over the past years. Statistics show that there were some 6,205 enrolments of new and returning graduate students among faculties in 2013/14 as compared to 4,991 in 2012/13 (see Table 2).

Table 3 shows a comparison of distribution of graduate enrolments from 2008/09 to 2013/14, with an attempt to separate the enrolments between taught and research programmes at the St Augustine campus. Students enrolled in the Taught PG/Masters programmes made up about 85.4% of the graduate population (i.e., some 5,301 students) among various faculties in 2013/2014. The rest, i.e., 904 MPhil/PhD students were being enrolled in 2013/14. As compared with that of 2012/13, there was a growth of 16.6% for research programmes and 25.7% for taught programmes, respectively (Pun, 2014b). It shows a significant growth of 24.3% on total enrolments, representing a growth of 25.7% for taught programmes and16.6% for research programmes, respectively (Pun, 2014b, 2014c).

Table 2. Enrolments of New and Returning Graduate Students Amongst Faculties, 2008/09 - 2013/14 Faculty (including Centres) 2008/09 2009/10 2010/11 2011/12 2012/13 2013/14* Engineering 873 848 907 1,023 1,005 1,248 Food and Agriculture n.a. n.a. n.a. 0 184 254 Gender and Dev Studies 0 17 22 19 28 29 Humanities and Education 585 744 843 858 872 1,038 Medical Sciences 283 263 337 373 399 500 Science and Agriculture 362 421 505 558 0 0 Science and Technology + n.a. n.a. n.a. 2 420 553 Seismic Research Centre 0 0 3 2 2 4 Social Sciences 1,372 1,442 1,757 2,057 2,081 2,579 TOTAL 3,475 3,735 4,374 4,892 4,991 6,205

Remarks: * - Available figures as at May 20th, 2014; n.a. – not available + - The University took the decision to separate the Faculty of Science and Agriculture into two new Faculties: Food and Agriculture, and Science

and Technology, starting from August 2012.

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 83: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

K.F. Pun: A Strategic Initiative on Enhancing Postgraduate Throughputs at The UWI St. Augustine Campus

81

Table 3. Distribution of Taught and Research Graduate Students Enrolled at UWI, St Augustine 2008/09 to 2013/14 Graduate Students

2008/09 2009/10 2010/11 2011/12 2012/13 2013/14* % Growth in 2013/14 compared with 2012/13

Research 802 645 722 766 775 904 16.6% Taught 2,845 3,171 3,652 4,126 4,218 5,301 25.7% Total: 3,647 3,816 4,374 4,892 4,991 6,205 24.3%

Remarks: * - Available figures as at 20 May, 2014.

Table 4. Graduands of Taught Masters and PG Diplomas by Faculties, 2008/09 - 2013/14 Faculty 2008/09 2009/10 2010/11 2011/12 2012/13 2013/14* Engineering 138 165 163 167 175 194 Food and Agriculture n.a. n.a. n.a. n.a. 22 28 Humanities and Education 182 155 375 314 360 383 Medical Sciences 49 18 99 86 116 116 Science and Agriculture 11 33 58 60 0 0 Science and Technology n.a. n.a. n.a. n.a. 67 73 Social Sciences 330 381 376 417 522 515 Total: 710 752 1,071 1,044 1,262 1,332

Remarks: * - Available figures at November1st, 2014; n.a. – Not available.

Despite an increase in the enrolment of graduate

research students being positive in the anticipated track since 2009/10, the throughput rate of PG students (in both taught and research programmes) has still been lagging behind. For 2013/2014, a total of 1,332 awards of Masters, Certificates and Diplomas were granted to PG graduands. The figure of 2012/13 was 1,262 (see Table 4). Besides, 46 MPhil/PhD graduands were produced in 2013/2014. There has been a 5.65% increase in the numbers of PG graduands from 1,262 in 2012/2013 to 1,332 in 2013/2014. Nevertheless, the growth of PG throughputs could not match with the increase in PG enrolment in both relative and absolute terms. There has been a gap between student enrolments and student throughputs.

There have been many factors and/or causes attributing to the mismatch of PG enrolments versus throughputs over the years. The increase in PG enrolment would imply that there would have increased the demand on teaching, human and institutional resources at various fronts. Many faculties, department and academic units would have been encountering serious constraints in utilising and expanding their teaching, human and institutional resources (including assignment/allocation of teaching versus student supervision responsibilities to faculty members, the provision of infrastructural (classroom, laboratory) supports, and the increase/improvement in student/administrative services, etc). In other words, this would certainly add further burden on the existing capability in respective faculties, department and academic units (Pun, 2014a).

The persistence of unsatisfactory throughput indicates that the challenge is substantial at St Augustine Campus. How to close/narrow the throughput gap and enhance the throughput rate of PG programmes had been recognised as one of the top priorities among others for

the University to address, particularly the taught programmes which attributed to majority of PG student population. 4. Needs for Enhancing PG Throughputs At UWI St Augustine Campus, most of the existing PG taught Master’s have research papers/projects components (particularly those individual projects/papers) and require structured and/or intense supervision. Difficulties in finding/assigning supervisors and over-burdening the existing supervisors have been the problems that affected adversely the PG throughputs. Moreover, supervision of these research papers/projects (particularly those individual ones) would be very resource-demanded in terms of supervisors’ (faculty members’) time. There is also a pressing need to refine and expand the design and delivery of many PG taught programmes so as to meet flexibly the changing needs from stakeholders (including the students, the employers/industry/government, and the accreditation bodies, etc).

From the students’ prospective, as shared from many PG students, they claimed that lacking of sufficient research motivation and supervisor guidance (or supervision) would affect significantly their studies and throughputs at St Augustine Campus. However, one major bottleneck was to fulfill the requirement of undertaking a research project/paper/practicum before they could graduate in respective taught PG Master’s programmes being offered at the St Augustine Campus. Statistics show that many students had stayed in the system for one reason and another. However, a great large number could not manage in completing the research project/paper/practicum in the designated timeline despite that they could have completed all their taught courses for graduation. As a consequence, many students would have been staying in the system for many

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 84: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

K.F. Pun: A Strategic Initiative on Enhancing Postgraduate Throughputs at The UWI St. Augustine Campus

82

years. Most of these PG students have their commitments from work and family and/or other personal goals. In some cases, it seemed that there have no ways out for them to complete the research project/paper/practicum component (Pun, 2014a, 2014c). 5. A Throughput Enhancement Project/Initiative In May-June 2014, a working group of Research and Innovation and Graduate Studies (RIGS) reviewed the decisions taken at the Campus Management Retreat in May 2014, on the key areas of priority for the Campus for the period 2014-2017. The group had recognised the pressing need to improving throughput, and recommended that such improvements must not take precedence over equity of access and outcomes. Enhancing the throughput rate would stress the improvement in the effectiveness of the teaching and learning process for the diverse student body, and there would be no compromising standards (Pun, 2014a). 5.1 The TEP Proposal The RIGS group initiated a draft proposal of ‘Throughput Enhancement Project’ (TEP) for Taught PG Programme. The project explored the possibility of restructuring the design of some postgraduate (such as taught Master’s) programmes that would allow students to opt to taking more taught courses in lieu of the credit requirements of research project/papers in respective programmes.

The credit requirements of a taught Master’s programme vary from one faculty to another faculty and from one department to another department. For instance, for a typical 45-credit programme, it would normally comprise of taught courses (i.e., 33-36 credits) and a research project/paper/practicum (i.e., 9-12 credits). Under the proposed TEP, a PG student enrolled in the said programme would be allowed to take a special option of obtaining 9-12 credits from the enrolment of additional 3-4 courses in lieu of the credit requirements of research project/paper, subject to the recommendations from the programme coordinator and approval from the Head of Department as well as the availability of additional courses being offered in respective programme.

The Faculty/Department is to assume the responsibility of 1) assessing the eligibility/suitability of students to take this option and 2) assuring the availability of additional courses that are relevant to the discipline and/or specialised/associated subject areas of respective programme that the student is being enrolled.

Moreover, several key performance indicators are identified to determine quantitative/qualitative impact of the TEP initiative (Pun, 2014c). These are 1) the number of PG programmes revised/modified to meet the students’ and stakeholders’ needs, 2) the increase in the number of graduands per programme and per academic year, and 3) the increase in staff members’ teaching and

research productivity (in terms of number of students supervised successfully, and number of research publications generated, etc).

4.2 Structure of TEP The TEP is to be designed as a campus-wide improvement initiative (Pun, 2014c). The focal area of the project is to allow individual faculties, departments and academic units to modify, revise and/or refine the design, offer and delivery of their existing PG courses and programmes that suit better for the needs of the students and stakeholders, in line with the existing university’s regulations and guidelines. Individual faculties, departments, institutes, academic /research centres, and units, are responsible for the revision, implementation, monitoring and maintenance of their TEP initiative and associated programmes at their levels (via their own coordinator(s)), and also to update the progress and accomplishments of respective initiatives to respective faculties and the SGSR/OGSR.

There has been a structured route of responsibilities in relation to the implementation of the proposed TEP project. It is anticipated that individual departments, institutes, academic /research centres, units and the like, would 1) identify the PG courses and programmes to be enhanced and prepare their own ‘TEP Implementation Plans’ for the proposed changes and revisions, and 2) appoint their own coordinator(s) or a designated person(s), (being their project manager) to coordinate the implementation. Any revisions and/or changes in PG courses and programmes could be submitted separately via the Faculty, or be incorporated into the ‘TEP Implementation Plan’ of their respective faculty, for possible consideration and approval from the Board for Graduate Studies and Research (BGSR). Similarly, it is expected that individual faculties would appoint their own coordinator(s)/project manager(s) (such as the Deputy Deans, GSR) to coordinate the implementation of any revisions and/or changes of respective PG courses and programmes at the faculty level. 4.3 Schedule of TEP Implementation The TEP Implementation Plans in respective faculties would be consolidated and then absorbed into the TEP Implementation Plan for the Campus. In essence, the revision(s) and change(s) of relevant programmes would be undergone through the normal submissions to the Board for Graduate Studies for approvals via the respective Faculty Boards and the CCGSR. The SGSR and OGSR would facilitate the approval process in a timely manner with respect to the scheduled meetings of the BGSR.

Upon approval, the actual implementation of modified PG programmes would fall under the responsibilities of individual faculties, departments, institutes, academic/research centres, and units. The Deans of respective faculties and the Heads or Directors

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 85: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

K.F. Pun: A Strategic Initiative on Enhancing Postgraduate Throughputs at The UWI St. Augustine Campus

83

of respective departments, school, institutes, centres and/or units are to assume the role as the project leaders. The Campus Coordinator of GSR is to assume the role as the project manager responsible for overseeing the coordination/ implementation of TEP initiative and associated programmes that are proposed, developed and consolidated at the campus level. Therefore, the Campus Coordinator GSR (and extending to SGSR/OGSR and CCGSR) is to provide facilitative supports on the issues and matters associated with TEP.

The TEP initiative is on-going in nature, and its implementation could adopt the ‘Plan-Do-Check-Act’ cycle (Pun, 2014c). Individual component of the cycle would take 6-9 months to complete. It is expected that each cycle would span for 2-3 academic years. In other words, the first cycle would start from 2014-2015 and then end by 2016-2017. Nevertheless, the pace of TEP implementation in individual faculties, departments, institutes, and other academic /research centres may vary. 4.4 Project Leaders and Process Owners Individual faculties, departments, institutes, academic/ research centres, and units, would be responsible for the revision, implementation, monitoring and maintenance of their TEP initiatives at their levels, via their own coordinator(s) who are to assume the role of process owners and update the progress and accomplishments of respective initiatives to the Campus Coordinator GSR. The Deans of respective faculties and the Heads or Directors of respective departments, school, institutes, centres and/or units would be to assume the role of project leaders.

The Campus Coordinator of GSR (and extending to SGSR/OSGR and CCGSR) would provide facilitative supports on the coordination/ implementation of TEP initiative and associated programmes that are proposed, developed and consolidated at the campus level. In other words, this is a partnership relationship between SGSR and faculties/departments/units in 1) preparing the TEP Implementation Plans at various levels (i.e., Campus, Faculty, and Departments/ Institutes/ Units), 2) fostering the execution of TEP initiative and programmes in these plans, and 3) monitoring and measuring the performance of respective TEP initiative and programmes at various levels (Pun, 2014a). 4.5 Resource Requirements versus Savings Any programme refinements, changes and expansions should have resource implications in terms of the support/provision of proper institutional and departmental (including infrastructural and administrative) supports across various faculties and departments on campus. It is proposed that individual faculties, departments, institutes, academic/research centres, units and the like, would be responsible for working out their budget for resource requirements (e.g., human, infrastructural, etc). Nevertheless, the resource

savings from undertaking the TEP initiative would be substantial as a result of enhanced throughput rates, in terms of better utilisation of existing human and infrastructural resources at individual faculties and departments/institutes/units.

The SGSR/OGSR would work closely with Faculties, and Departments/ Institutes/ Units on this venture. Hence, resource requirements at both SGSR and OGSR would be absorbed in their annual operational budgets for the coming years. 5. Discussions and Conclusion Recent research in educational development suggests that many factors affecting students’ throughput rates. These include flexible curriculum structures, encouraging student engagement through varied learning opportunities and a range of teaching styles, curriculum alignment, the embedding of academic literacies in curricula, enhanced effective support and social connectedness, as well as staff and tutor development (Biggs and Tang, 2009; Lubben et al., 2010; UCT, 2012; Pun, 2014c). It has been argued that student learning could be improved through concerted action that would foster the teaching and learning process.

Providing a supportive environment for creativity and innovation, and facilitating effective knowledge transfer can only serve to complement and enrich the teaching and learning experience for students both at the undergraduate and postgraduate levels. This paper focuses on addressing the issues associated with improving the throughputs of taught Master’s programmes, and explores the introduction of the proposed TEP project at St Augustine Campus for the academic year of 2014-2017 aligning the current University’s Strategic Plan.

There is evidence internationally, for instance, some Commonwealth Universities (e.g., City University of Hong Kong and The Hong Kong Polytechnic University) had since the mid-1990s undertaken similar projects as TEP. The successful experiences shared that a majority of students (i.e., some 70-90%) enrolled in their taught PG/Master’s programmes would have completed their studies by taking extra credits in relevant/practical courses in lieu of undertaking the research project/paper component. The remaining 10-30% of students who were of the research abilities would have been channeled to take the research project/paper option to complete their studies (CityUHK, 2014; HKPolyU, 2014). As a result, these projects had significantly enhanced the throughput rates of students enrolled in their taught PG/Master’s programmes and also strengthened strategically both teaching and research productivity of faculty members who would have engaged in teaching courses and supervising students in these programmes.

Drawn from the successful experience/lessons from abroad, it is anticipated that as a substantial amount of PG students at St Augustine would likely opt to taking

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 86: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

K.F. Pun: A Strategic Initiative on Enhancing Postgraduate Throughputs at The UWI St. Augustine Campus

84

extra courses to complete their studies. The saving of faculty members’ time on supervising students would be huge as compared to the situation where students need to complete their respective research projects/papers/ practicums to graduate. Besides, faculty members who engage in teaching courses and supervising students in these programmes could improve their teaching efficiency and effectiveness and foster their research productivity. Acknowledgements: This paper is a modified version of a Conference proceedings paper presented at the Third Industrial Engineering and Management Conference (IEM3) that was held at The University of the West Indies in December 2014 (ISBN 978-976-620-289-7). The author would like to thank members of Research and Innovation and Graduate Studies (RIGS) who contributed views towards the development of the proposed TEP initiative at The UWI St Augustine Campus. References: Biggs, J. and Tang, C. (2009), Teaching for Quality Learning at

University: What the Student Does, 3rd Edition, McGraw-Hill/ Society for Research into Higher Education and Open University Press, Maidenhead

CityUHK (2014), City University of Hong Kong, Accessed November 2014, from http://www.cityu.edu.hk/

HKPolyU (2014), The Hong Kong Polytechnic University, Accessed November 2014, http://www.polyu.edu.hk/cpa/polyu/index.php

Lubben, F., Davidowitz, B., Buffler, A., Allie, S., and Scott, I. (2010), “Factors influencing access students’ persistence in an undergraduate science programme: A South African case study”, International Journal of Educational Development, Vol.30, pp.351-358

Pun, K.F. (2013), Annual Report of The School for Graduate

Studies and Research 2011/2012, UWI, St Augustine, January, p.1-2

Pun, K.F. (2014a), Research and Innovation and Graduate Studies, presentation made at the University Retreat 2014 on 22-23 May 2014, UWI, St Augustine

Pun, K.F. (2014b), Report of the Campus Coordinator (St Augustine), The School for Graduate Studies and Research, Prepared for the University Meeting of the Board for Graduate Studies and Research, UWI, Mona, Jamaica, 29 May

Pun, K.F. (2014c), “Closing the Gap between Graduate Intakes and Throughputs at UWI – A Throughput Enhancement Project”, Proceedings of the IEM3-2014 Conference, Faculty of Engineering, The University of the West Indies, Trinidad and Tobago, December 2014, pp.19-24

UCT (2012), ADP Position Paper: Factors That Affect Throughput, University of Cape Town, South Africa, Accessed November 2014 from http://www.ched.uct.ac.za/usr/ched/docs/ADP_Position.pdf

UWI (2014), UWI Mission Statement, Accessed September 2014 from https://sta.uwi.edu/aboutuwi/mission.asp

Author’s Biographical Notes: Kit Fai Pun is Professor of Industrial Engineering of the Faculty of Engineering and the Chair and Campus Coordinator for Graduate Studies and Research at The University of the West Indies. He is a Registered Professional Engineer in Australia, Europe, Hong Kong, and The Republic of Trinidad and Tobago. Professor Pun is a member of Caribbean Academy of Science and a Fellow/member of several professional bodies and learned societies. His research interests and activities include industrial engineering, engineering management, quality systems, performance measurement, innovation, and information systems. ■

WIJE, ISSN 0511-5728; http://sta.uwi.edu/eng/wije/

Page 87: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

Subscription Order Form Please enter my Annual subscription to the forthcoming Volume (2 Issues). • US$15.00 or equivalent in Trinidad and Tobago Dollars (Local subscription) • US$25.00 (By airmail) • Payment enclosed. Amount equivalent TT$/US$ (Cheque, Bank draft) • Send pro-forma invoice for despatch Please make cheque payable to The West Indian Journal of Engineering. Name: ______________________________________________________________________ Profession: ___________________________________________________________________ Areas of Specialisation: _________________________________________________________ Address:______________________________________________________________________

Telephone: _______________ Fax: __________________ E-mail: ______________________ ---------------------------------------------------------------------------------------------------------------------

(Cut along this line)

Change of Address Please note that my new postal address is:- __________________________________________

instead of:- ___________________________________________________________________ ---------------------------------------------------------------------------------------------------------------------

(Cut along this line) Send orders and/or change of address forms to: The Editorial Office, The West Indian Journal of Engineering, Block #1, Faculty of Engineering, The University of the West Indies, St. Augustine, Trinidad and Tobago, West Indies. Orders must be accompanied by payment which should be in Trinidad and Tobago (TT) Dollars or its equivalent at the time of order in US Dollars. Despatch of issues will commence only after receipt of payment.

For further information, contact:

Professor Kit Fai Pun, The Editor-in-Chief, The Editorial Office, West Indian Journal of Engineering, c/o Faculty of Engineering,

The University of the West Indies, St Augustine, Trinidad and Tobago, W.I. Tel: (868) 662-2002 Exts-83459/82069 • Fax: (868) 662-4414

E-mails: [email protected]; [email protected]

Page 88: Volume 37 Number 2 (ISSN 0511 5728) January 2015 WEST ...by Ricardo J. Rodriguez and Winston G Lewis . 23 A Model for Extending the Analysis of the Heptanes Plus Fraction for Trinidad

Notes and Guidance to Authors

The West Indian Journal of Engineering, WIJE (ISSN 0511-5728)

Copyright: Articles submitted to The West Indian Journal of Engineering, WIJE (ISSN

0511-5728) should be original contributions and should not be under

consideration for any other publication at the same time. Authors

submitting articles for publication warrant that the work is not an

infringement of any existing copyright and will indemnify the publisher

against any breach of such warranty. For ease of dissemination and to

ensure proper policing of use, papers and contributions become the legal

copyright of the publisher unless otherwise agreed. Submissions should be

sent to:

The Editor-in-Chief: Professor Kit Fai Pun, c/o WIJE, Faculty of Engineering, The University of the

West Indies, St Augustine, Trinidad and Tobago, West Indies. Tel: 1-868-662-

2002 exts-82069/83459; Fax: 1-868-662-4414; E-mails: [email protected];

[email protected]

Editorial Aim and Policy: The WIJE is an international journal which has a focus on the Caribbean

region. Since its inception in 1967, it is published twice yearly by the Faculty

of Engineering at The University of the West Indies and the Council of

Caribbean Engineering Organisations in Trinidad and Tobago.

WIJE aims at contributing to the development of viable engineering

skills, techniques, management practices and strategies relating to

improving the performance of enterprises, community, and the quality of

life of human beings at large.

Apart from its international focus, WIJE also addresses itself specifically

to the Caribbean development by identifying and supporting emerging

research areas and promoting various engineering disciplines and their

applications in the region.

WIJE welcomes the submission of papers in various engineering

disciplines and related areas. Emphasis is placed on the publication of

articles which seek to link theory with application or critically analyse real

situations with the objective of identifying good practice cross different

engineering and related disciplines.

Articles may be of a theoretical nature, be based on practical

experience, report a case study situation or report experimental results. The

prime requirement for acceptance of an article will not be its form but

rather that it:

(1) makes a significant original contribution to the field of engineering

and the advancement of engineering practices;

(2) is directly relevant to engineering, engineering management and

technology, and related areas;

(3) contains elements which have general application;

(4) is within the scope of the journal; and

(5) has generally not been published previously except in very limited

circulation.

The reviewing process: Each paper is to be reviewed by the Editor-in-Chief and, if it is judged

suitable for this publication, it is then sent to two referees for double-blind

peer-review. Based on their recommendations, the Editor-in-Chief then

decides whether the paper should be accepted as is, revised or rejected.

Manuscript requirements: Full manuscript should be submitted in double line spacing with wide

margins. The names of author(s) and their details-- brief autobiographical

note, affiliation, e-mail address and full international contact details must

appear on a sheet separate from the article. The author(s) should not be

identified anywhere else in the article. To facilitate the reviewing processes,

submissions via e-mail are advisable.

As a guide, technical/research papers should be between 3,000 and

6,000 words in length. Shorter articles (Communications, Discussions, Book

Reviews, etc.) should be between 500 and 2,000 words. Please provide the

word count on the first page of your paper. A title of not more than eight

words should be provided.

Authors must supply a structured abstract. Maximum is 250 words in

total. In addition provide up to six keywords which encapsulate the

principal topics of the paper and categorise your paper. Headings must be

short, clearly defined and not numbered. Notes or Endnotes should be

used only if absolutely necessary and must be identified in the text by

consecutive numbers, enclosed in square brackets and listed at the end of

the article.

All Figures (charts, diagrams and line drawings) and Plates

(photographic images) should be submitted in both electronic form and

hard-copy originals. Figures should be of clear quality, in black and white

and numbered consecutively with Arabic numerals.

Figures created in MS Word, MS PowerPoint, MS Excel, Illustrator and

Freehand should be saved in their native formats.

Electronic figures created in other applications should be copied from

the origination software and pasted into a blank MS Word document or

saved and imported into an MS Word document by choosing "Insert" from

the menu bar, "Picture" from the drop-down menu and selecting "From

File..." to select the graphic to be imported.

For figures which cannot be supplied in MS Word, acceptable

standard image formats are: pdf, ai, wmf and eps. If you are unable to

supply graphics in these formats then please ensure they are tif, jpeg, or

bmp at a resolution of at least 300dpi and at least 10cm wide.

To prepare screen shots, simultaneously press the "Alt" and "Print

screen" keys on the keyboard, open a blank Microsoft Word document

and simultaneously press "Ctrl" and "V" to paste the image. (Capture all the

contents/windows on the computer screen to paste into MS Word, by

simultaneously pressing "Ctrl" and "Print screen".)

For photographic images (plates) good quality original photographs

should be submitted. If supplied electronically they should be saved as tif

or jpeg riles at a resolution of at least 3oodpi and at least 10cm wide.

Digital camera settings should be set at the highest resolution/quality

possible.

In the text of the paper the preferred position of all tables, figures and

plates should be indicated by typing on a separate line the words "Take in

Figure (No.)" or "Take in Plate (No.)". Tables should be typed and included

as part of the manuscript. They should not be submitted as graphic

elements. Supply succinct and clear captions for all tables, figures and

plates. Ensure that tables and figures are complete with necessary

superscripts shown, both next to the relevant items and with the

corresponding explanations or levels of significance shown as footnotes in

the tables and figures.

References to other publications must be in Harvard style and

carefully checked for completeness, accuracy and consistency. This is very

important in an electronic environment because it enables your readers to

exploit the Reference Linking facility on the database and link back to the

works you have cited through CrossRef. You should include all author

names and initials and give any journal title in full.

You should cite publications in the text: (Adams, 2008) using the first

named author's name or (Adams and Brown, 2008) citing both names of

two, or (Adams et al., 2008), when there are three or more authors. At the

end of the paper, a reference list in alphabetical order should be supplied:

• For books: Surname, initials, (year), title of book, publisher, place of

publication, e.g., Walesh, S. G. (2012), Engineering Your Future: The

Professional Practice of Engineering, 3rd Edition, ASCE Press/John Wiley

& Sons, New Jersey, NJ.

• For book chapters: Surname, initials, (year), "chapter title", editor's

surname, initials, title of book, publisher, place of publication, pages,

e.g., Liebowitz, J. (2005), "Conceptualising and implementing

knowledge management", in Love, P. E. D., Fong, P. S. W. and Irani, Z.,

(ed.), Management of Knowledge in Project Environments, Elsevier,

New York, NY, pp. 1-18

• For journals: Surname, initials, (year), "title of article", journal name,

volume, number, pages, e.g. Tsang, A. H. C. (2012), “A review on trend

tests for failure data analysis", West Indian Journal of Engineering, Vol.

35, No.1, July, pp.4-9.

• For electronic sources: Surname, initials, or institution, (year), name of

website, website address, date updated (if any) or date visited, e.g.,

EFQM (2012), European Foundation for Quality Management, available

at: http://www.EFQM.org/ (Dated: 1 January 2012)

Final submission of the article: Once accepted for publication, the Editor may request the final version as

an attached file to an e-mail or to be supplied on a diskette or a CD-ROM

labelled with author name(s); title of article; journal title; file name.

The manuscript will be considered to be the definitive version of the

article. The author must ensure that it is complete, grammatically correct

and without spelling or typographical errors.

The preferred file format is Word. Another acceptable format for

technical/mathematics content is Rich text format.