makale dows 2
Post on 08-Aug-2018
234 Views
Preview:
TRANSCRIPT
-
8/22/2019 Makale Dows 2
1/12
Journal of Loss Prevention in the Process Industries 20 (2007) 7990
The integration of Dows fire and explosion index (F&EI) into process
design and optimization to achieve inherently safer design
Jaffee Suardina, M. Sam Mannana,, Mahmoud El-Halwagib
aMary Kay OConnor Process Safety Center, Artie McFerrin Department of Chemical Engineering, Texas A&M University,
College Station, TX 77843-3122, USAbArtie McFerrin Department of Chemical Engineering, Texas A&M University, College Station, TX 77843-3122, USA
Received 14 March 2006; received in revised form 27 September 2006; accepted 16 October 2006
Abstract
For the processing industries, it is critically to have an economically optimum and inherently safer design and operation. The basic
concept is to achieve the best design based on technical and business performance criteria while performing within acceptable safety
levels. Commonly, safety is examined and incorporated typically as an after-thought to design. Therefore, systematic and structured
procedure for integrating safety into process design and optimization that is compatible with currently available optimization and safety
analysis methodology must be available.
The objective of this paper is to develop a systematic procedure for the incorporation of safety into the conceptual design and
optimization stage. We propose the inclusion of the Dow fire and explosion index (F&EI) as the safety metric in the design and
optimization framework by incorporating F&EI within the design and optimization framework. We first develop the F&EI computer
program to calculate the F&EI value and to generate the mathematical expression of F&EI as a function of material inventory and
operating pressure. The proposed procedure is applied to a case study involving reaction and separation. Then, the design and
optimization of the system are compared for the cases with and without safety as the optimization constraint. The final result is the
optimum economic and inherently safer design for the reactor and distillation column system.r 2006 Published by Elsevier Ltd.
Keyword: Fire and explosion index; Inherently safer design; Process safety; Process design and optimization
1. Introduction
Adapted from the Center for Chemical Process Safety
(CCPS), hazard is defined as physical or chemical
characteristic that has the potential for causing harm to
people, the environment, or property (Crowl, 1996). It is
very important to note that the hazards are intrinsic andare the basic properties of the material or its conditions of
use. For example, at a certain condition and concentration,
10,000 lb of propane holds the same amount of energy
which could be released by 28 tons of TNT. Those energies
are inherent to the propane, cannot be changed, and will be
released when equipment or other failure happens and
leads to an incident.
While an inherently safe plant infers a plant that has
no hazards on an absolute basis, such plant with zero
risk might be impossible to design and to operate.
Therefore, the need to manage hazards and risks strategi-
cally and systematically arises and one of the strategies is
the inherently safer design concept (as opposed to
inherently safe plant). In addition, the best strategy seeksto combine inherently safer design with process design and
optimization at the early stages of design where the degree
of freedom for modification is still high.
Mansfield and Cassidy (1994) presented an inherently
safer approach to plant design and general theory on how
it can be built into the design process. Palaniappan,
Srinivasan, and Tan (2004) applied inherently safety index
for identifying hazards and generating alternative designs.
Similar to the aforementioned efforts, others have been
concentrating on the safety analysis methodology without
ARTICLE IN PRESS
www.elsevier.com/locate/jlp
0950-4230/$ - see front matterr 2006 Published by Elsevier Ltd.
doi:10.1016/j.jlp.2006.10.006
Corresponding author. Tel.: +1 979862 3985; fax: +1 979845 6446.
E-mail address: mannan@tamu.edu (M. Sam Mannan).
http://www.elsevier.com/locate/jlphttp://localhost/var/www/apps/conversion/tmp/scratch_15/dx.doi.org/10.1016/j.jlp.2006.10.006mailto:mannan@tamu.edumailto:mannan@tamu.eduhttp://localhost/var/www/apps/conversion/tmp/scratch_15/dx.doi.org/10.1016/j.jlp.2006.10.006http://www.elsevier.com/locate/jlp -
8/22/2019 Makale Dows 2
2/12
applying it into design and optimization in a single
framework. Therefore, there is a need to systematize the
incorporation of safety metrics in the design stage. This is
the focus of this paper. Section 2 describes the problem
statement and the proposed approach. Later, the proposed
method is applied to a case study and the results are
analyzed.
2. Inherently safer design
Inherently safer design infers the elimination of hazards
as much as possible out of a chemical or physical process
permanently as opposed to using layers of protection.
There are four primary principles of inherently safer design
concept proposed by Kletz (1991):
1. Intensificationto reduce the inventories of hazardous
materials as more inventory of hazardous chemicals
means more hazards.
2. Substitutionto use less hazardous materials in the
process.
3. Attenuationto operate a process at less dangerous
process conditions (pressure, temperature, flow rate,
etc.).
4. Limitation of effectsto design the process according to
the hazards offered by the process in order to reduce the
effects of the hazards.
In the US, inherently safer design started receiving more
attention following a highly-praised paper presented by
Kletz in 1985 at the 19th Loss Prevention Symposium of
the American Institute of Chemical Engineers (AIChE)(Hendershot, 1999).
3. Safety studies
The most common and traditional approach has focused
on layers of protection (LOP) where additional safety
devices and features are added to the process, as shown in
Fig. 1 (American Institute of Chemical Engineers (AIChE),
1994). The LOP method has been successful in analyzing
safety systems. However, this approach has several
disadvantages as listed below (Crowl, 1996):
LOP increase the complexity of the process, and hence thecapital and operating cost. In the oil and gas industries,
1530% of the capital cost goes to safety issues and
pollution prevention (Palaniappan et al., 2004).
The hazards within the process remain, even when LOPare installed and are built based on the anticipation of
incidents, as shown in Fig. 2(a). Since nature might find
creative ways to release hazards, there are always
dangers from unanticipated failure mechanisms that
the LOP are not ready for, as shown in Fig. 2(b).
Since no LOP can be perfect, failures or degradation inLOP may pose risks offered by the hazards that lead to
incidents, as shown in Fig. 2(c).
Other efforts by the industries and researchers toward
safety studies tend to focus on hazard identification and
control. There has been some work in developing more
advanced hazard and risk analysis methods such as Failure
Modes and Effects Analysis (FMEA), Fault Tree Analysis(FTA), Event Tree Analysis (ETA), CauseConsequence
Analysis (CCA), Preliminary Hazard Analysis, Human
Reliability Analysis (HRA), and Hazard and Operability
Study (HAZOP) in addition to traditional methods such as
check list, safety review, relative ranking, and Whatif
analysis (Wang, 2004).
Several inherent safety efforts taken by US corporations
and US affiliates of European companies are listed below:
Dow Chemical CompanyDeveloped the Dow Fireand Explosion Index (American Institute of Chemical
Engineers (AIChE) (1994)) and the Dow Chemical
ARTICLE IN PRESS
Basic
Control
Critical Alarms, Human/Manual Intervention
Process
Design
Feed,F,X
F
Condenser
Reboiler
Top
Product,X
D, D
BottomProduct,
B,XB
Safety
Instrumented
System
Physical Protection
(Relief devices, etc)
Emergency
Responses
Fig. 1. Typical layers of protection for CPI (adapted from Hendershot,
1997).
J. Suardin et al. / Journal of Loss Prevention in the Process Industries 20 (2007) 799080
-
8/22/2019 Makale Dows 2
3/12
Exposure Index (AIChE, 1993) as hazard ranking
methodology based on inherent safety principles.
Exxon Chemical CompanyDescribed inherent safety,
health and environment review process based on a lifecycle approach (Hendershot, 1999).
Rohm and Haas Major Incident Prevention Programused checklist based on inherent safety principles for
hazard elimination and risk reduction (Hendershot,
1999).
In addition to actions taken by the CPI, actions have
also been taken by government in the form of federal
regulations such as the Process Safety Management (PSM)
regulation promulgated by the Occupational Safety and
Health Administration (OSHA) and the Risk Management
Program (RMP) regulation promulgated by the Environ-
mental Protection Agency (EPA).
Overall impression these efforts is that inherently safer
design principles have not been systematically applied. As
opposed to layers of protection concept, the concept of
inherently safer design is to reduce the inherent hazards
rather than to control them. There are two advantages
about having lower hazards: they need lesser LOP, less
complex LOP, and offer lower magnitude of hazards, as
shown in Figs. 3 and 4.
Another impression on the traditional approaches is that
the efforts focus on hazard identification and control. In
addition, currently optimization is performed as an attempt
to enhance the process design and the operation conditions
of equipment to achieve the largest production, the greatest
profit, minimum production cost, and the least energy
usage. Whereas, neither objective functions nor constraint
conditions contain safety parameters in the traditionalprocess optimization.
4. Hazard indices
There are several hazard indices available as tools for
chemical process loss prevention and risk management.
Although no index methodology can cover all safety
parameters, Dow fire and explosion index (F&EI), and
safety weighted hazard index (SWeHI) are found to be
robust (Khan & Amyotte, 2003). The F&EI is the most
widely known and used in the chemical industries. The
following are indices available in the industries and
research:
F&EI (American Institute of Chemical Engineers(AIChE), 1994) and Dows chemical exposure hazards
(Dow, 1993) as tools to determine relative ranking of
fire, explosion, and chemical exposure hazards. Etowa,
Amyotte, Pegg, and Khan (2002) have developed a
computer program to automate F&EI calculation and
perform sensitivity analysis using Microsofts Visual
Basic. However, their program was not intended to
determine business interruption and loss control credit
factors, to conduct process unit risk analyses, to
automate the sensitivity analysis in order to integrate
ARTICLE IN PRESS
Anticipated Potential Incidents Unanticipated Potential Incidents
unanticipated
Layers do not
work for
mechanisms
anticipated
LOP for
Degraded LOP
LOP
incidents
Actual Risk
Actual Risk
Potential Incidents
a b
c
Fig. 2. Layers of protection characteristics. (a) LOP reduces the anticipated potential incidents, (b) LOP does not reduce unanticipated potential incidents,
(c) degraded LOP does not reduce any potential incidents (adapted from Hendershot, 1997).
J. Suardin et al. / Journal of Loss Prevention in the Process Industries 20 (2007) 7990 81
-
8/22/2019 Makale Dows 2
4/12
F&EI calculation into process design and optimization
framework.
SWeHI as a tool to define fire, explosion, and toxicrelease hazards (Khan, Sadiq, & Amyotte, 2003).
Environmental Risk Management Screening Tools(ERMSTs) from Four Elements, Inc. for ranking
environmental hazards including air, ground water,
and surface water pollution. (Hendershot, 1999).
Mond Index as a tool to define fire, explosion, and toxicrelease hazard (Hendershot, 1999).
Hazardous waste index (HWI) as a tool for flamm-ability, reactivity, toxicity, and corrosivity hazard of
waste materials (Khan et al., 2003; Khan & Amyotte,
2003).
Transportation Risk Screening Model (ADLTRSs
) as atool for determining risk to people and environment
posed by chemical transportation operations (Khan
et al., 2003).
Inherent safety index was developed by Heikkila (1999)of Helsinki University of Technology. This method
classifies safety factors into two categories: chemical and
process inherent safety. The chemical inherent safety
includes the choice of material used in the whole process
by looking at its heat of reaction, flammability,
explosiveness, toxicity, corrosivity, and incompatibility
of chemicals. The process inherent safety covers the
process equipment and its conditions such as inventory,
pressure, temperature, type of process equipment, and
structure of the process.
Overall inherent safety index was developed by Edwardand Lawrence (1993) to measure the inherent safety
potential for different routes of reaction to obtain the
same product.
Fuzzy logic-based inherent safety index (FLISI) wasdeveloped by Gentile (2004). One of the major problems
in applying inherent safety is that safety mostly based on
the qualitative principles that cannot be easily be
evaluated and analyzed. FLISI was an attempt to use
hierarchical fuzzy logic to measure inherent safety and
provide conceptual framework for inherent safetyanalysis. Fuzzy logic is very helpful for combining
qualitative information (expert judgment) and quanti-
tative data (numerical modeling) by using fuzzy
IFTHEN rules.
5. Problem statement and overall approach
The problem to be addressed in this paper may be stated
as follows: Given a processing system that requires
economic optimization, devise a procedure that achieves
optimum process design while insuring that the design
meets certain safety criteria. In order to address the
problem, several challenges have to be overcome. These
include the following:
What is the best design based on technical and businessperformance within acceptable safety level ?
How to quantify safety and incorporate the safetymetric during design?
How to perform the conceptual design in a computa-tionally efficient manner?
This paper attempts to perform this optimization and
analyze the result by modifying common process optimiza-
ARTICLE IN PRESS
LOP 1
LOP 2
ProcessDesign
Feed
,F,X
F
Condenser
Reboiler
TopProduct,
XD, D
BottomProduct,B, X
B
Fig. 3. Inherently safer process design requires no or less additional LOP
adapted from Hendershot (1997).
Potential Incidents
No or Less
LOP needed
by applying
ISDActual
Risk
Fig. 4. Potential incidents for inherently safer design (adapted from
Hendershot, 1997; Hendershot, 1999).
J. Suardin et al. / Journal of Loss Prevention in the Process Industries 20 (2007) 799082
-
8/22/2019 Makale Dows 2
5/12
tion which focuses only on the technical and business
performance. The modified procedure for the use of F&EI
as safety parameter in the optimization is given in Fig. 6.
The following four steps were conducted to illustrate the
proposal methodology:
1. Computerize Dows fire and explosion index calcula-tion.
2. Generate F&EI mathematical expressions as a function
of operating pressure and the amount of materials in the
process units.
3. Propose a general procedure for integrating safety
parameters into process design and optimization.
4. Optimize the reactor and distillation column as a case
study with economic, performance, and safety para-
meters as the constraints to verify the procedure.
Even though some of the data shown came from the
F&EI computer program that we developed, this paperfocuses only on the methodology, thus the development of
the F&EI computer program is not shown.
6. Dows fire and explosion index (F&EI) methodology
F&EI is the most widely used hazard index calculation
and has been used and revised for six times since 1967. The
last revision is the seventh edition which was published in
1994 and is applied for this research. Fig. 5 shows the
F&EI procedure.
The F&EI calculation is done as the following. First,
material factor (MF, the measure of the potential energy
released by material under evaluation) is obtained from
databases, material safety data sheet (MSDS), or manual
calculation (using flammability, NF, and reactivity value,
NR). Then, determine the sum of penalties that contributes
to loss probability and its magnitude (general process
hazard factor, F1) and the sum of factors that the factor
that can increase the probability and historically contri-
butes to major fire and explosion incidents (special process
hazards factor, F2).
General process hazards cover six items, namely,
exothermic chemical reactions, endothermic processes,
material handling and transfer, enclosed or indoor process
units, access and drainage and spill control, although itmay not be necessary to apply all of them. Special process
hazards cover twelve items: toxic material, sub-atmo-
spheric pressure, operation in or near flammable range,
dust explosion, relief pressure, low temperature, quantity
of flammable/unstable material, corrosion and erosion,
leakage-joints and packing, use of fired equipment, hot oil
heat exchange system, and rotating equipment. Each of the
items is represented in terms of penalties and credit
factors.
The fire and explosion index (F&EI) is calculated using
(American Institute of Chemical Engineers (AIChE), 1994)
F3 F1 F2, (1)
F&EI MF F3. (2)
The next step is business interruption calculation (BI)
that is done based on F&EI calculated. F&EI will
determine the radius and the area of exposure. Any
equipment within this area will be exposed to the hazard.
The damage factor is then calculated that represents the
overall effect of fire and blast damage produced by releaseof fuel or reactivity energy from unit equipment. By having
original equipment cost and value of production per month
(VPM) as an input, the actual minimum probable property
damage (Actual MPPD) can be determined and then BI is
calculated by Eq. (3) (American Institute of Chemical
Engineers (AIChE), 1994):
BI US$ MPDO
30 VPM 0:7. (3)
7. Case study: overview
Procedure in Fig. 6 is examined in order to support the
argument that integrating safety into process design and
optimization gives benefits without necessarily violating the
economic and technical parameter. Hence, the final design
is the optimum economics and the inherently safer design
for the reactor-distillation column system.
Basic chemical engineering processes include reaction,
separation, and mixing. It is very common in the chemical
process industries that reactor is followed by separator to
separate the un-reacted raw materials and the specified
products. Thus, reactordistillation column system is acommon system used in the chemical process industries and
studying its optimization is very important.
It is also a fact that in performing economic analysis of a
reactor, the separator should be included since there is
trade-off between reactorseparator systems as shown in
Fig. 7. Economic balance between a high reactor cost at
high conversion and a high separation cost at low
conversion will determine the optimum reactor conversion
based on the total cost. Therefore, it is necessary to have a
procedure to improve reactor performance and/or reac-
tordistillation column system to produce desired products
while in the range of acceptable economic profit and safety
level.
8. Case study: reactor and distillation column system
The reactordistillation system is shown in Fig. 8. In
addition, it is important to note that the data presented in
this problem statement are adapted from several sources
without specifically representing a certain process. The
reason behind it is that this research focuses on the concept
of integrating F&EI value, and not in the complexities of
the calculation of F&EI where expert judgment is really
needed for the optimization process which includes a lot of
variables.
ARTICLE IN PRESS
J. Suardin et al. / Journal of Loss Prevention in the Process Industries 20 (2007) 7990 83
-
8/22/2019 Makale Dows 2
6/12
The reactor is to produce 645 million pounds of chemical
B per year from chemical A following the reaction of:
A ! B gas phase
The reaction properties allow only a portion of the
chemical A to be converted into chemical B. Then the output
of the reactor in the form of mixture of A and B will be fed to
the distillation column. Distillation column separates the
chemical A and B in order to have product A in a certain
number of purity. The data used in this case study are:
A-B (gas phase reaction)
Hazardous material: chemical A
Product of the reactor: 645 million pounds of chemicalB per year
Pressure range: 28 atm Isothermal and plug flow reactor Feasible optimum conversion:4070% Distillation column operating pressure:1016 atm
9. Objective function and optimization model
Optimization requires mathematical modeling. This
research employs F&EI as the safety parameter which its
mathematical expressions are available by using F&EI
program through its sensitivity analysis feature. The
ARTICLE IN PRESS
Select Pertinent Process
Unit
Calculate F1
General Process Hazards Factor
Detemine F & EI
F & EI = F3 x Material Factor
Calculate F2
Special Process Haxards Factor
Determine Area of Exposure
Determine Process Unit Hazards
Factor F3 = F1 x F2
Determine material factor
Calculate Loss Control
Credit Factor = C1 x C2 x C3
Determine Replacement Value in Exposure Area
Determine Base MPPD
Determine Actual MPPD
Determine MPDO
Determine BI
Determine Damage Factor
Fig. 5. F&EI procedure (AIChE, 1994).
J. Suardin et al. / Journal of Loss Prevention in the Process Industries 20 (2007) 799084
-
8/22/2019 Makale Dows 2
7/12
mathematical expressions are presented in the next section
along with other data required for the optimization.
9.1. Reactor optimization data
In this paper we use the plug flow reactor (PFR) as a case
study. The performance of the reactor can be determined
by using the following data:
1. The rate of reaction and the mass transfer characteristics
of the reacting fluid. This determines the volume of the
reactor needed to produce the specified product.
2. The constraints dictated by the reactor are set up such as
the type and geometry of the reactor. This determines the
cost of the reactor and thus the economic parameters.
Economic variables of reactor are type, diameter, height,
design pressure, materials of construction, and capacity
(Edgar, Himmeblau, & Lasdon, 2001).
PFR is a cylindrical vessel that can be determined in the
same way as that of the distillation column with several
modifications, only in different orientation. The reactor is a
horizontal pressure vessel in cylindrical form. The free on
board (f.o.b) purchase cost (CP) of horizontal pressure
vessel including the nozzles, the manholes, a skirt, and the
internals (not plates and/or packing) are described by
Seider, Seader, and Lewin (2004).
Conversion (X) is the measure on how far the reaction
has proceeded and is in the range of 01 (100%
conversion). In optimizing a reactor, the conversion might
not reach 100% conversion due to other constraints such
as economic factors. For reactions with more than one
reactant, the material which the conversion is based on
ARTICLE IN PRESS
Dows F & EI Method
By using F & EI Program
F & EI value < 128
DESIGN & OPTIMIZATION
ACCEPTABLE
LEVEL
FINAL DESIGN
OBJECTIVE FUNCTIONS:Economic Parameter
MODELING:Safety (Inherently Safer Design Principle )
Material Balance
Sizing and Costing of Equipment
Constraints:Safety (Inherently Safer Design Principle )
Technical Performance
Constraints Adjustment
Reactions and Materials Selection
Equipment Selection
Operating Condition Selection
etc
PROPOSED DESIGN
YES
NO
Red : improvement in optimization
Black : Current optimization
Fig. 6. The integration of safety parameter into process design and optimization.
Separator
Reactor
Total
1.0
X optimum
Cost($)
Reactor Conversion (X)
0
Fig. 7. Costs of reactor and distillation column as a function of reactor
conversion (Smith, 1995).
J. Suardin et al. / Journal of Loss Prevention in the Process Industries 20 (2007) 7990 85
-
8/22/2019 Makale Dows 2
8/12
must be specified. Conversion expression is
Conversion Conversion amountof material consumed
amount material provided Conversion amount in inlet streamamount in outlet streamamount in inlet stream .
4
The data for the reactor design and optimization are:
Objective function: Minimize Total reactor cost
Total cost Cv Cpl. (5)
Technical constraints (Fogler, 2002):
Volume fn X; FAo; CAo,
Volume p
4D2iL
FAo
kCAo2 ln
1
1 X X
. 6
Economic constraints (Seider et al., 2004)
Cv fn W,
CV expf8:717 0:2330 lnW 0:04333 lnW 2g,
(7)
W fnD; ts; L; Di,
W pDi tsL 0:8Di tsr, (8)
ts fnPd; Di,
tp PdDi
2SE 1:2Pd, (9)
Cpl fnDi,
CPL 1580Di0:20294, (10)
Pd fnPo,
Pd expf0:60608 0:91615lnPo 0:0015655lnP02g.
(11)
With Cv as cost of vessel, ts as vessel thickness, X as
conversion, Pd as design pressure, V as volume, CAo as initial
concentration ofA, Cplas cost of the platform, Po as operating
pressure, W as weight of the vessels, FAo as input material.
9.2. Distillation column optimization data:
Distillation column consists of tower vessel and plates/
packing. The capital cost of the distillation column is the
summation of the vessel cost and the installed plates/
packing cost. The data for the distillation column design
and optimization are:
Objective function Minimize Total reactor cost
Total distillation cost Cv Cpl Ct, (12)
Technical Constraint (Peters & Timmerhaus, 1991)
uf K1
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffirL rG
rG
r, (13)
Di 4Vt
uf rG p
0:5, (14)
L trayspacing N. (15)
ARTICLE IN PRESS
Steam
Cooling Water
Liquid Phase FlowVapor Phase Flow
F, XF
Condenser
Reboiler
Top Product,Chemical A
Reflux(Liquid)
Bottom Product,Chemical B
Vapor
Vapor
(Liquid)
(Liquid)
Tray
A
B (gas phase)Volume, Conversion
Chemical A
Chemical Aand B
Fig. 8. Reactor-distillation column system for the case study.
J. Suardin et al. / Journal of Loss Prevention in the Process Industries 20 (2007) 799086
-
8/22/2019 Makale Dows 2
9/12
Economic Constraints (Seider et al., 2004) Cv fn (W)
CV expf7:0374 0:18255lnW 0:02297lnW2g,
(16)
W fn (D, ts, L, Di)
W pDi tsL 0:8Di tsr, (17)
ts fn (Pd, Di)
tp PdDi
2SE 1:2Pd, (18)
Cpl fn(Di)
Cpl 237:1Di0:62216L0:80161, (19)
Pd expf0:60608 0:91615lnPo
0:0015655lnP02g, 20
With
Cpl 237:1Di0:62216L0:80161, (21)
CT NTFNTFTTFTMCBT, (22)
CBT 369 exp0:1739Di. (23)
With Cv as cost of vessel, ts as vessel thickness, Ct as cost of
the tray, Pd as design pressure, Cpl as cost of the platform,
Po as operating pressure, and W as weight of the vessels.
10. Result: F&EI mathematical expression
F&EI calculation is performed on the case study using
the F&EI program. The sensitivity analysis feature on the
F&EI program provides the mathematical expression of
F&EI as a function of pressure and material inventory. As
shown in Figs. 11 and 12, by varying the operating pressure
at constant material inventory and vice versa, the F&EI
program will automatically generate a chart. By using least
squares method, the mathematical expression of the chart
can be determined. This procedure is applied for both the
reactor and the distillation column.
For reactor, the expressions are:
F&EI 3 108Inventory2 0:0012Inventory 88:46,
(24)
F&EI 0:1176 pressure 109:8. (25)
For the distillation column, the expressions are:
F&EI 1 108Inventory2 0:0018
Inventory 101:16, 26
F&EI 5 105pressure2 0:1072
pressure 106:83. 27
Those expressions are the safety constraint in the
optimization and are applied according to the procedure
as shown in Fig. 6.
11. Result: optimization
Optimization is performed by LINGO optimization
software and uses principles of process integration (e.g.
El-Halwagi, 2006). For the reactor-distillation column
system, the total cost is the total of the reactor cost and
the distillation cost. As shown in Table 1, in the case study
of reactor and distillation column presented in this paper,
the feasible optimization solution without F&EI (safety
parameter) as the constraint is in the range of 4070% of
reaction conversion. American Institute of Chemical
Engineers (AIChE) (1994) recommends that Dows Fire
and Explosion Index as the safety constraint should not be
more than 128 as shown by the vertical line in the Fig. 10,
where the conversion is 49%. Therefore, applying F&EI
gives the new conversion range which is 4970%.
The vertical line also shows the conversion which gives
the F&EI value of 128. If the safety parameter is not
considered, the total cost will be available for the
conversion in the range of 4070%, as shown in Fig. 9.
However, safety parameter will not allow the process to
apply those conversions since at this point the process is
not inherently safer according to Dows F&EI methodol-
ogy. The feasible range of conversion after safety
parameter has been included is in the range of 4970%,
ARTICLE IN PRESS
5.40E+05
5.50E+05
5.60E+05
5.70E+05
5.80E+05
5.90E+05
6.00E+05
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Reactor Conversion
ReactorandDistillation
cost
1.08E+06
1.10E+06
1.12E+06
1.14E+06
1.16E+06
1.18E+06
1.20E+06
1.22E+06
Total CostRight Axis
Distillation Column Cost
Reactor Cost
Feasible Area
Fig. 9. Reactor-distillation column system cost without safety constraint.
Table 1Optimization result
Constraints Optimal conversion range
No safety constraint 4070%
F&EI o128 (recommended by
AIChE, 1994)
4970% (730% less)
J. Suardin et al. / Journal of Loss Prevention in the Process Industries 20 (2007) 7990 87
-
8/22/2019 Makale Dows 2
10/12
as shown in Fig. 10. This is 30% less than the original
conversion range which affects the economic performance
of the system. F&EI asks for higher conversion of reaction
due to the fact that higher conversion reaction produces
more products with less reactant compared to low
conversion reaction. In addition, the lower the conversion
the higher the reactant inventory needed thus the higher
material inventory required by the reactor and the higher
distillation column capacity required to perform specified
separation. This significant decrease in the conversion
range shows that the F&EI as safety constraint will affect
the overall system in this case study.
Fig. 10 also shows that the safety parameter is employedonly as one of the constraint for the optimization. It will
not change any of the design value such as the cost,
the reactor volume, the number of trays, etc. As a
constraint, safety will only limit the feasible area for the
optimization solution. Thus, if the optimization with
constraint is performed, the result will be unacceptable
and the designer has to adjust the constraint or the other
design variables.
In other word, F&EI is incorporated as a cutting point
between inherently safer and non-inherently safer based on
F&EI target value and not a variable. In this paper,
conversion range of less than 49% requires the amount of
reactant that produces F&EI value higher than 128 while
conversion range of bigger than 49% yields F&EI value of
less than 128. The F&EI values for conversion range of
4970% are absolutely less than 128 and considered as
inherently safer according to Dow F&EI method with
target value of 128. Hence the F&EI value for conversion
range after the cutting point does not affect design decision
significantly as it is already meet the objective function of
the optimization process, which are technical, business and
safety aspect of the system. However, it is important to
note that the F&EI value as safety constraint is not
restricted to 128 as it depends on the target value and will
change move the cutting point to a different one.
Figs. 11 and 12 show that F&EI value decreases with
lower material inventory. Hence, higher reaction conver-
sion produces lower F&EI value and it will move thecutting point of 128 showed in Fig. 10 more to the right,
and vice versa.
The advantages of integrating F&EI as safety parameter,
into process design and optimization are:
The final design is an inherently safer design. The trade off between safety and other constraints can
be adjusted according to the policy of the owner/
designer.
Since both safety studies and design and optimizationare performed at the same time, safety and other
constraints will affect each other significantly.
The safety level of the design is known even before theoptimal design is achieved. Thus, detail design is
worked after the safety level is acceptable.
F&EI value as safety constrain is flexible andcan be determine based on the objective of the
design and safety level needed. This changes the cutting
point.
12. Conclusion
Mathematical expression represented safety parameter is
required when safety is included in the optimization. A
ARTICLE IN PRESS
5.40E+05
5.50E+05
5.60E+05
5.70E+05
5.80E+05
5.90E+05
6.00E+05
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Reactor Conversion
ReactorandDistillationcost
1.08E+06
1.10E+06
1.12E+06
1.14E+06
1.16E+06
1.18E+06
1.20E+06
1.22E+06
Dow's F & EI= 128
Total Cost
Right Axis
Distillation Column Cost
Reactor Cost
Fig. 10. Reactor-distillation column with safety constraint.
Operating Pressure vs Fire and Explosion Index
REACTOR
REACTOR
Material Inventory vs Fire and Explosion Index
FireandE
xplosionIndex
FireandE
xplosionIndex
Pressure (Psig)
y = 3E-08x2 + 0.0012x + 88.46
180
160
15 25 35
140
120
100
10
108
110
112
114116
118
120
122
124
126
128
20 30
2040
60
80
5
20 40 60 80 100 120 140 160
00
0
y = 1E-16x2 + 0.1176x + 109.8
Material Inventory (1000 Ibs)
Fig. 11. Sensitivity analysis and F&EI mathematical expression for
reactor.
J. Suardin et al. / Journal of Loss Prevention in the Process Industries 20 (2007) 799088
-
8/22/2019 Makale Dows 2
11/12
simple way to have it for Dows Fire and Explosion Index
can be done by using the F&EI computer program
developed for this paper.
The case study on reactor-distillation column system
proves that the proposed procedures of integrating
safety parameter (Dows F&EI in this research) into
process design and optimization framework quant-
itatively and systematically are very useful. The safety
parameter acts as a constraint rather than as the
process variable. It only limits the feasible area of the
conversion optimization solution. It does not change any
of the process variables. In the case study of reactor
and distillation column presented in this paper, the
feasible optimization solution without safety as the
constraint is in the range of 4070% of reaction conver-
sion. F&EI application as the safety parameter narrows the
conversion range into 4970%. The conversion range of
4049% is not inherently safer according to F&EI
methodology.
When F&EI value of 128 as constraint is applied, safety
constraint is not significantly affecting the decision making
any further within the conversion range of 49%70%. By
applying the different F&EI value as needed, one can find
different and the right cutting point for the design. Based
on the fact that lower reaction conversion demands higher
amount of reactant, lower F&EI value selected reduces the
conversion ranges. This changes the feasible conversion
range for the process under evaluation.
There are several contributions presented by this
methodology:
Getting safety parameter as a mathematical expressionhas been a problem in safety thus inhibits the integration
of safety parameter into process design systematically.
This paper presents a simple way of generating
expressions from available hazard analysis which can
be useful in modeling and predicting the hazard of the
specific process.
Proving that there is a possibility for Dows fire andexplosion index method to be integrated into process
design and optimization framework while satisfying the
specified technical and economic parameters.
Presenting general idea on how to integrate safety intoprocess design and optimization. Instead of using only
Dows fire and explosion index, reader might assign
other methodology that fits their specific process.
However, the idea is still the same which is having the
mathematical expression of the safety study undergo
and utilize it in the optimization.
ARTICLE IN PRESS
DISTILLATION COLUMN
Operating Pressure vs Fire and Explosion Index
DISTILLATION COLUMN
Material Inventory vs Fire and Explosion Index
FireandExplosionIndex
FireandExplos
ionIndex
Operating Pressure (Psig)
y = 5E-05x2 + 0.1072x + 106.83
160
150
150
150
250 350
140
130
120
110
100
100
100
200
200
300 400
90
8050
10 20 30 40 50 60 70 80
50
0
0
0
y = 1E-08x2 + 0.0018x + 101.16
Weight (1000 Ibs)
Fig. 12. Sensitivity analysis and F&EI mathematical expression for distillation column.
J. Suardin et al. / Journal of Loss Prevention in the Process Industries 20 (2007) 7990 89
-
8/22/2019 Makale Dows 2
12/12
References
American Institute of Chemical Engineers (AIChE). (1994). Dows fire and
explosion index hazard classification guide (7th ed.). New York:
AIChE.
Crowl, D. A. (1996). Inherently safer chemical processes: A life cycle
approach. New York: AIChE.
Edgar, T. F., Himmeblau, D. M., & Lasdon, L. S. (2001). Optimization ofchemical processes (2nd ed). Singapore: McGraw-Hill.
Edward, D. W., & Lawrence, D. (1993). Assessing the inherent safety of
chemical process routes: is there a relation between plant cost and
inherent safety. TransIChemE, Pt. B(Process Safety and Enviromental
Protection), 71B, 252.
El-Halwagi, M. M. (2006). Process integration, process systems engineer-
ing, Vol. 7. Amsterdam: Academic Press.
Etowa, C. B., Amyotte, P. R., Pegg, M. J., & Khan, F. I. (2002).
Quantification of inherent safety aspects of the dow indices. Journal of
Loss Prevention in the Process Industries, 15, 477487.
Fogler, H. S. (2002). Elements of chemical reaction engineering (3rd ed).
New Delhi, India: Prentice-Hall of India.
Gentile, M. (2004). Development of a hierarchical fuzzy model for the
evaluation of inherent safety. Ph.D. Dissertation. Texas A&M
University, College Station.Heikkila, A. M. (1999). Inherent safety in process plant design: an index
based approach. Ph.D Dissertation, Technical Research Center of
Finland, VTT, Espoo: Finland.
Hendershot, D. (1997). Safety through design in the chemical process
industry: Inherently safer process design, Benchmarks for World Class
Safety Through Design Symposium, Rohm and Haas Company,
Bloomingdale, Illinois.
Hendershot, D. (1999). Inherently safer process design philosophy, Rohm
and Haas Company, Herzlia, Israel.
Khan, F. I., & Amyotte, P. R. (2003). How to make inherent safety
practice a reality. The Canadian Journal of Chemical Engineering, 81,
116.
Khan, F. I., Sadiq, R., & Amyotte, P. R. (2003). Evaluation of availableindices for inherently safer design options. Process Safety Progress,
22(2), 8398.
Kletz, T. (1991). Plant design for safety: a user-friendly approach.
Hemisphere New York, New York: Publishing Corporation.
Mansfield, D., & Cassidy, K. (1994). Inherently safer approach to plant
design: the benefits of an inherently safer approach and how this can
be built into the design process. Institution of Chemical Engineer
Symposium Series, 134, 285299.
Palaniappan, C., Srinivasan, R., & Tan, R. (2004). Selection of inherently
safer process routes: a case study. Chemical Engineering and Proces-
sing, 43, 647653.
Peters, M. S., & Timmerhaus, K. D. (1991). Plant design and economics for
chemical engineers (4th ed). New York: McGraw-Hill, Inc.
Seider, W. D., Seader, J. D., & Lewin, D. R. (2004). Product & process
design principles: synthesis, analysis and evaluation (2nd Ed). NewYork: Wiley.
Smith, R. (1995). Chemical process design. New York: McGraw-Hill.
Wang, Y. (2004). Development of computer-aided fault tree synthesis
methodology for quantitative risk analysis in the chemical process
industry. Ph.D.Dissertation, A&M University, College Station, TX.
ARTICLE IN PRESS
J. Suardin et al. / Journal of Loss Prevention in the Process Industries 20 (2007) 799090
top related