coulson & richardson’s - dl4a.orgdl4a.org/uploads/pdf/coulson richardson's chemical engineering...

Coulson & Richardsons

CHEMICAL ENGINEERING

VOLUME 6

Coulson & Richardsons Chemical Engineering

Chemical Engineering, Volume 1, Sixth editionFluid Flow, Heat Transfer and Mass TransferJ. M. Coulson and J. F. Richardsonwith J. R. Backhurst and J. H. Harker

Chemical Engineering, Volume 2, Fifth editionParticle Technology and Separation ProcessesJ. F. Richardson and J. H. Harkerwith J. R. Backhurst

Chemical Engineering, Volume 3, Third editionChemical & Biochemical Reactors & Process ControlEdited by J. F. Richardson and D. G. Peacock

Chemical Engineering, Second editionSolutions to the Problems in Volume 1J. R. Backhurst and J. H. Harker with J. F. Richardson

Chemical Engineering, Solutions to the Problemsin Volumes 2 and 3J. R. Backhurst and J. H. Harker with J. F. Richardson

Chemical Engineering, Volume 6, Fourth editionChemical Engineering DesignR. K. Sinnott

Coulson & Richardsons

CHEMICAL ENGINEERING

VOLUME 6FOURTH EDITION

Chemical Engineering Design

R. K. SINNOTT

AMSTERDAM BOSTON HEIDELBERG LONDON NEW YORK OXFORDPARIS SAN DIEGO SAN FRANCISCO SINGAPORE SYDNEY TOKYO

Elsevier Butterworth-HeinemannLinacre House, Jordan Hill, Oxford OX2 8DP30 Corporate Drive, MA 01803

First published 1983Second edition 1993Reprinted with corrections 1994Reprinted with revisions 1996Third edition 1999Reprinted 2001, 2003Fourth edition 2005

Copyright 1993, 1996, 1999, 2005 R. K. Sinnott. All rights reserved

The right of R. K. Sinnott to be identified as the author of this workhas been asserted in accordance with the Copyright, Designs andPatents Act 1988

No part of this publication may be reproduced in any material form (includingphotocopying or storing in any medium by electronic means and whetheror not transiently or incidentally to some other use of this publication) withoutthe written permission of the copyright holder except in accordance with theprovisions of the Copyright, Designs and Patents Act 1988 or under the terms ofa licence issued by the Copyright Licensing Agency Ltd, 90 Tottenham Court Road,London, England W1T 4LP. Applications for the copyright holders writtenpermission to reproduce any part of this publication should be addressedto the publisher

Permissions may be sought directly from Elseviers Science & Technology RightsDepartment in Oxford, UK: phone: (C44) (0)1865 843830; fax: (C44) (0)1865 853333;e-mail: [email protected]. You may also complete your request on-line viathe Elsevier homepage (http://www.elsevier.com), by selecting Customer Supportand then Obtaining Permissions

British Library Cataloguing in Publication DataA catalogue record for this book is available from the British Library

Library of Congress Cataloguing in Publication DataA catalogue record for this book is available from the Library of Congress

ISBN 0 7506 6538 6

For information on all Elsevier Butterworth-Heinemannpublications visit our website at http://books.elsevier.com

Typeset by Laserwords Private Limited, Chennai, India

Contents

PREFACE TO FOURTH EDITION xvii

PREFACE TO THIRD EDITION xx

PREFACE TO SECOND EDITION xxi

PREFACE TO FIRST EDITION xxiii

SERIES EDITORS PREFACE xxiv

ACKNOWLEDGEMENT xxv

1 Introduction to Design 1

1.1 Introduction 11.2 Nature of design 1

1.2.1 The design objective (the need) 31.2.2 Data collection 31.2.3 Generation of possible design solutions 31.2.4 Selection 4

1.3 The anatomy of a chemical manufacturing process 51.3.1 Continuous and batch processes 7

1.4 The organisation of a chemical engineering project 71.5 Project documentation 101.6 Codes and standards 121.7 Factors of safety (design factors) 131.8 Systems of units 141.9 Degrees of freedom and design variables. The mathematical representation

of the design problem 151.9.1 Information flow and design variables 151.9.2 Selection of design variables 191.9.3 Information flow and the structure of design problems 20

1.10 Optimisation 241.10.1 General procedure 251.10.2 Simple models 251.10.3 Multiple variable problems 271.10.4 Linear programming 291.10.5 Dynamic programming 291.10.6 Optimisation of batch and semicontinuous processes 29

1.11 References 301.12 Nomenclature 311.13 Problems 32

2 Fundamentals of Material Balances 34

2.1 Introduction 342.2 The equivalence of mass and energy 342.3 Conservation of mass 342.4 Units used to express compositions 352.5 Stoichiometry 36

v

vi CONTENTS

2.6 Choice of system boundary 372.7 Choice of basis for calculations 402.8 Number of independent components 402.9 Constraints on flows and compositions 412.10 General algebraic method 422.11 Tie components 442.12 Excess reagent 462.13 Conversion and yield 472.14 Recycle processes 502.15 Purge 522.16 By-pass 532.17 Unsteady-state calculations 542.18 General procedure for material-balance problems 562.19 References (Further Reading) 572.20 Nomenclature 572.21 Problems 57

3 Fundamentals of Energy Balances (and Energy Utilisation) 60

3.1 Introduction 603.2 Conservation of energy 603.3 Forms of energy (per unit mass of material) 61

3.3.1 Potential energy 613.3.2 Kinetic energy 613.3.3 Internal energy 613.3.4 Work 613.3.5 Heat 623.3.6 Electrical energy 62

3.4 The energy balance 623.5 Calculation of specific enthalpy 673.6 Mean heat capacities 683.7 The effect of pressure on heat capacity 703.8 Enthalpy of mixtures 71

3.8.1 Integral heats of solution 723.9 Enthalpy-concentration diagrams 733.10 Heats of reaction 75

3.10.1 Effect of pressure on heats of reaction 773.11 Standard heats of formation 793.12 Heats of combustion 803.13 Compression and expansion of gases 81

3.13.1 Mollier diagrams 823.13.2 Polytropic compression and expansion 843.13.3 Multistage compressors 903.13.4 Electrical drives 93

3.14 Energy balance calculations 933.15 Unsteady state energy balances 993.16 Energy recovery 101

3.16.1 Heat exchange 1013.16.2 Heat-exchanger networks 1013.16.3 Waste-heat boilers 1023.16.4 High-temperature reactors 1033.16.5 Low-grade fuels 1053.16.6 High-pressure process streams 1073.16.7 Heat pumps 110

3.17 Process integration and pinch technology 1113.17.1 Pinch technology 1113.17.2 The problem table method 1153.17.3 The heat exchanger network 1173.17.4 Minimum number of exchangers 1213.17.5 Threshold problems 123

CONTENTS vii

3.17.6 Multiple pinches and multiple utilities 1243.17.7 Process integration: integration of other process operations 124


4 Flow-sheeting 1334.1 Introduction 1334.2 Flow-sheet presentation 133

4.2.1 Block diagrams 1344.2.2 Pictorial representation 1344.2.3 Presentation of stream flow-rates 1344.2.4 Information to be included 1354.2.5 Layout 1394.2.6 Precision of data 1394.2.7 Basis of the calculation 1404.2.8 Batch processes 1404.2.9 Services (utilities) 1404.2.10 Equipment identification 1404.2.11 Computer aided drafting 140

4.3 Manual flow-sheet calculations 1414.3.1 Basis for the flow-sheet calculations 1424.3.2 Flow-sheet calculations on individual units 143

4.4 Computer-aided flow-sheeting 1684.5 Full steady-state simulation programs 168

4.5.1 Information flow diagrams 1714.6 Manual calculations with recycle streams 172

4.6.1 The split-fraction concept 1724.6.2 Illustration of the method 1764.6.3 Guide rules for estimating split-fraction coefficients 185


5 Piping and Instrumentation 1945.1 Introduction 1945.2 The P and I diagram 194

5.2.1 Symbols and layout 1955.2.2 Basic symbols 195

5.3 Valve selection 1975.4 Pumps 199

5.4.1 Pump selection 1995.4.2 Pressure drop in pipelines 2015.4.3 Power requirements for pumping liquids 2065.4.4 Characteristic curves for centrifugal pumps 2085.4.5 System curve (operating line) 2105.4.6 Net positive suction head (NPSH) 2125.4.7 Pump and other shaft seals 213

5.5 Mechanical design of piping systems 2165.5.1 Wall thickness: pipe schedule 2165.5.2 Pipe supports 2175.5.3 Pipe fittings 2175.5.4 Pipe stressing 2175.5.5 Layout and design 218

5.6 Pipe size selection 2185.7 Control and instrumentation 227

5.7.1 Instruments 2275.7.2 Instrumentation and control objectives 2275.7.3 Automatic-control schemes 228

viii CONTENTS

5.8 Typical control systems 2295.8.1 Level control 2295.8.2 Pressure control 2295.8.3 Flow control 2295.8.4 Heat exchangers 2305.8.5 Cascade control 2315.8.6 Ratio control 2315.8.7 Distillation column control 2315.8.8 Reactor control 233

5.9 Alarms and safety trips, and interlocks 2355.10 Computers and microprocessors in process control 2365.11 References 2385.12 Nomenclature 2395.13 Problems 240

6 Costing and Project Evaluation 2436.1 Introduction 2436.2 Accuracy and purpose of capital cost estimates 2436.3 Fixed and working capital 2446.4 Cost escalation (inflation) 2456.5 Rapid capital cost estimating methods 247

6.5.1 Historical costs 2476.5.2 Step counting methods 249

6.6 The factorial method of cost estimation 2506.6.1 Lang factors 2516.6.2 Detailed factorial estimates 251

6.7 Estimation of purchased equipment costs 2536.8 Summary of the factorial method 2606.9 Operating costs 260

6.9.1 Estimation of operating costs 2616.10 Economic evaluation of projects 270

6.10.1 Cash flow and cash-flow diagrams 2706.10.2 Tax and depreciation 2726.10.3 Discounted cash flow (time value of money) 2726.10.4 Rate of return calculations 2736.10.5 Discounted cash-flow rate of return (DCFRR) 2736.10.6 Pay-back time 2746.10.7 Allowing for inflation 2746.10.8 Sensitivity analysis 2746.10.9 Summary 275

6.11 Computer methods for costing and project evaluation 2786.12 References 2796.13 Nomenclature 2796.14 Problems 280

7 Materials of Construction 2847.1 Introduction 2847.2 Material properties 2847.3 Mechanical properties 285

7.3.1 Tensile strength 2857.3.2 Stiffness 2857.3.3 Toughness 2867.3.4 Hardness 2867.3.5 Fatigue 2867.3.6 Creep 2877.3.7 Effect of temperature on the mechanical properties 287

7.4 Corrosion resistance 2877.4.1 Uniform corrosion 2887.4.2 Galvanic corrosion 289

CONTENTS ix

7.4.3 Pitting 2907.4.4 Intergranular corrosion 2907.4.5 Effect of stress 2907.4.6 Erosion-corrosion 2917.4.7 High-temperature oxidation 2917.4.8 Hydrogen embrittlement 292

7.5 Selection for corrosion resistance 2927.6 Material costs 2937.7 Contamination 294

7.7.1 Surface finish 2957.8 Commonly used materials of construction 295

7.8.1 Iron and steel 2957.8.2 Stainless steel 2967.8.3 Nickel 2987.8.4 Monel 2997.8.5 Inconel 2997.8.6 The Hastelloys 2997.8.7 Copper and copper alloys 2997.8.8 Aluminium and its alloys 2997.8.9 Lead 3007.8.10 Titanium 3007.8.11 Tantalum 3007.8.12 Zirconium 3007.8.13 Silver 3017.8.14 Gold 3017.8.15 Platinum 301

7.9 Plastics as materials of construction for chemical plant 3017.9.1 Poly-vinyl chloride (PVC) 3027.9.2 Polyolefines 3027.9.3 Polytetrafluroethylene (PTFE) 3027.9.4 Polyvinylidene fluoride (PVDF) 3027.9.5 Glass-fibre reinforced plastics (GRP) 3027.9.6 Rubber 303

7.10 Ceramic materials (silicate materials) 3037.10.1 Glass 3047.10.2 Stoneware 3047.10.3 Acid-resistant bricks and tiles 3047.10.4 Refractory materials (refractories) 304

7.11 Carbon 3057.12 Protective coatings 3057.13 Design for corrosion resistance 3057.14 References 3057.15 Nomenclature 3077.16 Problems 307

8 Design Information and Data 309

8.1 Introduction 3098.2 Sources of information on manufacturing processes 3098.3 General sources of physical properties 3118.4 Accuracy required of engineering data 3128.5 Prediction of physical properties 3138.6 Density 314

8.6.1 Liquids 3148.6.2 Gas and vapour density (specific volume) 315

8.7 Viscosity 3168.7.1 Liquids 3168.7.2 Gases 320

8.8 Thermal conductivity 3208.8.1 Solids 3208.8.2 Liquids 321

x CONTENTS

8.8.3 Gases 3218.8.4 Mixtures 322

8.9 Specific heat capacity 3228.9.1 Solids and liquids 3228.9.2 Gases 325

8.10 Enthalpy of vaporisation (latent heat) 3288.10.1 Mixtures 329

8.11 Vapour pressure 3308.12 Diffusion coefficients (diffusivities) 331

8.12.1 Gases 3318.12.2 Liquids 333

8.13 Surface tension 3358.13.1 Mixtures 335

8.14 Critical constants 3368.15 Enthalpy of reaction and enthalpy of formation 3398.16 Phase equilibrium data 339

8.16.1 Experimental data 3398.16.2 Phase equilibria 3398.16.3 Equations of state 3418.16.4 Correlations for liquid phase activity coefficients 3428.16.5 Prediction of vapour-liquid equilibria 3468.16.6 K -values for hydrocarbons 3488.16.7 Sour-water systems (Sour) 3488.16.8 Vapour-liquid equilibria at high pressures 3488.16.9 Liquid-liquid equilibria 3488.16.10 Choice of phase equilibria for design calculations 3508.16.11 Gas solubilities 3518.16.12 Use of equations of state to estimate specific enthalpy and density 353


9 Safety and Loss Prevention 360

9.1 Introduction 3609.2 Intrinsic and extrinsic safety 3619.3 The hazards 361

9.3.1 Toxicity 3619.3.2 Flammability 3639.3.3 Explosions 3659.3.4 Sources of ignition 3669.3.5 Ionising radiation 3689.3.6 Pressure 3689.3.7 Temperature deviations 3699.3.8 Noise 370

9.4 Dow fire and explosion index 3719.4.1 Calculation of the Dow F & EI 3719.4.2 Potential loss 3759.4.3 Basic preventative and protective measures 3779.4.4 Mond fire, explosion, and toxicity index 3789.4.5 Summary 379

9.5 Hazard and operability studies 3819.5.1 Basic principles 3829.5.2 Explanation of guide words 3839.5.3 Procedure 384

9.6 Hazard analysis 3899.7 Acceptable risk and safety priorities 3909.8 Safety check lists 3929.9 Major hazards 394

9.9.1 Computer software for quantitative risk analysis 395

CONTENTS xi

9.10 References 3969.11 Problems 398

10 Equipment Selection, Specification and Design 400

10.1 Introduction 40010.2 Separation processes 40110.3 Solid-solid separations 401

10.3.1 Screening (sieving) 40110.3.2 Liquid-solid cyclones 40410.3.3 Hydroseparators and sizers (classifiers) 40510.3.4 Hydraulic jigs 40510.3.5 Tables 40510.3.6 Classifying centrifuges 40610.3.7 Dense-medium separators (sink and float processes) 40610.3.8 Flotation separators (froth-flotation) 40710.3.9 Magnetic separators 40710.3.10 Electrostatic separators 408

10.4 Liquid-solid (solid-liquid) separators 40810.4.1 Thickeners and clarifiers 40810.4.2 Filtration 40910.4.3 Centrifuges 41510.4.4 Hydrocyclones (liquid-cyclones) 42210.4.5 Pressing (expression) 42610.4.6 Solids drying 426

10.5 Separation of dissolved solids 43410.5.1 Evaporators 43410.5.2 Crystallisation 437

10.6 Liquid-liquid separation 44010.6.1 Decanters (settlers) 44010.6.2 Plate separators 44510.6.3 Coalescers 44510.6.4 Centrifugal separators 446

10.7 Separation of dissolved liquids 44610.7.1 Solvent extraction and leaching 447

10.8 Gas-solids separations (gas cleaning) 44810.8.1 Gravity settlers (settling chambers) 44810.8.2 Impingement separators 44810.8.3 Centrifugal separators (cyclones) 45010.8.4 Filters 45810.8.5 Wet scrubbers (washing) 45910.8.6 Electrostatic precipitators 459

10.9 Gas liquid separators 46010.9.1 Settling velocity 46110.9.2 Vertical separators 46110.9.3 Horizontal separators 463

10.10 Crushing and grinding (comminution) equipment 46510.11 Mixing equipment 468

10.11.1 Gas mixing 46810.11.2 Liquid mixing 46810.11.3 Solids and pastes 476

10.12 Transport and storage of materials 47610.12.1 Gases 47710.12.2 Liquids 47910.12.3 Solids 481

10.13 Reactors 48210.13.1 Principal types of reactor 48310.13.2 Design procedure 486


xii CONTENTS

11 Separation Columns (Distillation, Absorption and Extraction) 49311.1 Introduction 49311.2 Continuous distillation: process description 494

11.2.1 Reflux considerations 49511.2.2 Feed-point location 49611.2.3 Selection of column pressure 496

11.3 Continuous distillation: basic principles 49711.3.1 Stage equations 49711.3.2 Dew points and bubble points 49811.3.3 Equilibrium flash calculations 499

11.4 Design variables in distillation 50111.5 Design methods for binary systems 503

11.5.1 Basic equations 50311.5.2 McCabe-Thiele method 50511.5.3 Low product concentrations 50711.5.4 The Smoker equations 512

11.6 Multicomponent distillation: general considerations 51511.6.1 Key components 51611.6.2 Number and sequencing of columns 517

11.7 Multicomponent distillation: short-cut methods for stage and reflux requirements 51711.7.1 Pseudo-binary systems 51811.7.2 Smith-Brinkley method 52211.7.3 Empirical correlations 52311.7.4 Distribution of non-key components (graphical method) 526

11.8 Multicomponent systems: rigorous solution procedures (computer methods) 54211.8.1 Lewis-Matheson method 54311.8.2 Thiele-Geddes method 54411.8.3 Relaxation methods 54511.8.4 Linear algebra methods 545

11.9 Other distillation systems 54611.9.1 Batch distillation 54611.9.2 Steam distillation 54611.9.3 Reactive distillation 547

11.10 Plate efficiency 54711.10.1 Prediction of plate efficiency 54811.10.2 OConnells correlation 55011.10.3 Van Winkles correlation 55211.10.4 AIChE method 55311.10.5 Entrainment 556

11.11 Approximate column sizing 55711.12 Plate contactors 557

11.12.1 Selection of plate type 56011.12.2 Plate construction 561

11.13 Plate hydraulic design 56511.13.1 Plate-design procedure 56711.13.2 Plate areas 56711.13.3 Diameter 56711.13.4 Liquid-flow arrangement 56911.13.5 Entrainment 57011.13.6 Weep point 57111.13.7 Weir liquid crest 57211.13.8 Weir dimensions 57211.13.9 Perforated area 57211.13.10 Hole size 57311.13.11 Hole pitch 57411.13.12 Hydraulic gradient 57411.13.13 Liquid throw 57511.13.14 Plate pressure drop 57511.13.15 Downcomer design [back-up] 577

11.14 Packed columns 58711.14.1 Types of packing 589

CONTENTS xiii

11.14.2 Packed-bed height 59311.14.3 Prediction of the height of a transfer unit (HTU) 59711.14.4 Column diameter (capacity) 60211.14.5 Column internals 60911.14.6 Wetting rates 616

11.15 Column auxiliaries 61611.16 Solvent extraction (liquid liquid extraction) 617

11.16.1 Extraction equipment 61711.16.2 Extractor design 61811.16.3 Extraction columns 62311.16.4 Supercritical fluid extraction 624


12 Heat-transfer Equipment 634

12.1 Introduction 63412.2 Basic design procedure and theory 635

12.2.1 Heat exchanger analysis: the effectiveness NTU method 63612.3 Overall heat-transfer coefficient 63612.4 Fouling factors (dirt factors) 63812.5 Shell and tube exchangers: construction details 640

12.5.1 Heat-exchanger standards and codes 64412.5.2 Tubes 64512.5.3 Shells 64712.5.4 Tube-sheet layout (tube count) 64712.5.5 Shell types (passes) 64912.5.6 Shell and tube designation 64912.5.7 Baffles 65012.5.8 Support plates and tie rods 65212.5.9 Tube sheets (plates) 65212.5.10 Shell and header nozzles (branches) 65312.5.11 Flow-induced tube vibrations 653

12.6 Mean temperature difference (temperature driving force) 65512.7 Shell and tube exchangers: general design considerations 660

12.7.1 Fluid allocation: shell or tubes 66012.7.2 Shell and tube fluid velocities 66012.7.3 Stream temperatures 66112.7.4 Pressure drop 66112.7.5 Fluid physical properties 661

12.8 Tube-side heat-transfer coefficient and pressure drop (single phase) 66212.8.1 Heat transfer 66212.8.2 Tube-side pressure drop 666

12.9 Shell-side heat-transfer and pressure drop (single phase) 66912.9.1 Flow pattern 66912.9.2 Design methods 67012.9.3 Kerns method 67112.9.4 Bells method 69312.9.5 Shell and bundle geometry 70212.9.6 Effect of fouling on pressure drop 70512.9.7 Pressure-drop limitations 705

12.10 Condensers 70912.10.1 Heat-transfer fundamentals 71012.10.2 Condensation outside horizontal tubes 71012.10.3 Condensation inside and outside vertical tubes 71112.10.4 Condensation inside horizontal tubes 71612.10.5 Condensation of steam 71712.10.6 Mean temperature difference 71712.10.7 Desuperheating and sub-cooling 717

xiv CONTENTS

12.10.8 Condensation of mixtures 71912.10.9 Pressure drop in condensers 723

12.11 Reboilers and vaporisers 72812.11.1 Boiling heat-transfer fundamentals 73112.11.2 Pool boiling 73212.11.3 Convective boiling 73512.11.4 Design of forced-circulation reboilers 74012.11.5 Design of thermosyphon reboilers 74112.11.6 Design of kettle reboilers 750

12.12 Plate heat exchangers 75612.12.1 Gasketed plate heat exchangers 75612.12.2 Welded plate 76412.12.3 Plate-fin 76412.12.4 Spiral heat exchangers 765

12.13 Direct-contact heat exchangers 76612.14 Finned tubes 76712.15 Double-pipe heat exchangers 76812.16 Air-cooled exchangers 76912.17 Fired heaters (furnaces and boilers) 769

12.17.1 Basic construction 77012.17.2 Design 77112.17.3 Heat transfer 77212.17.4 Pressure drop 77412.17.5 Process-side heat transfer and pressure drop 77412.17.6 Stack design 77412.17.7 Thermal efficiency 775

12.18 Heat transfer to vessels 77512.18.1 Jacketed vessels 77512.18.2 Internal coils 77712.18.3 Agitated vessels 778


13 Mechanical Design of Process Equipment 79413.1 Introduction 794

13.1.1 Classification of pressure vessels 79513.2 Pressure vessel codes and standards 79513.3 Fundamental principles and equations 796

13.3.1 Principal stresses 79613.3.2 Theories of failure 79713.3.3 Elastic stability 79813.3.4 Membrane stresses in shells of revolution 79813.3.5 Flat plates 80513.3.6 Dilation of vessels 80913.3.7 Secondary stresses 809

13.4 General design considerations: pressure vessels 81013.4.1 Design pressure 81013.4.2 Design temperature 81013.4.3 Materials 81113.4.4 Design stress (nominal design strength) 81113.4.5 Welded joint efficiency, and construction categories 81213.4.6 Corrosion allowance 81313.4.7 Design loads 81413.4.8 Minimum practical wall thickness 814

13.5 The design of thin-walled vessels under internal pressure 81513.5.1 Cylinders and spherical shells 81513.5.2 Heads and closures 81513.5.3 Design of flat ends 81713.5.4 Design of domed ends 81813.5.5 Conical sections and end closures 819

CONTENTS xv

13.6 Compensation for openings and branches 82213.7 Design of vessels subject to external pressure 825

13.7.1 Cylindrical shells 82513.7.2 Design of stiffness rings 82813.7.3 Vessel heads 829

13.8 Design of vessels subject to combined loading 83113.8.1 Weight loads 83513.8.2 Wind loads (tall vessels) 83713.8.3 Earthquake loading 83913.8.4 Eccentric loads (tall vessels) 84013.8.5 Torque 841

13.9 Vessel supports 84413.9.1 Saddle supports 84413.9.2 Skirt supports 84813.9.3 Bracket supports 856

13.10 Bolted flanged joints 85813.10.1 Types of flange, and selection 85813.10.2 Gaskets 85913.10.3 Flange faces 86113.10.4 Flange design 86213.10.5 Standard flanges 865

13.11 Heat-exchanger tube-plates 86713.12 Welded joint design 86913.13 Fatigue assessment of vessels 87213.14 Pressure tests 87213.15 High-pressure vessels 873

13.15.1 Fundamental equations 87313.15.2 Compound vessels 87713.15.3 Autofrettage 878

13.16 Liquid storage tanks 87913.17 Mechanical design of centrifuges 879

13.17.1 Centrifugal pressure 87913.17.2 Bowl and spindle motion: critical speed 881


14 General Site Considerations 892

14.1 Introduction 89214.2 Plant location and site selection 89214.3 Site layout 89414.4 Plant layout 896

14.4.1 Techniques used in site and plant layout 89714.5 Utilities 90014.6 Environmental considerations 902

14.6.1 Waste management 90214.6.2 Noise 90514.6.3 Visual impact 90514.6.4 Legislation 90514.6.5 Environmental auditing 906

14.7 References 906

APPENDIX A: GRAPHICAL SYMBOLS FOR PIPING SYSTEMS AND PLANT 908

APPENDIX B: CORROSION CHART 917

APPENDIX C: PHYSICAL PROPERTY DATA BANK 937

APPENDIX D: CONVERSION FACTORS FOR SOME COMMON SI UNITS 958

xvi CONTENTS

APPENDIX E: STANDARD FLANGES 960

APPENDIX F: DESIGN PROJECTS 965

APPENDIX G: EQUIPMENT SPECIFICATION (DATA) SHEETS 990

APPENDIX H: TYPICAL SHELL AND TUBE HEAT EXCHANGER TUBE-SHEET LAYOUTS 1002

AUTHOR INDEX 1007

SUBJECT INDEX 1017

CHAPTER 1

Introduction to Design

1.1. INTRODUCTION

This chapter is an introduction to the nature and methodology of the design process, andits application to the design of chemical manufacturing processes.

1.2. NATURE OF DESIGN

This section is a general, somewhat philosophical, discussion of the design process; how adesigner works. The subject of this book is chemical engineering design, but the method-ology of design described in this section applies equally to other branches of engineeringdesign.

Design is a creative activity, and as such can be one of the most rewarding and satisfyingactivities undertaken by an engineer. It is the synthesis, the putting together, of ideas toachieve a desired purpose. The design does not exist at the commencement of the project.The designer starts with a specific objective in mind, a need, and by developing andevaluating possible designs, arrives at what he considers the best way of achieving thatobjective; be it a better chair, a new bridge, or for the chemical engineer, a new chemicalproduct or a stage in the design of a production process.

When considering possible ways of achieving the objective the designer will beconstrained by many factors, which will narrow down the number of possible designs;but, there will rarely be just one possible solution to the problem, just one design. Severalalternative ways of meeting the objective will normally be possible, even several bestdesigns, depending on the nature of the constraints.

These constraints on the possible solutions to a problem in design arise in many ways.Some constraints will be fixed, invariable, such as those that arise from physical laws,government regulations, and standards. Others will be less rigid, and will be capable ofrelaxation by the designer as part of his general strategy in seeking the best design. Theconstraints that are outside the designers influence can be termed the external constraints.These set the outer boundary of possible designs; as shown in Figure 1.1. Within thisboundary there will be a number of plausible designs bounded by the other constraints,the internal constraints, over which the designer has some control; such as, choice ofprocess, choice of process conditions, materials, equipment.

Economic considerations are obviously a major constraint on any engineering design:plants must make a profit.

Time will also be a constraint. The time available for completion of a design willusually limit the number of alternative designs that can be considered.

1

2 CHEMICAL ENGINEERING

Plausibledesigns

Government controls

Eco

nom

ic c

onst

rain

ts

Safe

ty re

gulat

ions

Resources

Physical law

sStandards and codes

Personnel

Materials

ProcessconditionsCh

oice o

f

proces

s

Met

hods

Time

External constraintsInternal constraints

Region of all designs

Possible designs

Figure 1.1. Design constraints

Objective(design

specification)

Collection of data,physical

properties designmethods

Generation ofpossible designs

Selection andevaluation

(optimisation)

Finaldesign

Figure 1.2. The design process

The stages in the development of a design, from the initial identification of the objectiveto the final design, are shown diagrammatically in Figure 1.2. Each stage is discussed inthe following sections.

Figure 1.2 shows design as an iterative procedure; as the design develops the designerwill be aware of more possibilities and more constraints, and will be constantly seekingnew data and ideas, and evaluating possible design solutions.

INTRODUCTION TO DESIGN 3

1.2.1. The design objective (the need)

Chaddock (1975) defined design as, the conversion of an ill-defined requirement into asatisfied customer.

The designer is creating a design for an article, or a manufacturing process, to fulfil aparticular need. In the design of a chemical process, the need is the public need for theproduct, the commercial opportunity, as foreseen by the sales and marketing organisation.Within this overall objective the designer will recognise sub-objectives; the requirementsof the various units that make up the overall process.

Before starting work the designer should obtain as complete, and as unambiguous, astatement of the requirements as possible. If the requirement (need) arises from outside thedesign group, from a client or from another department, then he will have to elucidate thereal requirements through discussion. It is important to distinguish between the real needsand the wants. The wants are those parts of the initial specification that may be thoughtdesirable, but which can be relaxed if required as the design develops. For example, aparticular product specification may be considered desirable by the sales department, butmay be difficult and costly to obtain, and some relaxation of the specification may bepossible, producing a saleable but cheaper product. Whenever he is in a position to do so,the designer should always question the design requirements (the project and equipmentspecifications) and keep them under review as the design progresses.

Where he writes specifications for others, such as for the mechanical design or purchaseof a piece of equipment, he should be aware of the restrictions (constraints) he is placingon other designers. A tight, well-thought-out, comprehensive, specification of the require-ments defines the external constraints within which the other designers must work.

1.2.2. Data collection

To proceed with a design, the designer must first assemble all the relevant facts anddata required. For process design this will include information on possible processes,equipment performance, and physical property data. This stage can be one of the mosttime consuming, and frustrating, aspects of design. Sources of process information andphysical properties are reviewed in Chapter 8.

Many design organisations will prepare a basic data manual, containing all the processknow-how on which the design is to be based. Most organisations will have designmanuals covering preferred methods and data for the more frequently used, routine, designprocedures.

The national standards are also sources of design methods and data; they are also designconstraints.

The constraints, particularly the external constraints, should be identified early in thedesign process.

1.2.3. Generation of possible design solutions

The creative part of the design process is the generation of possible solutions to theproblem (ways of meeting the objective) for analysis, evaluation and selection. In thisactivity the designer will largely rely on previous experience, his own and that of others.


It is doubtful if any design is entirely novel. The antecedence of most designs can usuallybe easily traced. The first motor cars were clearly horse-drawn carriages without thehorse; and the development of the design of the modern car can be traced step by stepfrom these early prototypes. In the chemical industry, modern distillation processes havedeveloped from the ancient stills used for rectification of spirits; and the packed columnsused for gas absorption have developed from primitive, brushwood-packed towers. So,it is not often that a process designer is faced with the task of producing a design for acompletely novel process or piece of equipment.

The experienced engineer will wisely prefer the tried and tested methods, rather thanpossibly more exciting but untried novel designs. The work required to develop newprocesses, and the cost, is usually underestimated. Progress is made more surely in smallsteps. However, whenever innovation is wanted, previous experience, through prejudice,can inhibit the generation and acceptance of new ideas; the not invented here syndrome.

The amount of work, and the way it is tackled, will depend on the degree of noveltyin a design project.

Chemical engineering projects can be divided into three types, depending on the noveltyinvolved:

1. Modifications, and additions, to existing plant; usually carried out by the plant designgroup.

2. New production capacity to meet growing sales demand, and the sale of establishedprocesses by contractors. Repetition of existing designs, with only minor designchanges.

3. New processes, developed from laboratory research, through pilot plant, to acommercial process. Even here, most of the unit operations and process equipmentwill use established designs.

The first step in devising a new process design will be to sketch out a rough blockdiagram showing the main stages in the process; and to list the primary function (objective)and the major constraints for each stage. Experience should then indicate what types ofunit operations and equipment should be considered.

Jones (1970) discusses the methodology of design, and reviews some of the specialtechniques, such as brainstorming sessions and synectics, that have been developed tohelp generate ideas for solving intractable problems. A good general reference on the artof problem solving is the classical work by Polya (1957); see also Chittenden (1987).Some techniques for problem solving in the Chemical Industry are covered in a short textby Casey and Frazer (1984).

The generation of ideas for possible solutions to a design problem cannot be separatedfrom the selection stage of the design process; some ideas will be rejected as impracticalas soon as they are conceived.

1.2.4. Selection

The designer starts with the set of all possible solutions bounded by the externalconstraints, and by a process of progressive evaluation and selection, narrows down therange of candidates to find the best design for the purpose.


The selection process can be considered to go through the following stages:

Possible designs (credible) within the external constraints.Plausible designs (feasible) within the internal constraints.Probable designs likely candidates.Best design (optimum) judged the best solution to the problem.

The selection process will become more detailed and more refined as the design progressesfrom the area of possible to the area of probable solutions. In the early stages a coarsescreening based on common sense, engineering judgement, and rough costings will usuallysuffice. For example, it would not take many minutes to narrow down the choice of rawmaterials for the manufacture of ammonia from the possible candidates of, say, wood,peat, coal, natural gas, and oil, to a choice of between gas and oil, but a more detailedstudy would be needed to choose between oil and gas. To select the best design from theprobable designs, detailed design work and costing will usually be necessary. However,where the performance of candidate designs is likely to be close the cost of this furtherrefinement, in time and money, may not be worthwhile, particularly as there will usuallybe some uncertainty in the accuracy of the estimates.

The mathematical techniques that have been developed to assist in the optimisation ofdesigns, and plant performance, are discussed briefly in Section 1.10.

Rudd and Watson (1968) and Wells (1973) describe formal techniques for the prelim-inary screening of alternative designs.

1.3. THE ANATOMY OF A CHEMICAL MANUFACTURINGPROCESS

The basic components of a typical chemical process are shown in Figure 1.3, in whicheach block represents a stage in the overall process for producing a product from the rawmaterials. Figure 1.3 represents a generalised process; not all the stages will be needed forany particular process, and the complexity of each stage will depend on the nature of theprocess. Chemical engineering design is concerned with the selection and arrangementof the stages, and the selection, specification and design of the equipment required toperform the stage functions.

Rawmaterialstorage

Feedpreparation Reaction

Productseparation

Productpurification

Productstorage

Sales

Recycle of unreactedmaterial

By-products

Wastes

Stage 1 Stage 2 Stage 3 Stage 4 Stage 5 Stage 6

Figure 1.3. Anatomy of a chemical process

Stage 1. Raw material storage

Unless the raw materials (also called essential materials, or feed stocks) are suppliedas intermediate products (intermediates) from a neighbouring plant, some provision will


have to be made to hold several days, or weeks, storage to smooth out fluctuations andinterruptions in supply. Even when the materials come from an adjacent plant someprovision is usually made to hold a few hours, or even days, supply to decouple theprocesses. The storage required will depend on the nature of the raw materials, the methodof delivery, and what assurance can be placed on the continuity of supply. If materials aredelivered by ship (tanker or bulk carrier) several weeks stocks may be necessary; whereasif they are received by road or rail, in smaller lots, less storage will be needed.

Stage 2. Feed preparationSome purification, and preparation, of the raw materials will usually be necessary beforethey are sufficiently pure, or in the right form, to be fed to the reaction stage. For example,acetylene generated by the carbide process contains arsenical and sulphur compounds, andother impurities, which must be removed by scrubbing with concentrated sulphuric acid(or other processes) before it is sufficiently pure for reaction with hydrochloric acid toproduce dichloroethane. Liquid feeds will need to be vaporised before being fed to gas-phase reactors, and solids may need crushing, grinding and screening.

Stage 3. ReactorThe reaction stage is the heart of a chemical manufacturing process. In the reactor theraw materials are brought together under conditions that promote the production of thedesired product; invariably, by-products and unwanted compounds (impurities) will alsobe formed.

Stage 4. Product separationIn this first stage after the reactor the products and by-products are separated from anyunreacted material. If in sufficient quantity, the unreacted material will be recycled tothe reactor. They may be returned directly to the reactor, or to the feed purification andpreparation stage. The by-products may also be separated from the products at this stage.

Stage 5. PurificationBefore sale, the main product will usually need purification to meet the product specifi-cation. If produced in economic quantities, the by-products may also be purified for sale.

Stage 6. Product storageSome inventory of finished product must be held to match production with sales. Provisionfor product packaging and transport will also be needed, depending on the nature of theproduct. Liquids will normally be dispatched in drums and in bulk tankers (road, rail andsea), solids in sacks, cartons or bales.

The stock held will depend on the nature of the product and the market.

Ancillary processesIn addition to the main process stages shown in Figure 1.3, provision will have to bemade for the supply of the services (utilities) needed; such as, process water, cooling


water, compressed air, steam. Facilities will also be needed for maintenance, firefighting,offices and other accommodation, and laboratories; see Chapter 14.

1.3.1. Continuous and batch processes

Continuous processes are designed to operate 24 hours a day, 7 days a week, throughoutthe year. Some down time will be allowed for maintenance and, for some processes,catalyst regeneration. The plant attainment; that is, the percentage of the available hoursin a year that the plant operates, will usually be 90 to 95%.

Attainment % D hours operated8760

100Batch processes are designed to operate intermittently. Some, or all, the process units

being frequently shut down and started up.Continuous processes will usually be more economical for large scale production. Batch

processes are used where some flexibility is wanted in production rate or product speci-fication.

Choice of continuous versus batch productionThe choice between batch or continuous operation will not be clear cut, but the followingrules can be used as a guide.

Continuous

1. Production rate greater than 5 106 kg/h2. Single product3. No severe fouling4. Good catalyst life5. Proven processes design6. Established market

Batch

1. Production rate less than 5 106 kg/h2. A range of products or product specifications3. Severe fouling4. Short catalyst life5. New product6. Uncertain design

1.4. THE ORGANISATION OF A CHEMICAL ENGINEERINGPROJECT

The design work required in the engineering of a chemical manufacturing process can bedivided into two broad phases.

Phase 1. Process design, which covers the steps from the initial selection of the processto be used, through to the issuing of the process flow-sheets; and includes the selection,


Project specification

Initial evaluation.Process selection.Preliminary flow diagrams.

Detailed process design.Flow-sheets.Chemical engineering equipmentdesign and specifications.Reactors, Unit operations, Heat exchangers,Miscellaneous equipment.Materials selection.Process manuals

Material and energy balances.Preliminary equipment selectionand design.Process flow-sheeting.

Preliminary cost estimation.Authorisation of funds.

Piping and instrument design

Instrument selectionand specification

Pumps and compressors.Selection and specification

Vessel design Heat exchanger design Utilities and other services.Design and specification

Electrical,Motors, switch gear,substations, etc.

Piping design Structural design Plant layout General civil work.Foundations, drains,roads, etc.

Buildings.Offices, laboratories,control rooms, etc.

Project cost estimation.Capital authorisation

Purchasing/procurement Raw material specification.(contracts)

Construction

Start-up Operating manuals

Operation

Sales

Figure 1.4. The structure of a chemical engineering project


specification and chemical engineering design of equipment. In a typical organisation,this phase is the responsibility of the Process Design Group, and the work will be mainlydone by chemical engineers. The process design group may also be responsible for thepreparation of the piping and instrumentation diagrams.

Phase 2. The detailed mechanical design of equipment; the structural, civil and electricaldesign; and the specification and design of the ancillary services. These activities will bethe responsibility of specialist design groups, having expertise in the whole range ofengineering disciplines.

Other specialist groups will be responsible for cost estimation, and the purchase andprocurement of equipment and materials.

The sequence of steps in the design, construction and start-up of a typical chemicalprocess plant is shown diagrammatically in Figure 1.4 and the organisation of a typicalproject group in Figure 1.5. Each step in the design process will not be as neatly separatedfrom the others as is indicated in Figure 1.4; nor will the sequence of events be as clearlydefined. There will be a constant interchange of information between the various designsections as the design develops, but it is clear that some steps in a design must be largelycompleted before others can be started.

A project manager, often a chemical engineer by training, is usually responsible for theco-ordination of the project, as shown in Figure 1.5.

Specialist design sectionsVessels Layout Piping Heat exchangers valves fired heatersControl Civil workand instruments structures Electrical buildingsCompressorsand turbines Utilitiespumps

Process sectionProcess evaluationFlow-sheetingEquipment specifications

Construction sectionConstructionStart-up

Projectmanager

ProcurementsectionEstimatingInspectionScheduling

Figure 1.5. Project organisation

As was stated in Section 1.2.1, the project design should start with a clear specificationdefining the product, capacity, raw materials, process and site location. If the project isbased on an established process and product, a full specification can be drawn up atthe start of the project. For a new product, the specification will be developed from aneconomic evaluation of possible processes, based on laboratory research, pilot plant testsand product market research.


The organisation of chemical process design is discussed in more detail by Rase andBarrow (1964) and Baasel (1974).

Some of the larger chemical manufacturing companies have their own project designorganisations and carry out the whole project design and engineering, and possiblyconstruction, within their own organisation. More usually the design and construction, andpossibly assistance with start-up, is entrusted to one of the international contracting firms.

The operating company will often provide the know-how for the process, and willwork closely with the contractor throughout all stages of the project.

1.5. PROJECT DOCUMENTATION

As shown in Figure 1.5 and described in Section 1.4, the design and engineering ofa chemical process requires the co-operation of many specialist groups. Effective co-operation depends on effective communications, and all design organisations have formalprocedures for handling project information and documentation. The project documen-tation will include:

1. General correspondence within the design group and with:government departmentsequipment vendorssite personnelthe client

2. Calculation sheets design calculationscostingcomputer print-out

3. Drawings flow-sheetspiping and instrumentation diagramslayout diagramsplot/site plansequipment detailspiping diagramsarchitectural drawingsdesign sketches

4. Specification sheets for equipment, such as:heat exchangerspumps

5. Purchase orders quotationsinvoices

All documents should be assigned a code number for easy cross referencing, filing andretrieval.

Calculation sheets

The design engineer should develop the habit of setting out calculations so that they canbe easily understood and checked by others. It is good practice to include on calculation


sheets the basis of the calculations, and any assumptions and approximations made, insufficient detail for the methods, as well as the arithmetic, to be checked. Design calcula-tions are normally set out on standard sheets. The heading at the top of each sheet shouldinclude: the project title and identification number and, most importantly, the signature(or initials) of the person who checked the calculation.

Drawings

All project drawings are normally drawn on specially printed sheets, with the companyname; project title and number; drawing title and identification number; draughtsmansname and person checking the drawing; clearly set out in a box in the bottom right-handcorner. Provision should also be made for noting on the drawing all modifications to theinitial issue.

Drawings should conform to accepted drawing conventions, preferably those laid downby the national standards. The symbols used for flow-sheets and piping and instrumentdiagrams are discussed in Chapter 4. Drawings and sketches are normally made ondetail paper (semi-transparent) in pencil, so modifications can be easily made, and printstaken.

In most design offices Computer Aided Design (CAD) methods are now used to producethe drawings required for all the aspects of a project: flow-sheets, piping and instrumen-tation, mechanical and civil work.

Specification sheets

Standard specification sheets are normally used to transmit the information required forthe detailed design, or purchase, of equipment items; such as, heat exchangers, pumps,columns.

As well as ensuring that the information is clearly and unambiguously presented,standard specification sheets serve as check lists to ensure that all the information requiredis included.

Examples of equipment specification sheets are given in Appendix G.

Process manuals

Process manuals are often prepared by the process design group to describe the process andthe basis of the design. Together with the flow-sheets, they provide a complete technicaldescription of the process.

Operating manuals

Operating manuals give the detailed, step by step, instructions for operation of the processand equipment. They would normally be prepared by the operating company personnel,but may also be issued by a contractor as part of the contract package for a less experiencedclient. The operating manuals would be used for operator instruction and training, andfor the preparation of the formal plant operating instructions.


1.6. CODES AND STANDARDS

The need for standardisation arose early in the evolution of the modern engineeringindustry; Whitworth introduced the first standard screw thread to give a measure ofinterchangeability between different manufacturers in 1841. Modern engineering standardscover a much wider function than the interchange of parts. In engineering practicethey cover:

1. Materials, properties and compositions.2. Testing procedures for performance, compositions, quality.3. Preferred sizes; for example, tubes, plates, sections.4. Design methods, inspection, fabrication.5. Codes of practice, for plant operation and safety.

The terms STANDARD and CODE are used interchangeably, though CODE shouldreally be reserved for a code of practice covering say, a recommended design or operatingprocedure; and STANDARD for preferred sizes, compositions, etc.

All of the developed countries, and many of the developing countries, have nationalstandards organisations, responsible for the issue and maintenance of standards for themanufacturing industries, and for the protection of consumers. In the United Kingdompreparation and promulgation of national standards are the responsibility of the BritishStandards Institution (BSI). The Institution has a secretariat and a number of technicalpersonnel, but the preparation of the standards is largely the responsibility of committeesof persons from the appropriate industry, the professional engineering institutions andother interested organisations.

In the United States the government organisation responsible for coordinating infor-mation on standards is the National Bureau of Standards; standards are issued by Federal,State and various commercial organisations. The principal ones of interest to chemicalengineers are those issued by the American National Standards Institute (ANSI), theAmerican Petroleum Institute (API), the American Society for Testing Materials (ASTM),and the American Society of Mechanical Engineers (ASME) (pressure vessels). Burklin(1979) gives a comprehensive list of the American codes and standards.

The International Organization for Standardization (ISO) coordinates the publication ofinternational standards.

All the published British standards are listed, and their scope and application described,in the British Standards Institute Catalogue; which the designer should consult. Thecatalogue is available online, go to the BSI group home page, www.bsi-global.com.

As well as the various national standards and codes, the larger design organisationswill have their own (in-house) standards. Much of the detail in engineering design workis routine and repetitious, and it saves time and money, and ensures a conformity betweenprojects, if standard designs are used whenever practicable.

Equipment manufacturers also work to standards to produce standardised designs andsize ranges for commonly used items; such as electric motors, pumps, pipes and pipefittings. They will conform to national standards, where they exist, or to those issued bytrade associations. It is clearly more economic to produce a limited range of standardsizes than to have to treat each order as a special job.


For the designer, the use of a standardised component size allows for the easy integrationof a piece of equipment into the rest of the plant. For example, if a standard range ofcentrifugal pumps is specified the pump dimensions will be known, and this facilitates thedesign of the foundations plates, pipe connections and the selection of the drive motors:standard electric motors would be used.

For an operating company, the standardisation of equipment designs and sizes increasesinterchangeability and reduces the stock of spares that have to be held in maintenancestores.

Though there are clearly considerable advantages to be gained from the use of standardsin design, there are also some disadvantages. Standards impose constraints on the designer.The nearest standard size will normally be selected on completing a design calculation(rounding-up) but this will not necessarily be the optimum size; though as the standardsize will be cheaper than a special size, it will usually be the best choice from the point ofview of initial capital cost. Standard design methods must, of their nature, be historical,and do not necessarily incorporate the latest techniques.

The use of standards in design is illustrated in the discussion of the pressure vesseldesign standards (codes) in Chapter 13.

1.7. FACTORS OF SAFETY (DESIGN FACTORS)

Design is an inexact art; errors and uncertainties will arise from uncertainties in the designdata available and in the approximations necessary in design calculations. To ensure thatthe design specification is met, factors are included to give a margin of safety in thedesign; safety in the sense that the equipment will not fail to perform satisfactorily, andthat it will operate safely: will not cause a hazard. Design factor is a better term to use,as it does not confuse safety and performance factors.

In mechanical and structural design, the magnitude of the design factors used to allowfor uncertainties in material properties, design methods, fabrication and operating loadsare well established. For example, a factor of around 4 on the tensile strength, or about2.5 on the 0.1 per cent proof stress, is normally used in general structural design. Theselection of design factors in mechanical engineering design is illustrated in the discussionof pressure vessel design in Chapter 13.

Design factors are also applied in process design to give some tolerance in the design.For example, the process stream average flows calculated from material balances areusually increased by a factor, typically 10 per cent, to give some flexibility in processoperation. This factor will set the maximum flows for equipment, instrumentation, andpiping design. Where design factors are introduced to give some contingency in a processdesign, they should be agreed within the project organisation, and clearly stated in theproject documents (drawings, calculation sheets and manuals). If this is not done, thereis a danger that each of the specialist design groups will add its own factor of safety;resulting in gross, and unnecessary, over-design.

When selecting the design factor to use a balance has to be made between the desireto make sure the design is adequate and the need to design to tight margins to remaincompetitive. The greater the uncertainty in the design methods and data, the bigger thedesign factor that must be used.


1.8. SYSTEMS OF UNITS

To be consistent with the other volumes in this series, SI units have been used in thisbook. However, in practice the design methods, data and standards which the designer willuse are often only available in the traditional scientific and engineering units. Chemicalengineering has always used a diversity of units; embracing the scientific CGS and MKSsystems, and both the American and British engineering systems. Those engineers in theolder industries will also have had to deal with some bizarre traditional units; such asdegrees Twaddle (density) and barrels for quantity. Desirable as it may be for industryworld-wide to adopt one consistent set of units, such as SI, this is unlikely to come aboutfor many years, and the designer must contend with whatever system, or combination ofsystems, his organisation uses. For those in the contracting industry this will also meanworking with whatever system of units the client requires.

It is usually the best practice to work through design calculations in the units in whichthe result is to be presented; but, if working in SI units is preferred, data can be convertedto SI units, the calculation made, and the result converted to whatever units are required.Conversion factors to the SI system from most of the scientific and engineering units usedin chemical engineering design are given in Appendix D.

Some license has been taken in the use of the SI system in this volume. Temperatures aregiven in degrees Celsius (C); degrees Kelvin are only used when absolute temperatureis required in the calculation. Pressures are often given in bar (or atmospheres) ratherthan in the Pascals (N/m2), as this gives a better feel for the magnitude of the pressures.In technical calculations the bar can be taken as equivalent to an atmosphere, whateverdefinition is used for atmosphere. The abbreviations bara and barg are often used to denotebar absolute and bar gauge; analogous to psia and psig when the pressure is expressedin pound force per square inch. When bar is used on its own, without qualification, it isnormally taken as absolute.

For stress, N/mm2 have been used, as these units are now generally accepted byengineers, and the use of a small unit of area helps to indicate that stress is the intensity offorce at a point (as is also pressure). For quantity, kmol are generally used in preferenceto mol, and for flow, kmol/h instead of mol/s, as this gives more sensibly sized figures,which are also closer to the more familiar lb/h.

For volume and volumetric flow, m3 and m3/h are used in preference to m3/s, whichgives ridiculously small values in engineering calculations. Litres per second are used forsmall flow-rates, as this is the preferred unit for pump specifications.

Where, for convenience, other than SI units have been used on figures or diagrams, thescales are also given in SI units, or the appropriate conversion factors are given in thetext. The answers to some examples are given in British engineering units as well as SI,to help illustrate the significance of the values.

Some approximate conversion factors to SI units are given in Table 1.1. These areworth committing to memory, to give some feel for the units for those more familiar withthe traditional engineering units. The exact conversion factors are also shown in the table.A more comprehensive table of conversion factors is given in Appendix D.

Engineers need to be aware of the difference between US gallons and imperial gallons(UK) when using American literature and equipment catalogues. Equipment quoted in an


Table 1.1. Approximate conversion units

Quantity British SI unitEng. unit approx. exact

Energy 1 Btu 1 kJ 1.05506

Specific enthalpy 1 Btu/lb 2 kJ/kg 2.326

Specific heat capacity 1 Btu/lbF 4 kJ/kgC 4.1868(CHU/lbC)

Heat transfer coeff. 1 Btu/ft2hF 6 W/m2 C 5.678(CHU/ft2hC)

Viscosity 1 centipoise 1 mNs/m2 1.0001 lbf/ft h 0.4 mNs/m2 0.4134

Surface tension 1 dyne/cm 1 mN/m 1.000

Pressure 1 lbf/in2 7 kN/m2 6.8941 atm 1 bar 1.01325

105 N/m2

Density 1 lb/ft3 16 kg/m3 16.01901 g/cm3 1 kg/m3

Volume 1 imp gal. 4.5 103 m3 4.5461 103Flow-rate 1 imp gal/m 16 m3/h 16.366

Note:1 US gallon D 0.84 imperial gallons (UK)1 barrel (oil) D 50 US gall 0.19 m3 (exact 0.1893)1 kWh D 3.6 MJ

American catalogue in US gallons or gpm (gallons per minute) will have only 80 per centof the rated capacity when measured in imperial gallons.

The electrical supply frequency in these two countries is also different: 60 Hz in the USand 50 Hz in the UK. So a pump specified as 50 gpm (US gallons), running at 1750 rpm(revolutions per second) in the US would only deliver 35 imp gpm if operated in the UK;where the motor speed would be reduced to 1460 rpm: so beware.

1.9. DEGREES OF FREEDOM AND DESIGN VARIABLES.THE MATHEMATICAL REPRESENTATION OF

THE DESIGN PROBLEMIn Section 1.2 it was shown that the designer in seeking a solution to a design problemworks within the constraints inherent in the particular problem.

In this section the structure of design problems is examined by representing the generaldesign problem in a mathematical form.

1.9.1. Information flow and design variables

A process unit in a chemical process plant performs some operation on the inlet materialstreams to produce the desired outlet streams. In the design of such a unit the designcalculations model the operation of the unit. A process unit and the design equations


Inputstreams

Inputinformation

Outputstreams

Outputinformation

Unit

Calculationmethod

Figure 1.6. The design unit

representing the unit are shown diagrammatically in Figure 1.6. In the design unit theflow of material is replaced by a flow of information into the unit and a flow of derivedinformation from the unit.

The information flows are the values of the variables which are involved in the design;such as, stream compositions, temperatures, pressure, stream flow-rates, and streamenthalpies. Composition, temperature and pressure are intensive variables: independent ofthe quantity of material (flow-rate). The constraints on the design will place restrictions onthe possible values that these variables can take. The values of some of the variables willbe fixed directly by process specifications. The values of other variables will be determinedby design relationships arising from constraints. Some of the design relationships willbe in the form of explicit mathematical equations (design equations); such as thosearising from material and energy balances, thermodynamic relationships, and equipmentperformance parameters. Other relationships will be less precise; such as those arisingfrom the use of standards and preferred sizes, and safety considerations.

The difference between the number of variables involved in a design and the numberof design relationships has been called the number of degrees of freedom; similar to theuse of the term in the phase rule. The number of variables in the system is analogous to thenumber of variables in a set of simultaneous equations, and the number of relationshipsanalogous to the number of equations. The difference between the number of variablesand equations is called the variance of the set of equations.

If Nv is the number of possible variables in a design problem and Nr the number ofdesign relationships, then the degrees of freedom Nd is given by:

Nd D Nv Nr 1.1Nd represents the freedom that the designer has to manipulate the variables to find thebest design.

If Nv D Nr,Nd D 0 and there is only one, unique, solution to the problem. The problemis not a true design problem, no optimisation is possible.

If Nv < Nr,Nd < 0, and the problem is over defined; only a trivial solution is possible.If Nv > Nr,Nd > 0, and there is an infinite number of possible solutions. However,

for a practical problem there will be only a limited number of feasible solutions. Thevalue of Nd is the number of variables which the designer must assign values to solvethe problem.

How the number of process variables, design relationships, and design variables definesa system can be best illustrated by considering the simplest system; a single-phase, processstream.


Process stream

Consider a single-phase stream, containing C components.

Variable Number

Stream flow-rate 1Composition (component concentrations) CTemperature 1Pressure 1Stream enthalpy 1

Total, Nv D CC 4Relationships between variables Number

Composition1 1Enthalpy2 1

Total, Nr D 2

Degrees of freedom Nd D Nv Nr D CC 4 2 D CC 2

(1) The sum of the mass or mol, fractions, must equal one.(2) The enthalpy is a function of stream composition, temperature and pressure.

Specifying (CC 2) variables completely defines the stream.

Flash distillation

The idea of degrees of freedom in the design process can be further illustrated by consid-ering a simple process unit, a flash distillation. (For a description of flash distillation seeVolume 2, Chapter 11).

F2, P2, T2, (xi)2

F3, P3, T3, (xi)3

F1, P1, T1, (xi)1

q

Figure 1.7. Flash distillation

The unit is shown in Figure 1.7, where:F D stream flow rate,P D pressure,T D temperature,xi D concentration, component i,q D heat input.

Suffixes, 1 D inlet, 2 D outlet vapour, 3 D outlet liquid.


Variable Number

Streams (free variables)1 3CC 21Still

pressure 1temperature 1heat input 1

Nr D 3CC 9Relationship Number

Material balances (each component) CHeat balance, overall 1v l e relationships2 CEquilibrium still restriction3 4

2CC 5

Degrees of freedom Nd D 3CC 9 2CC 5 D CC 4

(1) The degrees of freedom for each stream. The total variables in each stream could have been used, andthe stream relationships included in the count of relationships.

This shows how the degrees of freedom for a complex unit can be built up from the degrees of freedom ofits components. For more complex examples see Kwauk (1956).

(2) Given the temperature and pressure, the concentration of any component in the vapour phase can beobtained from the concentration in the liquid phase, from the vapour liquid equilibrium data for the system.

(3) The concept (definition) of an equilibrium separation implies that the outlet streams and the still are atthe same temperature and pressure. This gives four equations:

P2 D P3 D PT2 D T3 D T

Though the total degrees of freedom is seen to be (CC 4) some of the variables willnormally be fixed by general process considerations, and will not be free for the designerto select as design variables. The flash distillation unit will normally be one unit in aprocess system and the feed composition and feed conditions will be fixed by the upstreamprocesses; the feed will arise as an outlet stream from some other unit. Defining the feedfixes (CC 2) variables, so the designer is left with:

CC 4 CC 2 D 2as design variables.

Summary

The purpose of this discussion was to show that in a design there will be a certainnumber of variables that the designer must specify to define the problem, and which hecan manipulate to seek the best design. In manual calculations the designer will rarely


need to calculate the degrees of freedom in a formal way. He will usually have intuitivefeel for the problem, and can change the calculation procedure, and select the designvariables, as he works through the design. He will know by experience if the problem iscorrectly specified. A computer, however, has no intuition, and for computer-aided designcalculations it is essential to ensure that the necessary number of variables is specified todefine the problem correctly. For complex processes the number of variables and relatingequations will be very large, and the calculation of the degrees of freedom very involved.Kwauk (1956) has shown how the degrees of freedom can be calculated for separationprocesses by building up the complex unit from simpler units. Smith (1963) uses Kwauksmethod, and illustrates how the idea of degrees of freedom can be used in the designof separation processes.

1.9.2. Selection of design variables

In setting out to solve a design problem the designer has to decide which variables are tobe chosen as design variables; the ones he will manipulate to produce the best design.The choice of design variables is important; careful selection can simplify the designcalculations. This can be illustrated by considering the choice of design variables for asimple binary flash distillation.

For a flash distillation the total degrees of freedom was shown to be (CC 4), so fortwo components Nd D 6. If the feed stream flow, composition, temperature and pressureare fixed by upstream conditions, then the number of design variables will be:

N0d D 6 CC 2 D 6 4 D 2So the designer is free to select two variables from the remaining variables in order toproceed with the calculation of the outlet stream compositions and flows.

If he selects the still pressure (which for a binary system will determine the vapourliquid equilibrium relationship) and one outlet stream flow-rate, then the outlet compo-sitions can be calculated by simultaneous solution of the mass balance and equilibriumrelationships (equations). A graphical method for the simultaneous solution is given inVolume 2, Chapter 11.

However, if he selects an outlet stream composition (say the liquid stream) instead ofa flow-rate, then the simultaneous solution of the mass balance and v l e relationshipswould not be necessary. The stream compositions could be calculated by the followingstep-by-step (sequential) procedure:

1. Specifying P determines the v l e relationship (equilibrium) curve from experi-mental data.

2. Knowing the outlet liquid composition, the outlet vapour composition can be calcu-lated from the v l e relationship.

3. Knowing the feed and outlet compositions, and the feed flow-rate, the outlet streamflows can be calculated from a material balance.

4. An enthalpy balance then gives the heat input required.

The need for simultaneous solution of the design equations implies that there is arecycle of information. Choice of an outlet stream composition as a design variable in


x3F2F3T

PF2 (or F3)

Feed

Select

(a)

(b)

F3 (or F2)x2x3T

x2 (or x3)

x2 (or x3)

Direction of calculation

F1x1P1T1Px2 (or x3)

Feed

Select

Direction of calculation

F1x1P1T1

Figure 1.8. Information flow, binary flash distillation calculation (a) Information recycle (b) Information flowreversal

effect reverses the flow of information through the problem and removes the recycle; thisis shown diagrammatically in Figure 1.8.

1.9.3. Information flow and the structure of design problems

It was shown in Section 1.9.2. by studying a relatively simple problem, that the wayin which the designer selects his design variables can determine whether the designcalculations will prove to be easy or difficult. Selection of one particular set of variablescan lead to a straightforward, step-by-step, procedure, whereas selection of another setcan force the need for simultaneous solution of some of the relationships; which oftenrequires an iterative procedure (cut-and-try method). How the choice of design variables,inputs to the calculation procedure, affects the ease of solution for the general designproblem can be illustrated by studying the flow of information, using simple informationflow diagrams. The method used will be that given by Lee et al. (1966) who used a formof directed graph; a biparte graph, see Berge (1962).

The general design problem can be represented in mathematical symbolism as a seriesof equations:

fivj D 0where j D 1, 2, 3, . . . , Nv,

i D 1, 2, 3, . . . , NrConsider the following set of such equations:

f1v1, v2 D 0f2v1, v2, v3, v5 D 0


f3v1, v3, v4 D 0f4v2, v4, v5, v6 D 0f5v5, v6, v7 D 0

There are seven variables, Nv D 7, and five equations (relationships) Nr D 5, so thenumber of degrees of freedom is:

Nd D Nv Nr D 7 5 D 2

The task is to select two variables from the total of seven in such a way as to give thesimplest, most efficient, method of solution to the seven equations. There are twenty-oneways of selecting two items from seven.

In Lees method the equations and variables are represented by nodes on the bipartegraph (circles), connected by edges (lines), as shown in Figure 1.9.

f1

v1 v1

f node

v node

Figure 1.9. Nodes and edges on a biparte graph

Figure 1.9 shows that equation f1 contains (is connected to) variables v1 and v2. Thecomplete graph for the set of equations is shown in Figure 1.10.

f1 f2 f3 f4

v1 v2 v3 v4 v5 v6 v7

f5

Figure 1.10. Biparte graph for the complete set of equations

The number of edges connected to a node defines the local degree of the node p.For example, the local degree of the f1 node is 2, pf1 D 2, and at the v5 node it is 3,pv5 D 3. Assigning directions to the edges of Figure 1.10 (by putting arrows on thelines) identifies one possible order of solution for the equations. If a variable vj is definedas an output variable from an equation fi, then the direction of information flow is fromthe node fi to the node vj and all other edges will be oriented into fi. What this means,mathematically, is that assigning vj as an output from fi rearranges that equation so that:

fiv1, v2, . . . , vn D vjvj is calculated from equation fi.


The variables selected as design variables (fixed by the designer) cannot therefore beassigned as output variables from an f node. They are inputs to the system and their edgesmust be oriented into the system of equations.

If, for instance, variables v3 and v4 are selected as design variables, then Figure 1.11shows one possible order of solution of the set of equations. Different types of arrowsare used to distinguish between input and output variables, and the variables selected asdesign variables are enclosed in a double circle.

f1 f2 f3 f4 f5

v1 v2 v5 v6 v7

v3 v4

Design variables (inputs)InputsOutputs

Figure 1.11. An order of solution

Tracing the order of the solution of the equations as shown in Figure 1.11 shows howthe information flows through the system of equations:

1. Fixing v3 and v4 enables f3 to be solved, giving v1 as the output. v1 is an input tof1 and f2.

2. With v1 as an input, f1 can be solved giving v2; v2 is an input to f2 and f4.3. Knowing v3, v1 and v2, f2 can be solved to give v5; v5 is an input to f4 and f5.4. Knowing v4, v2 and v5, f4 can be solved to give v6; v6 is an input to f5.5. Knowing v6 and v5, f5 can be solved to give v7; which completes the solution.

This order of calculation can be shown more clearly by redrawing Figure 1.11 as shownin Figure 1.12.

f3 f1 f2 f4 f5v1 v2 v5 v6 v7

v3

v4

v2

v5

v3

v4

Figure 1.12. Figure 1.11 redrawn to show order of solution


With this order, the equations can be solved sequentially, with no need for the simul-taneous solution of any of the equations. The fortuitous selection of v3 and v4 as designvariables has given an efficient order of solution of the equations.

If for a set of equations an order of solution exists such that there is no need for thesimultaneous solution of any of the equations, the system is said to be acyclic, norecycle of information.

If another pair of variables had been selected, for instance v5 and v7, an acyclic orderof solution for the set of equations would not necessarily have been obtained.

For many design calculations it will not be possible to select the design variables so asto eliminate the recycle of information and obviate the need for iterative solution of thedesign relationships.

For example, the set of equations given below will be cyclic for all choices of the twopossible design variables.

f1x1, x2 D 0f2x1, x3, x4 D 0f3x2, x3, x4, x5, x6 D 0f4x4, x5, x6 D 0Nd D 6 4 D 2

The biparte graph for this example, with x3 and x5 selected as the design variables(inputs), is shown in Figure 1.13.

f1 f2 f3 f4

x6x4x2x1

x3 x5

Figure 1.13.

One strategy for the solution of this cyclic set of equations would be to guess (assigna value to) x6. The equations could then be solved sequentially, as shown in Figure 1.14,to produce a calculated value for x6, which could be compared with the assumed valueand the procedure repeated until a satisfactory convergence of the assumed and calculatedvalue had been obtained. Assigning a value to x6 is equivalent to tearing the recycleloop at x6 (Figure 1.15). Iterative methods for the solution of equations are discussed byHenley and Rosen (1969).

When a design problem cannot be reduced to an acyclic form by judicious selection ofthe design variables, the design variables should be chosen so as to reduce the recycle of


f1 f2 f3 f4

x6

x6x4x2x1

x3 x5

Assumedvalue

Calculatedvalue

Figure 1.14.

f4 f2 f1

f3

x6

x5

x3

x5

x6

x4

x4 x1

x3

x2

Figure 1.15.

information to a minimum. Lee and Rudd (1966) and Rudd and Watson (1968) give analgorithm that can be used to help in the selection of the best design variables in manualcalculations.

The recycle of information, often associated with the actual recycle of process material,will usually occur in any design problem involving large sets of equations; such as in thecomputer simulation of chemical processes. Efficient methods for the solution of sets ofequations are required in computer-aided design procedures to reduce the computer timeneeded. Several workers have published algorithms for the efficient ordering of recycleloops for iterative solution procedures, and some references to this work are given in thechapter on flow-sheeting, Chapter 4.

1.10. OPTIMISATION

Design is optimisation: the designer seeks the best, the optimum, solution to a problem.Much of the selection and choice in the design process will depend on the intuitive

judgement of the designer; who must decide when more formal optimisation techniquescan be used to advantage.

The task of formally optimising the design of a complex processing plant involvingseveral hundred variables, with complex interactions, is formidable, if not impossible.The task can be reduced by dividing the process into more manageable units, identifyingthe key variables and concentrating work where the effort involved will give the greatest


benefit. Sub-division, and optimisation of the sub-units rather than the whole, will notnecessarily give the optimum design for the whole process. The optimisation of one unitmay be at the expense of another. For example, it will usually be satisfactory to optimisethe reflux ratio for a fractionating column independently of the rest of the plant; but if thecolumn is part of a separation stage following a reactor, in which the product is separatedfrom the unreacted materials, then the design of the column will interact with, and maywell determine, the optimisation of the reactor design.

In this book the discussion of optimisation methods will, of necessity, be limited to abrief review of the main techniques used in process and equipment design. The extensiveliterature on the subject should be consulted for full details of the methods available, andtheir application and limitations; see Beightler and Wilde (1967), Beveridge and Schechter(1970), Stoecker (1989), Rudd and Watson (1968), Edgar and Himmelblau (2001). Thebooks by Rudd and Watson (1968) and Edgar and Himmelblau (2001) are particularlyrecommended to students.

1.10.1. General procedure

When setting out to optimise any system, the first step is clearly to identify the objective:the criterion to be used to judge the system performance. In engineering design theobjective will invariably be an economic one. For a chemical process, the overall objectivefor the operating company will be to maximise profits. This will give rise to sub-objectives,which the designer will work to achieve. The main sub-objective will usually be tominimise operating costs. Other sub-objectives may be to reduce investment, maximiseyield, reduce labour requirements, reduce maintenance, operate safely.

When choosing his objectives the designer must keep in mind the overall objective.Minimising cost per unit of production will not necessarily maximise profits per unit time;market factors, such as quality and delivery, may determine the best overall strategy.

The second step is to determine the objective function: the system of equations, andother relationships, which relate the objective with the variables to be manipulated tooptimise the function. If the objective is economic, it will be necessary to express theobjective function in economic terms (costs).

Difficulties will arise in expressing functions that depend on value judgements; forexample, the social benefits and the social costs that arise from pollution.

The third step is to find the values of the variables that give the optimum value of theobjective function (maximum or minimum). The best techniques to be used for this stepwill depend on the complexity of the system and on the particular mathematical modelused to represent the system.

A mathematical model represents the design as a set of equations (relationships) and, aswas shown in Section 1.9.1, it will only be possible to optimise the design if the numberof variables exceeds the number of relationships; there is some degree of freedom in thesystem.

1.10.2. Simple models

If the objective function can be expressed as a function of one variable (single degree offreedom) the function can be differentiated, or plotted, to find the maximum or minimum.


This will be possible for only a few practical design problems. The technique is illus-trated in Example 1.1, and in the derivation of the formula for optimum pipe diameter inChapter 5. The determination of the economic reflux ratio for a distillation column, whichis discussed in Volume 2, Chapter 11, is an example of the use of a graphical procedureto find the optimum value.

Example 1.1

The optimum proportions for a cylindrical container. A classical example of the optimi-sation of a simple function.

The surface area, A, of a closed cylinder is:

A D D L C 24D2

where D D vessel diameterL D vessel length (or height)

This will be the objective function which is to be minimised; simplified:

fD L D D L C D2

2equation A

For a given volume, V, the diameter and length are related by:

V D 4D2 L

and

L D 4VD2

equation B

and the objective function becomes

fD D 4VD

C D2

2

Setting the differential of this function zero will give the optimum value for D

4VD2

C D D 0

D D 3

4V

From equation B, the corresponding length will be:

L D 3

4V

So for a cylindrical container the minimum surface area to enclose a given volume isobtained when the length is made equal to the diameter.

In practice, when cost is taken as the objective function, the optimum will be nearerL D 2D; the proportions of the ubiquitous tin can, and oil drum. This is because the cost


will include that of forming the vessel and making the joints, in addition to cost of thematerial (the surface area); see Wells (1973).

If the vessel is a pressure vessel the optimum length to diameter ratio will be evengreater, as the thickness of plate required is a direct function of the diameter; seeChapter 13. Urbaniec (1986) gives procedures for the optimisation of tanks and vessel,and other process equipment.

1.10.3. Multiple variable problems

The general optimisation problem can be represented mathematically as:

f D fv1, v2, v3, . . . , vn 1.2

where f is the objective function and v1, v2, v3, . . . , vn are the variables.In a design situation there will be constraints on the possible values of the objective

function, arising from constraints on the variables; such as, minimum flow-rates, maximumallowable concentrations, and preferred sizes and standards.

Some may be equality constraints, expressed by equations of the form:

m D mv1, v2, v3, . . . , vn D 0 1.3

Others as inequality constraints:

p D pv1, v2, v3, . . . , vn Pp 1.4

The problem is to find values for the variables v1 to vn that optimise the objective function:that give the maximum or minimum value, within the constraints.

Analytical methods

If the objective function can be expressed as a mathematical function the classical methodsof calculus can be used to find the maximum or minimum. Setting the partial derivativesto zero will produce a set of simultaneous equations that can be solved to find the optimumvalues. For the general, unconstrained, objective function, the derivatives will give thecritical points; which may be maximum or minimum, or ridges or valleys. As with singlevariable functions, the nature of the first derivative can be found by taking the secondderivative. For most practical design problems the range of values that the variablescan take will be subject to constraints (equations 1.3 and 1.4), and the optimum of theconstrained objective function will not necessarily occur where the partial derivativesof the objective function are zero. This situation is illustrated in Figure 1.16 for a two-dimensional problem. For this problem, the optimum will lie on the boundary defined bythe constraint y D a.

The method of Lagranges undetermined multipliers is a useful analytical technique fordealing with problems that have equality constraints (fixed design values). Examples ofthe use of this technique for simple design problems are given by Stoecker (1989), Petersand Timmerhaus (1991) and Boas (1963a).


Feasible region

Minimum offunction

y = a

f(v)

v

Figure 1.16. Effect of constraints on optimum of a function

Search methods

The nature of the relationships and constraints in most design problems is such thatthe use of analytical methods is not feasible. In these circumstances search methods,that require only that the objective function can be computed from arbitrary values ofthe independent variables, are used. For single variable problems, where the objectivefunction is unimodal, the simplest approach is to calculate the value of the objectivefunction at uniformly spaced values of the variable until a maximum (or minimum) valueis obtained. Though this method is not the most efficient, it will not require excessivecomputing time for simple problems. Several more efficient search techniques have beendeveloped, such as the method of the golden section; see Boas (1963b) and Edgar andHimmelblau (2001).

Efficient search methods will be needed for multi-dimensional problems, as the numberof calculations required and the computer time necessary will be greatly increased,compared with single variable problems; see Himmelblau (1963), Stoecker (1971),Beveridge and Schechter (1970), and Baasel (1974).

Two variable problems can be plotted as shown in Figure 1.17. The values of theobjective function are shown as contour lines, as on a map, which are slices through thethree-dimensional model of the function. Seeking the optimum of such a fu

coulson & richardson’s - dl4a.orgdl4a.org/uploads/pdf/coulson richardson's chemical engineering...

Documents