This item is held in Loughborough University’s Institutional Repository (https://dspace.lboro.ac.uk/) and was harvested from the British Library’s EThOS service (http://www.ethos.bl.uk/). It is made available under the following Creative Commons Licence conditions. For the full text of this licence, please go to: http://creativecommons.org/licenses/by-nc-nd/2.5/ ENERGY RECOVERY IN DRYING BY ADSORPTION HEAT PUNPING. by NCHEKV1UBE JOHN DANIEL ERINNE. Submitted in partial fulfilment of the requirements for the Degree of Doctor of Philosophy of Loughborough University of Technology. 1982. Supervisor and Director of Research: Professor D. C. Freshwater. CO by N. J. D. Erinne, June 1982. PAGE NUMBERS CUT OFF IN ORIGINAL DAMAGED TEXT N ORIGINAL i DEDICATION Dedicated to the memory of Dad, a man of great vision, whose life remains a burning inspiration and to Mum, to whom I owe so much. ii ACKNOWLEDGEh'TS I would like to express my very profound gratitude to Professor D. C. Freshwater for his kind advice and invaluable guidance during the course of this study and for making available the excellent facilities for the work. My thanks also go to Messrs. Tyler and Son of Loughborough for providing the facilities for part of the study and to Mr. R. Aird for his assistance and advice at various stages. May I also extend my thanks to Messrs. Hugh Peters and Andy Milne as well as the workshop staff of the Department of Chem. Engineering, L. U. T., for their kind co-operation and help during the construction of the experimental apparatus and during the experimental work. Gratitude also to my friends and colleagues - Dele Makanjuola, Mike Adigweme, Teju Bogunjoko, Mike Gyamerah, Emmad Baddar and James Ajayi - for their various useful suggestions. Very special thanks also to Granny, Mum, Paul and Sam and the other members of my family back home, whose fervent prayers and encouragement have been invaluable and to Miss Felicia 'Uju Anyaegbunam, whose emotional support has contributed very immensely to the success of the work. Thanks to the Anambra State Government and Federal Government of Nigeria for making the grant available for the study. Also to Mrs. Joan Bakewell for her patience in typing the scripts and to Mrs. Christine Sharpe for her contribution in the typing. Finally, my inestimable gratitude to God for blessing me with the health and strength that enabled me to pull through. 111 LIST OF TABLES No. Title Page 1- I World Primary Energy Consumption (1965-80). 5 1 -2 USA Primary Energy Consumption (1965-80) 5 1- 3 West European Primary Energy Consumption (1965-80) 6 1- 4. U. K. Primary Energy Consumption (1950-80). 7 1 -5 Forecast of U. K. Primary Energy Consumption 6 (1975-2000) 1 -6 Current and Real U. S. and World Oil Prices 12 (1880-1977). 1 -7 Sectoral Energy Consumption in U. K. (1980). 13 1 -8 Breakdown of U. K. Industrial Energy Consumption 13 (1980). 1 -9 U. K. Chemical Industry Energy Consumption (1979). 15 1 - 10A. Comparison of U. K. Chemical Industry, All Industry 16 and Total Energy Consumption (1979). (Electricity Generation Losses not included). 1 - 10B Comparison of U. K. Chemical Industry, All Industry 16 and Total Energy Consumption (1979). (Electricity Generation Losses Included). 1 - 11 End Use of U. K. Chemical Industry Non-Feedstock 17 Energy (1977). 1 - 12 Water Removed in Drying in some Major U. K. 18 Industries (1975). 2 -1 Latent Heats of 'later and Organic Solvents. 20 2 -2 Maximum Evaporation Rates from Materials. 24. 2 -3 Performance of Tumble Dryer. 1.8 2 - 4. Breakdown of Energy Consumption in Tumble Dryer. 1+9 3 - 1A Dryer Optimum Performance (T = 1000C) 68 3 - 1B Dryer Optimum Performance (f = 0.6) 68 3 -2 Typical Overall Heat Transfer Coefficients for 79 Various Process Systems. 5 -I Properties of Adsorbents. 105 7 -1 Experimental Results. 161 7 -2 Calculated Results. 163 7 -3 Comparison of Equilibrium and Breakpoint Capacities. 181 8 -I Air Conditions from Conventional Dryer. 189 8 -2 Prediction of Adsorption by Lee-Cummings Procedure. 200 8 -3 Effect of Initial Bed Temperature on adsorption 207 (Lee Cummings Procedure). iv List of Tables cont'd. No. Title Page 8- 4 Effects of Velocity on Adsorption 208 (Lee-Cummings Procedure). 8- 5 Prediction of Adsorption by Dehler Method. 210 8- 6 Prediction of Adsorption by Modified Dehler Method. 217 8- 7 Prediction of Adsorption by BRH Model. 221 8- 8 Effect of Initial Bed Temperature on Adsorption 230 (BRH Model). 8- 9 Prediction of Performance of TSC Dryer 232 (BRH Model). 8- 10 Replacement Efficiencies of TSC Dryer. 234 8- 11 Effect of Initial Bed Tem erature on 243 Performance of TSC Dryer (BRH Model). 8- 12 Effect of Regeneration Temperature on 2/4 Performance of TSC Dryer (BRH Model). 8- 13 Comparison of Performance of TSC and PSC 246 Dryers (BRH Model) 8- 14 Effects of Total Pressure and Condenser Temperature on Performance of PSC Dryer 246 (BRH Model). B1 Tumbler Air Inlet and Outlet Temperatures. 266 D1 Raw Data (Experiment 45). 2 74 V LIST OF FIGURES Figure No. Page No. 1- I Trends in U. K. Energy Consumption (1950-80) 8 and Projections (1975-2000). (Million tonnes oil equivalent). 1- 2 Trends in U. K. Energy Consumption (1950-80) 9 and Projections (1975-2000). (Percentage Proportions). 1- 3 U. K. Price Indices, Including Fuels 11 (Inflation Corrected) 1960-76. 2- I Drying Equilibrium Curves. 22 2- 2 Moisture Content vs Drying Time. 23 2- 3 Drying Rate vs Moisture Content. 23 2- 1. Drying Rate vs Time. 23 2- 5 Temperature Profile in a Brick During Drying. 30 2- 6 Flowchart for Dryer Design. 34. 2 -7 Ideal Continuous Counter Flow Drying. 37 2- 8 Temperature Profiles in Counter Flow Drying. 37 2 -9 Concentration Profiles in Counter Flow Drying. 37 2 - 10 Ideal Continuous Parallel Flow Drying. 39 2 - 11 Temperature Profiles in Parallel Flow Drying. 39 2 - 12 Concentration Profiles in Parallel Flow Drying. 39 2 - 13 Convective Dryer with Air Recirculation. 1+2 2 - 14. Sankey Representation of Energy Use in Tumble 47 Dryer. 3 -I Air Recycle Convective Dryer. 53 3 -2 Thermal Efficiency as a Function of Recycle 56 Ratio, Fractional Saturation and Temperature. 3 -3 Thermal Efficiency and Wet-Bulb Temperature as 57 Functions of Recycle Ratio and Frational Saturation. 3 - 4. Thermal Efficiency as a Function of Recycle 60 Ratio and Temperature (Wet-Bulb Temperature = 2+00 C). T w 3 -5 Thermal Efficiency as aFunction of Recycle Ratio 61 and temperature (Wet-Bulb Temperature T =60 C). w vi List of Figures cont'd. Fig ure No . Pare No . 3- 6 Thermal Efficiency and Fractional Saturation 62 as Functions of Recycle Ratio and7, et-Bulb Temperature. 3- 7 Optimization of Recycle Rati8 jr. terms of 66 Efficiency and Rate (T = 100 C) 3- 8 Optimization of Recycle Ratio in terms of 67 Efficiency and Rate (f = 0.6) 3 -9 Optimization of Recycle Ratio in terms of 69 = 1+0 C). Efficiency and Rate T W 3 - 10 Combined Dryer/Heat Exchanger. 71 3 - 11 Thermal Efficiency of Combined Dryer/Heat Exchanger as a Function of Temperature and 7 Fractional Saturation. 3 - 12 Thermal Efficiency of Combined Dryer/Heat 75 Exchanger as a Function of Recycle Ratio and Fractional Saturation. 3- 13 Energy Requirements and Heat Transfer Units 78 in Combined Dryer/Heat Exchanger. 1ý -1 Closed Cycle Mechanical Vapour Compression 81 Heat Pump. 4. -2 Ideal Carnot Heat Pump Cycle. 82 4. -3 Real Heat Pump Cycle. 82 4. - 1+ Mechanical Heat Pump Dryer. 86 4. -5 Performance of ECRC Heat Pump Dryer. 88 4. -6 Open-Cycle Air Recompression Dryer. 90 4 -7 (A& B) Alternative Air Recompression Cycles. 91 4. -8 Open-Cycle Steam Recompression Dryer. 93 4. -9 Lyle's Cycle. 95 4. - 10 Absorption Heat Pump. 97 4 - 11 ERG/ERSU Experimental Heat Pump. 98 4. - 12 Carnot Cycle Representation of Heat Driven AbHP. 100 5 -I Adsorption Isotherms. 1 C7 5 -2 Adsorption Isobars. 107 5 -3 Adsorption Isosteres. 107 5 - 4. Loading vs Relative Humidity. 107 vii List of Figures cont'd. Figure No. Pa =-e No. 5- 5 Freundlich's Isotherm 108 5- 6 Multimolecular Layer Isotherm 110 5- 7 Capillary Condensation Augmented Isotherm. 111 5- 8 Hysteresis Effect. 113 5- 9 The Adsorption Wave and Breakthrough. 115 5- 10 Breakthrough Temperature and Concentration Curves. 116 5- 11 Adsorption Bed Temperature Profiles. 116 5- 12 Depletion Temperature and Concentration Curves. 116 5- 13 Desorption Bed Temperature Profiles. 116 5- 14. Miller/Roberts Data. 129 5- 15 Worthington's Data. 129 6- I TSC Heat Pump Dryer. 133 6- 2 PSC Heat Pump Dryer. 134 6- 3 p-T Representation of TSC Heat Pump Dryer. 135 6 -4 p-T Representation of PSC Heat Pump Dryer. 136 7 -I Schematic flow Diagram of Experimental Rig. 151 7 -2 Adsorption Column. 152 7 -3 Wet/Dry Bulb Thermometer for Humidity Measurement. 155 7 - 4a. Adsorption Bed Temperature Profiles (Run 37). 158 7 - lb. Adsorption Air Temperature and Humidity Profiles 159 (Run 37). 7 -5 Effect of Flowrate on Breakpoint Capacity and Bed 165 Relative Humidity. 7 -6 Effect of Temperature on Breakpoint Capacity and 167 Bed Relative Humidity. 7 -7 Effect of Inlet Temperature on Temperature Change. 168 7 -8 Effect of Humidity on Breakpoint Capacity. 170 7 -9 Effect of Humidity on Bed Relative Humidity. 171 7 - 10 Temperature Change vs Humidity Change. 172 7 - 11 Effect of Bed Height on Breakpoint Capacity. 171 7 - 12 Correlation of Breakpoint Capacity and Bed 176 Relative Humidity. 7 - 13 Log-Log Correlation of Breakpoint Capacity and 177 Bed Relative Humidity. 7 - 14 Scheme for Evaluating Silica-Gel Equilibrium and 179 Breakpoint capacies. viii List of Figures cont'd. Figure No. Page No. 7- 15a. Desorption Bed Temperature Profiles (Run 28) 1 ý-4 7- 15b. Desorption Air Temperature and Humidity Profiles 1 b5 (Run 28). 7- 16a. Desorption Bed Temperature Profiles (Run 38). 186 7- 16b. Desorption Air Temperature and Humidity Profiles 187 (Run 38). 8- 1 Air Conditions from Convective Dryer. 190 8- 2 Thermal Efficiency of TSC Dryer (Ecuilibrium Model). 192 8- 3 Effects of Regeneration Temperature and Exhaust 193 Condition on Thermal Efficiency of TSC Dryer (Equilibrium Model). 8 - 1+ Thermal Efficiency of PSC Dryer (Equilibrium 196 Model). 8 -5 Effect of Pressure on Efficiency of PSC Dryer 196 (Equilibrium Model). 8 -6 Scheme for Evaluation of Adsorption Models. 199 8 -7 Temperature Lift vs Fractional Saturation 202 (Lee-Cummings Procedure). 8 -8 Breakpoint Capacity vs Fractional Saturation 203 (Lee Cummings Procedure). 8 -9 Breakpoint Capacity vs Bed Relative Humidity 20J+ (Lee Cummings Procedure). 8 - 10 Temperature Lift vs Humidity Change 205 (Lee-Cummings Procedure) 8 - 11 Temperature Lift vs Fractional Saturation 211 (Dehler Method). 8 - 12 Breakpoint Capacity vs Fractional Saturation 213 (Dehler Method). 8 - 13 Breakpoint Capacity vs Bed Relative Humidity 211} (Dehler Method) 8 - 14 Temperature Lift vs Humidity Change 215 (Dehler Method) 8 - 15 Temperature Lift vs Fractional Saturation 218 (Modified Dehler Method) 8 - 16 Breakpoint Capacity vs Fractional Saturation 219 (Modified Dehler Method). 8 - 17 Breakpoint Capacity vs Bed Relative Humidity 220 (Modified De-, ' ler Method). 1X List of Figures contd. Figure No. Pare No, 8- 18 Temperature Lift vs Humidity Change 221 (Modified Dehler Method). 8- 19 Temperature Lift vs Fractional Saturation X25 (BRH Model). 8- 20 Breakpoint Capacity vs Fractional Saturation 226 (BRH Model) 8- 21 Breakpoint Capacity vs Bed Relative Humidity 227 (BRH Model). 8- 22 Temperature Lift vs Humidity Change 228 (BRH Model) . 8- 23 Thermal Efficiency of TSC Dryer (BRH Model. 233 8- 24 Loci of Replacement Efficiency and Fractional 235 Saturation of TSC Dryer. 8- 25 Theoretical COP of TSC Heat Pump. 238 8- 26 Actual COP of TSC Heat Pump. 239 8- 27 Thermodynamic Efficiency of TSC Heat Pump. 240 8- 28 Adsorbent Bed size for TSC Dryer. 212 B1 Tumbler Inlet and Outlet Temperatures. 268 B2 Scheme for Flash Calculation. 267 C1 Blower Performance Characteristics. 272 C2 Flow Calibration Chart. 273 D1 Bed Temperature Profiles (Run 4.5) . 275 D2 Air Temperature and Humidity Profiles (Run 4-5) . 276 " PLATE A Experimental Rig. x NOTATION A Cross sectional area, drying area a Specific surface area Ch Humid heat capacity of air Amax Maximum humid heat capacity of air Cmin Minimum humid heat capacity of air Cpg Specific heat of air C Pv Specific heat of water vapour COP Coefficient of performance COP max Ideal coefficient of performance DAB Gas phase diffusivity DL Liquid phase diffusivity DPF Dryer performace factor E comp Compression energy EP Primary energy E rec Energy recovery in condenser EvaP Energy for vaporisation of moisture E Exergy x f Fractional saturation of air in dryer fd Relative drying rate, N/N f Fraction of MTZ unused. Z G Air flowrate G Air mass flow velocity v (m2) (m2/m3 ) ý k g KJ kg_OC KJ ' kg_°C KJ (kg-0 ) KJ ýkg aC (m2/hr ) (m2/nr) kg 2N KJ_hr (KJ) (KJ) (am) (KJ) (KJ) (kg ) ( kg ) m2_ß, ) xi HTU Heat transfer units. 1000 Air humidity h Convective heat transfer coefficient c h Radiative heat transfer coefficient r ht Total heat transfer coefficient jD Colburn factor. k Thermal conductivity k Mass transfer coefficient g K Breakpoint capacity correlation constant KT Trouton's ratio Kt Moisture distribution ratio. L Length of continuous dryer LT Length of a transfer unit LUB Length of unused bed 1 Slab thickness M Mass of solid adsorbent s MZ Adsorption bed height modulus MER Specific moisture extraction rate MTU Mass transfer units. N Drying rate N Constant drying rate c NReb Reynold's number Nt Number of transfer units n Breakpoint capacity correlation constant. (lb s 1000f't3 KJ m2-hr-ýC KJ m2-hr-0 C KJ m2-hrýC KJ km-hr-°C h KJ ýkg C ( ) (Z) (m) (m) (kg) (kg ) (kg m -hr (kg m -hr xii p Partial pressure (mm Hg) p Vapour pressure (mm Hg) o P ob Saturation vapour pressure at bed ( Hg) temperature p Surface vapour pressure ( mm Hg) s Psat Saturation partial pressure (mm Hg) P Wet-bulb partial pressure (mm Hg) w Q Heat input (KJ) Qabs Absorber heat (KJ) Qcond Condenser heat (KIT) Qe rg Regeneration heat (KJ) QT Total heat input (KJ) R Universal gas constant 1 kg mole- K r Recycle ratio to dryer rA Rate of adsorption (-ý m -hr rD Dryer recycle ratio in AHP S Entropy (0 0 C S Approach to saturation (%) SER Specific energy requirement (g T Temperature ( °C) Tb Boiling temperature ( °C) T Critical temperature ( °C) c Tf Feed temperature (0C Tr Reduced temperature Ts Drying surface temperature (°c) Tsb Temperature of adsorption bed at breakpoint (°C) TSD Temperature of desorption bed (°C) xiii T si Initial temperature of adsorption bed Tsat Saturation temperature T Wet-bulb temperature w U Overall heat transfer coefficient V Gas volume ß VL Liquid volume y Gas flow velocity W Adsorbate loading on adsorbent wb Adsorbent breakpoint capacity wba Adiabatic breakpoint capacity Wbi Isothermal breakpoint capacity weD Initial capacity of desorption bed Initial capacity of adsorption bed X Moisture content of drying solids. X Relative humidity of air (°c) C°c) (°c) ( KJ o) m -hr- C (m3) (m3) ' hr ) () (&) kg () ( g) C: L) (h) ( g) Xb Adsorption bed relative humidity at breakpoint Xc Critical moisture content (kZ) x Equilibrium moisture content (hg) e kg x Equilibrium bed relative humidity e XeD Desorption bed equilibrium relative humidity Xf Modified relative humidity, X exp (1/500) x Distance inside a slab (m) xiv y Air humidity ( g) sat Saturation air humidity (. ýL) kg yW Wet-bulb air humidity (kg) kg Z Height of adsorbent bed (m) Zs Thickness of slab (m) General Subscripts Conventional Dryer: 0 Ambient conditions I Dryer inlet 2 Dryer outlet Experimental Adsorption Column: i inlet conditions e exit conditions Adsorption Heat Pump : 0 Ambient conditions Di Dryer inlet De Dryer exit Ai Adsorber inlet Ae Adsorber exit Ri Desorber inlet Re Desorber exit. xv Greek Letters a Dimensionless time constant Humid Heat ratio, C ho /C h2 AT Temperature lift Ay Humidity change 6 Polar parameter E Bed void fraction Er Thermal effectiveness A Latent heat aA Heat of adsorption of Latent heat at feed temperatures A Molar latent heat m A Latent heat at wet-bulb w n Dryer thermal efficiency (°c) Cg ( g) ( g) ( g) xJ kg mole () kg C%) T l. __,, Thermal efficiency of conventional dryer (C /-Q) %: VllV '1EM Efficiency of AHP predicted from Equilibrium Model ideal Ideal instantaneous efficiency of dryer (ö) T1 R Efficiency of combined dryer/heat exchanger (%) T1 RH Efficiency of AHP predicted from BRH Model T Thermodynamic efficiency (; ý) Characteristic moisture content Drying time (hrs) Ab Breakpoint time (hrs) 9 ba Adiabatic breakpoint time (hrs) 40,9 bi Isothermal breakpoint time (hrs) xvi A Constant drying time (hrs C Ad Dead time (hrs) A Equilibrium time (hrs) e 9f Falling rate drying time (hrs AT Total residence time (hrs) g Air density () m Ps Solids density (kg) 3 M u Fluid viscosity (g m-hr TI Total pressure (atmosphere) IIH Dimensionless heat transfer potential it M Dimensionless mass transfer potential SZ Collision integral v a Surface tension (N/m a Molecular diameter (A° Particle shape factor xvii CONTENTS Title Page DEDICATION. i ACKNOWLEDGP. i, I', TS. ii LIST OF TABLES. --ii LIST OF FIGURES. v NO'-'I TIONS. X CONTENTS. xvii ABSTRACT 1 1. INTRODUCTION AND GENERAL BACKGROUND. 3 1 .1. Background. 3 1.1.1. The Energy Problem. 3 1.1.2. Input Forms. 3 1 .1 . 3. Trends. 4 1.1.4.. Costs. g 1.2. Energy Use Pattern. 13 1.2.1. Overall Consumption. 1 1.2.2. Energy Use in Chemical Industry. 14 1.2.3. Energy Use in Drying. 17 1.3. Project Objective. 19 2. DRYING PROCESSES. 20 2.1. Scope of Application. 20 2.2. General Drying Behaviour. 21 2.2.1. Equilibrium Relations. 21 2.2.2. Drying Periods. 22 2.3. Theories of Drying. 27 2.3.1. Liquid Diffusion Mechanism. 27 2.3.2. Capillary Theory. 28 2.3.3. Receding Plane Model. 29 2.3-4. Vetted Surface Model. 31 2.3.5. Other Models. 32 2.4. Performance of Dryers. 32 2.4.1. Drying Time 32 2.1.2. Design. 33 xviii Contents cont. -I'd. Title Page 2.5. Energy Requirements in Dr-. ying. 4-0 2.5.1. Ideal Instantaneous Energy efficiency. 4-C 2.5.2. Ideal Overall Efficiency. 2.5.3. Actual Efficiency. 1+6 2.6. Measurement of Thermal Efficiency of Tumble Dryer. 46 2.7. Energy Conservation. 50 2.7.1. Housekeeping. 50 2.7.2. Practical Constraints. 51 2.7.3. Heat Recovery. 51 3. CONVENTIONAL HEAT RECOVERY METHODS. 52 3.1. Exhaust Air Recirculation. 52 3.1.1. Constant Fractional Saturation. 52 3.1.2. Constant Wet-Bulb. 58 3.1 . 3. Effects on Drying Rate. 63 3.2. Heat Exchange Recuperation. 70 3.2.1. Efficiency of Combined System. 70 3.2.2. Heat Transfer Units. 76 ! ý. HEAT PUMP DRYING. 80 4.1. Closed-Cycle Mechanical Vapour Compression 80 Heat Pump. 4.1-1. Coefficient of Performance (COP) 80 4.1.2. Thermodynamic Efficiency. 83 4.2. The ECRC Heat Pump Dryer. 85 4.3. Open-Cycle Vapour Recompression. 87 1.3.1. Air Recompression. 89 4.3.2. Steam Recompression. 92 4.4. Absorption Heat Pumps (AbHP) 94 4.4-1. Honigman's Locomotive. 94 4.. 4.2. Absorption Refrigeration Cycle. 96 1.4.3. The ERG/ERSU Heat Pump 96 4 4.4. Coefficient of Performance (COP) 99 4.4.5. Chemical Heat Pump (CHP) 102 x ix Contents cont'd. "'itle Page 4.5. Concept of Adsorption Heat Pump. 102 5. ADSORPTION 1 01+ 5.1. Introduction. 104. 5.1-1. Nature of Adsorption. 04. 5.1.2. Types of Adsorption. 10+ 5.1.3. Properties of Adsorbents. 105 5.2. Adsorption Equilibrium Relations. 106 5.2.1. Presentation of Equilibrium Relations. 1 06 5.2.2. Monolayer Theory 108 5.2.3. Multilayer. 109 5.2.1.. Capillary Condensation. 110 5.2.5" Water Adsorption Equilibrium 112 5.2.6. Hysteresis. 113 5.3. Breakthrough Phenomenom. 111+ 5.4. Kinetics of Adsorption. 117 5.4.1. External Diffusion Control. 119 5.4.2. Internal Pore Diffusion Control. 120 5.4.3. Physical Adsorption Control 121 5 4.4. Combined Resistances. 5.5. Adiabatic Adsorption. 121 5.6. Desorption. 123 5.7. KITZ Approach 121 5.8. Other Semi-Steady State Methods for 127 Non-Isothermal Adsorption. 5.8.1. Breakpoint Capacity. 127 5.8.2. Effects of Operating Variables on alb 128 5.8.3. Temperature Change. 130 5.8.4. Regeneration. 131 6. THE ADSORPTION HEAT PUMA' (AHP) 132 6.1. System Description. 132 6.1.1. Temperature Swing Cycle (TSC) 132 6.1.2. Pressure Swine Cycle (PSC) 132 xx Contents cont' d. Title 6.2. Performance Evaluation. 6.2.1. COP. 6.2.2. Thermal Efficiency. 6.3. Equilibrium Model. 6.3.1. Dryer. 6.3.2. Adsorption. 6.3.3. Regeneration 6.3-4. Cycle Performance . 6.3.5. PSC Dryer. 6.4. Lee-Cummings Procedure. 6.1f. 1. Adsorption. 6.1.2. Regeneration. 6.5. Dehler Method. 7. EXPERIMENTAL. 7.1. Objectives. 7.2. Design of Apparatus. 7.2.1. Air Blower 7.2.2. Air Heater 7 2.3. Adsorption Column. 7.3. Instrumentation and Procedure. 7 3.1. Flow Measurement. 7.3.2. Temperature Measurement. 7.3.3. Humidity Measurement. 7.3.4.. Procedure. 7 3.5. Materials. 7 4. Results. 7.5. Calculation Procedure. 7.6. Effects of Variables. 7.6.1. Flowrate. 7.6.2. Temperature. 7.6.3. Humidity 7.6.4. Bed Height. Pace 157 137 138 138 '139 140 141 141 11+2 143 143 146 147 149 149 149 150 150 150 153 153 153 154 156 157 157 162 164. 164 166 169 173 xxi Contents contd. "'itle Pare 7.7. Correlation of Data. 175 7.8. Comparison of Breakpoint and Equilibrium Capacities. 178 7.9. Regeneration. X63 8. PREDICTION OF PERFORMANCE. 186 8.1. Conditions of Air from Conventional Dryer. 168 8.2. Equilibrium Model. 191 8.2.1. Drying Conditions. 191 8.2.2. Regeneration Temperature. 191 8.2.3. Exhaust Conditions. 194 8.2.1}.. Comparison with Conventional Dryer. 194 8.2.5. Pressure Swing Cycle (PSC) 195 8.2.6. Comparison of TSC and PSC. 195 8 3. Evaluation of Adsorption Models. 198 8.3.1. Lee-Cummings Procedure. 19E 8.3.2. Dehler Method. 209 8.3.3. Modified Dehler Method. 216 8.4. Bed Relative Humidity (BRH) Model. 222 8.4-1. Outline of Model 223 8.4.2. Evaluation of Model. 223 8.5. Overall Cycle Prediction by BRH Model. 231 8.5.1. Thermal Efficiency. 231 8.5.2. Comparison with Mechanical Heat Pump 236 Drying. 8.5.3. Coefficient of Performance. 236 8.5.1. Adsorbent Bed Size. 241 8.5.5. Initial Bed Temperature. 241 8.5.6. Regeneration Temperature. 244 8.5.7. PSC Dryer. 245 9. CONCLUSIONS AND SUGGESTIONS FOR FURTHER ''i O RK 24 9.1. Conclusions 9.2. Suggestions for Further Tork 2ýG xxii Contents cont'd. Title Pace REFERENCES 251 APPENDICES 257 A. Thermodynamic and Transport Properties. 57 2-5 Al. Vapour Pressure 257 A2. Heat of Vaporisation. 259 A3. Heat of Adsorption. 260 A4. Viscosity. 261 A5. Diffusivity. 264 266 B. Sample Calculation of Tumbler Efficiency. C. Blower Characteristics and Flow Calibration. 272 C1 . Blower Operating Characteristics. 272 C2. Flow Calibration Chart. 273 D. Sample Calculation of Experimental Results. 274. E. Prediction Programme - BRH Model. 279 I ABSTRACT Drying is one of the most energy intensive operations of the chemical industry and accounts for about 6% of the total energy used by U. K. industry and about 2.5 of the overall energy demand of . he U. K. Measurements taken on a typical industrial tumble dryer confirmed that generally convective dryers operate at thermal efficiencies less than 4 and that over 5Y of the energy input is lost as sensible heat content of the moist exhaust air. Any significant improvement in the thermal performance of dryers would therefore require a means of gainfully recovering the heat lost in the exhaust air. Some conventional heat recovery methods have been considered. Recirculation of exhaust air was shown to improve thermal efficiency but at the cost of reduced drying rates. Heat recovery by heat exchange was found unattractive because very large heat exchange surface areas would be required. A new type of adsorption heat pump (AHP) which may be operated either as a temperature swing cycle (TSC) or a pressure swing cycle (PSC), has therefore been proposed here for heat recovery from dryers. An experimental rig was built and used to investigate the adsorption of moisture on silica-gel in a 1.5 m. high, 0.25m. diameter column under conditions that simulate an industrial dryer, including high temperatures and humidities. Correlation of the experimental data led to the derivation of a polynomial function, similar to the system equilibrium equation, which relates the breakpoint capacity of the adsorbent to the breakpoint bed relative humidity. This function was used to develop a new theoretical model for predicting the performance of the proposed heat pump dryer. The predictions of this model enabled similar predictions obtained from two other models synthesized from various proposals put forward by other investigators previously to be tested against experimental results. This new model was found to be the most appropriate for the conditions encountered and was therefore considered to be the most suitable for predicting the performance of the adsorption heat pump drying system. 2 Theoretical predictions based on this model indicate that the heat pump drL-er may be attractive for low temperature ( <1000C) drying. Under high temperature drying conditions low thermal efficiencies and unfavourably large adsorption bed size re _uirements make the heat pump unattractive. CHAPTER ONE INTRODUCTION AND GENERAL BACKGROUND 3 1. INTRODUCTION r. ý'D GENERAL B. =. C : G. RC UD. I . 1. Background 1 .1 .1. The Energy Problem The last decade witnessed two successive major global energy crises: (i) 1973/74 crisis, following the Middle-East war; (2) 1979 crisis, arising from the political upheaval and the subsequent revolution in Iran. These two successive crises have served to show how dependent the fortunes of the industrial' society are on the future of world energy supplies. It is hard to overstate the importance of energy to the sustenance of society. Most facets of human activity depend directly or indirectly on energy and this dependence becomes greater as machines are used more and more to-replace animal power and human power. The energy crises can be traced to a number of factors which include: (i) Excessive reliance on fossil fuels, especially oil and gas, as major sources of primary energy. (ii) The sudden escalation of fuel prices in the aftermath of the 1973 Middle-East war. (iii) Gross mismatch in terms of geographical location, of primary energy availability and its demand. (iv) The finite availability of fossil fuels, especially oil. 1 .1 . 2. Inpu t Forms Primary Energy: The final form in which energy is required for use is often not the form in which it is available naturally. Therefore, it often has to be converted from one form to another in order to be used. For instance, to obtain energy for househeating, oil or coal may be fired in a boiler. Electricity for lighting may be generated from a nuclear reactor. Each of these clearly involves transformation of a primary energy input, be it coal or nuclear fuel, into a convenient form in which it is required. The major primary energy 4. inputs are essentially oil, natural gas, coal, nuclear fuel and hydro- power. Fossil Fuels : Fossil fuel refers to primary energy sources which are formed by geo-degradation of biological material in the earth's crust over the years. Conventionally, this includes oil, natural gas and coal. 1 .1 . 3. Trends Tables 1-1 to 1-3 show the total and primary energy inputs for the whole world, U. S. A. and Western Europe respectively over the period 1965-1980. Table 1-4 shows the same data for U. K. between 1950 and 1980. The statistics show that there was a consistent upward trend in total energy consumption which subsequently levelled-off during the last decade as a result of the energy crisis. Evidently oil is the most important primary energy input. Its proportion in the energy mix has increased considerably while that of coal has shown a steady decline. The proportion of natural gas has also increased consistently. Overall it is a reflection of the dominance of oil and gas in the energy mix that together they account for over 6c of world energy consumption. The trends are more clearly illustrated in figs. 1-1 and 1-2. Obviously, between 1950 and 1970 there was a clear switch from coal to oil in the U. K; as the proportion of oil continued to rise at the expense of coal. Following the 1973/74 crisis, the proportion of oil started to fall and the slack was taken up more by natural gas than coal which more or less remained steady. Several reasons have been adduced to explain this trend, including cheapness, less environmental problems associated with extraction and use, ease of handling and transportation (especially oil) and ease of conversion to heat and electricity. Whatever the reason the significance is that any adverse development in the supply or price of oil and gas is immediately transmitted, with very considerable impact, to the overall energy situation, as was the case during most of the last decade. 5 TABLE 1-1 : WORLD PRIMARY ENERGY COIN3UMP ION (1965-1980)' 1965 1970 1975 1980 PRIMARY TONNES TONNES TONNES TONNES ENERGY OIL EQ. % OIL EQ. % OIL EQ. OIL EQ. x109 x109 x109 109 x OIL 1.53 39.0 2.28 144.0 2.71 4.6.0 3.00 43.6 GAS 0.65 16.5 0.96 ' 18.5 1.11 19.0 1.28 18.6 COAL 1.50 38.0 1.62 131.0 1.65 27.5 2.02 29.3 NUCLEAR ( 0.25 6.5 0.02 0.5 0.09 1.5 0.17 2.5 HYDRO 0.31 6.0 r 0.36 6.0 0.11 6.0 TOTAL 3.93 100 5.19 i loo 5.92 100 6.88 M 100 TABLE 1-2 : U. S A. PRIMARY ENERGY CONSUMPTION (1965-1980) 1 1965 1970 1975 1980 PRIMARY TONNES TONNES TONNE TONNES ENERGY OIL EQ. % OIL EQ % OIL E .% OIL EQ. % X109 x10 9 x10 9 x109 OIL 0.55 41.0 0.69 4.1.6 0.77 44.8 . 79 4-2.9 GAS 0.4-3 32.1 0.56 33.7 0.51 29.7 . 4.9 26.6 COAL 0.31 23.1 0.33 19.9 0.32 18.6 . 4.1 22.3 NUCLEAR 0.01 0.6 0.04 2.3 . 07 3.8 HYDRO 0.05 3.8 0.07 4.2 0.08 4.6 . 08 .4 TOTAL 1.31 100 1.66 100 1.72 100 1.84. 100 6 TABLE 1-3 : WEST EUROPEAN PRIMARY ENERGY CO:, -SJMP-7ION (1965-1980)1 1965 1970 1975 1980 PRIMARY TONNES TONNES TONNES TONNES ENERGY OIL EQ. % OIL EQ. ýo OIL EQ. % OIL EQ. x 109 x 109 ix 109 109 x OIL 0.39 x+5.9 0.63 58.9 0.66 - 58.4 0.68 53.1 GAS 0.02 2.4 0.07 6.5 0.15 13.3 0.18 14.1 COAL 0.36 42.3 0.27 25.2 0.19 16.8 0.27 '21 .1 NUCLEAR ( 0.01 1.0 0.03 2.7 0.05 3.9 HYDRO ý 0.08 9" 0.09 8.4 0.10 8.8 0.10 17.8 TOTAL 0.85 1100 1.07 100 1.13 100 1 1.24. 100 TABLE 1-5 : FORECAST OF U. K. PRSMhRY ENERGY CONSUMPTION (1975-2000) 1975 1990 2000 PRIMARY TONNES TONNES TONNES ENERGY OIL EQ. % OIL EQ. % OIL EQ. % x 10 x 10 6 x 10 OIL 92.0 45.0 81.0 34.5 75.0 29.1 GAS 32.9 16.1 15.0 19.3 30.0 11.6 COAL 71.9 35.2 84.0 35.9 96.0 37.2 NUCLEAR 6.3 3.1 21.0 10.0 4.8.0 18.6 ALTERNATIVE 1.2 0.6 3.0 1.3 9.0 3.5 ENERGY TOTAL 204.3 100 234.0 100 258.0 1 100 7 ö ö I- 0 t-, OD Co -ä. ON x0F. -3 AD LIZ al% 8 9 N O O " Q OD 0 -, J - py N -. b, O %Z -' O -` HO _ Ö " lJi -l N Z rN j IN ýU C/) " 1 J e O g O W " r v ?' CC) ö9 O N ' -. J ' lr+ '. O !Xp " - - N N H . . . ý N 11--m 1%D 11.0 0 Cl\ 2 e0 cri \O vi `_. O O W " V Vi OAI N ýl -l Co ix O O N - O -' H N) Co 0 t-' 1J Lw O, \ O\ CX) W N Lw O O ºr n O D ý; " oý ao -1 Lw vý _a _ H ýý C x `Cf s O 1 0 lm v ý]N NK n C z C O F. O cfl N 8ö Co %o 0000 CD ö 0 U 0 K CD 0 ýnp ýi' cf- 0 G 00 OD ýl IPxj F'ý a FL 0 -J. N- CD p n c+ H. . 0 7C to t ý. D CD J 11 N 0 Z fA tý 9 W N ý N" N" O 0 J 10 CO ). j" N cD lJ'i 0 U, U, --a N1 . s+ 9 1 F-'" N J D C n J CD N 1 CO H \I 1 \ "11 /1 1 /1 / I Cl% c7 O H ---j cc -ýýýý_ý 10 Projections of U. K. energy demand up to the end of this century. adapted from Energy Paper No. 11 published by U. K. Department of E-nergy 12, are presented in Table 1-5 and indicated in figs. 1-1 and 1-2 by dashed strokes. The main assumptions of the forecasts are: (1) A low economic growth rate of 1- 2/-c in U. K. up to the e nd of the century. (2) Energy prices would rise more slowly during the rest of the century than in the last decade (see section 1.1-4). (3) Increased implementation of energy saving me asures would yield about 151/"L reduction in energy consumption below what it would have otherwise reached. Clearly, the highlight of these forecasts iz a continued rise in the total energy consumption, though slower than was experienced in the 1950's and 1960's. There would be a switch to coal and nuclear power at the expense of oil and gas. 1.1 . 1. Costs For a long time energy costs remained very low. The relevant U. K. energy price indices up to 1977, corrected for inflation, originally presented by Kantyka3 and extended by Grant 4, are reproduced in fig. 1-3. The graph clearly shows the sharp rise in fuel prices between 1973 and 1974 relative to other prices. Similar data for U. S. A. and the world, indicating both actual transaction and inflation corrected real prices, are also presented in Table 1-65. While energy remained cheap the evolving industrial society fashioned itself accordingly, taking advantage of the situation, with more energy driven machinery and consumer goods being introduced. It is no surprise therefore that the sudden escalation of fuel prices came as a severe shock. More than anything else it revealed the inherent dangers associated with the excessive reliance on petroleum resources for energy and the need to develop means of reducing energy consumption without possibly reducing industrial production. 11 350 300 750 Index. 200 15a 5< Fig. 1-3k. U. K. Price Indices, Including Fuels (Inflation Corrected) 1960-76. O crude od 0 heavy fuel od x elecnnuty i plant e wages and satarin natural gas 66 67 -68 69 7.71 1? 73 71 75 76 77 78 79 8 X60 61 62 63 64 65, 100 ' Year. 12 TABLE 1-65 : CURRENT AND REAL U. S. AND WORLD CRUDE OIL PRICES (1880-1977) YEAR CURRENT U. S. PRICE Dollars/bbl) Average price) U. S. CONSUMER PRICE INDEX (1957-1959= 100) REAL U. S. price in 1957-59 (Dollars/ bbl). WORLD PRICE estimated actual trans- actions price. REAL '1VO LD price 1957-59 (Dollars/ bbl) 1880 . 94. 34. 2.76 1890 . 77 32 2.41 1900 1.19 39 4.10 1905 . 62 31 2.00 1910 . 61 33 1.85 1915 . 64- 35 1.83 1920 3.07 70 4.40 1925 1.68 61 2.75 1930 1.19 58 2.01} 1933 . 67 4.5 1.4.9 1910 1.02 49 2.09 19+5 1.22 63 1 . 95 1950 2.51 84. 3.01 1.71 2.04 1955 2.77 93 2.97 1.63 1.75 1960 2.88 107 2.70 1.53 1.43 1965 2.86 110 2.50 1.33 1.21 1970 3.18 13tß 2.37 1.26 . 9tß 1971 3.39 145 2.34 1.66 1.14 1972 3.39 150 2.26 1.84 1.23 1973 3.89 15. 2.53 2.91 1.89 1974 6.74. 175 3.85 10.77 6.15 1975 7.67 187 4.10 10.72 5.73 1976 8.11 202 4.01 11.51 5.70 1977 8.22 211 3.90 13.12 6.22 13 1.2 Energy Use Pattern 1 2.1. Overall Consumption A breakdown of the energy consumption in 1980 by all U. K. end users, in terms of sectors, is presented in Table 1-7 below. TABLE 1-7 : SECTORAL ENERGY CONSUMPTION IN U. K. (1980) 2 Energy Consumption SECTOR Therms x10 KJx1012 Industry 19,173 2,023 33.9 Transport 14,103 1,488 25.0 Domestic 15,784 1,665 27.9 Public Administration 3,51+5 374 6.3 Agriculture 591 63 1.0 Others 3,342 353 5.9 TOTAL 56,51+7 5,966 100.0 TABLE 1-8 : BREkKDOy'y N OF U. K. INDUSTRIAL ENERGY CONSUMPTION (1980) 2 GROUP ö i t Iron and Steel 2,867 302.5 15.0 Engineering, Metals 3,738 39+. 4 19.5 Food, Drinks, Tobacco 1,875 197.8 9.8 Chemicals and Allied 4,140 436.8 21.6 Textiles, Leather. 940 99.2 4.9 Clothing. Paper, Printing, Statione 1,102 116.3 5.7 Bricks, Tiles, Fireclay 454 47.9 2.4 other Building Materials. Ceramics, Glass 631 66.9 3.3 Cement 787 83.0 4.1 Other 2,636 278.1 13.7 Energy Consumption 1 Therms x10 KJx1012 i TOTAL 19,173 2,022.9 100.0 14 Clearly, industry, which accounts for about 34 of total demand, is the single most important energy user. further break- down of energy usage within industry, (Tablel-8), shows that among the different industrial groups the Chemical and Allied Products Industry is the most prominent user, accounting for 21.6ö of industrial consumption. 1.2.2. Energy Use in the Chemical Industry The chemical industry energy consumption consists essentially of two components : (i) fuel, power and others; (ii) feedstocks. The fuel demand comprises the primary energy inputs - coal, oil and gas - used essentially to provide process heating, including generation of steam, while the power demand is mainly in form of purchased electricity, which is used to provide motive power for driving mechanical equipment such as compressors, pumps, grinders/crushers, conveyors and agitators/ mixers. There is also an extra demand for energy, mainly in form of electricity, for other purposes such as lighting and electrolysis which are strictly neither process heating nor motive power. In addition to purchasedelectricity the chemical industry generates significant quantities of electricity and power internally. The primary energy for this comes mainly from heat generated in the combustion of primary fuels. The feedstock component of the energy requirement consists of naphtha, LPG and natural gas which serve as basic raw materials for various petrochemical processes. The U. K. Department of Energy 2 statistics on overall and industry sector energy consumption, reproduced in Tables 1-7 and 1-8, do not account for the primary energy used to generate purchased electricity. Besides, they include only a fraction of the chemical industry feedstock energy input. Alternative data from Conseil European des Federations de l'Industrie Chimique (CEFIC) sources give the following breakdown of U. K. chemical industry energy consumption after the feedstock input and the primary fuel equivalent of purchased electricity have been put into consideration. 15 TABLE 1 -9 : U. R. CHEMICAL II DUJTRY E, RCY COIýSUb: PTIOI; (1979) 7 Energy Consumption therms x10 KJx1012 of Fuel C5) of Feedstock TOTAL Fuel : coal and other soli d 158.3 16.7 3.7 fuels oil & other liquid fuels 1348.7 142.3 31.8 gaseons fuels 1166.7 123.1 27.6 purchased electricity 1561.9 164.8 36.9 Sub-total : Fuel 4234.9 446.9 100.0 50.3 Feedstock: Liquid fuels 3200.6 337.7 76.5 gaseons fuels 985.7 101+. 0 23.5 Sub-total : Feedstocks 4186.3 441.7 100.0 49.7 TOTAL 8421.2 888.6 100.0 100.0 Apparently, the fuel and feedstock energy demands of the chemical industry are about 5o each. However, it will be inappropriate to make a direct comparison between the data of Table 1-9 and those of Tables 1-7 and 1-8 owing to the different underlying assumptions. The comparison can be facilitated by adding the fuel energy demand of the chemical industry to the industry and overall consumption figures as well as discounting the primary fuel equivalent of purchased electricity in the chemical industry fuel energy input. Using a fuel to electricity conversion factor of 0.3, the data of Table 1-10A are obtained. On this basis therefore, the chemical industry would account for about 331% of U. K. industry sector energy consumption and 1 of overall consumption by all end users. 16 TABLE 1-10A : COMPARISON OF U. R. CHEMICAL INDUSTRY nLL ITO Sý'RY AND TOTAL. ENERGY CONSUMPTION 1979 E1ectricity Generation Losses not included) Energy Consumption therms x 10 KJ x 1012 Chem. Ind : Fuel 3,131.9 331.5 Feedstock 4,186.3 441.7 (a) Total Chem. Ind. 7,328.2 773.2 (b) All Industry 22,360.9 2.359.3 (c) Overall U. K. 59,732.7 6,302.4 (a) / (b) 32.8jo 32. (a) / (c) 12. 12. V/o Alternatively, the data may be compared after electricity generation losses have been included. The U. K. Department of Energy gives the electricity demands by U. K. industry and by all consumers in 1979 as 2,533 and 8028 x 106 therms respectively2. Based on this then, the chemical industry demand would come down to about 29/1-c of all industry ener&y consumption and 10. of total U. K. consumption (Table 1-10B) . TABLE 1-1 OB : COMPARISON OF U. K. CHEMICAL INDUSTRY, ALL INDUSTRY AND TOTAL ENERGY CONSUMPTION 1979 2'7 Electricity Generation Losses Included). Energy Consumption therms x 10 KJ x 1012 Chem. Ind : Fuel )+, 234.9 446.9 Feedstock 4,186.3 441.7 (a) Total Chem. Ind. 8,421.2 888.6 (b) All Industry 29,365.0 3,098.3 (c) All U. K. 79,558.3 8,394.2 (a)/(b) 28.28. (a)/(c) 1 0.6%0 1o. 6 17 Grant4 presented approximate estimates of the broad end-use pattern of U. K. chemical industry non-feedstock (fuel) energy, which is reproduced in Table 1-11. This shows that nearly 75`- of the fuel energy input goes into process heating, which includes a wide range of heat-using operations, such as direct heating of process streams, distillation, evaporation, drying etc. It is noted however that this data includes purchased electricity used in electrolysis and for electrical drives for which the generation losses are not included. TABLE 1-11 : END USE OF U. K. CHEMICAL INDUSTRY NON-FEEDSTOCK ENERGY( Therms x 10 Heat Electrolysis 2469 221+ Mechanical drives) electrical drives 1.58) and general direct 212) 670 services drives TOTAL 3363 Energy consumed KJ x 1012 260.5 73.4 23.6 6.7 48.3 ) 22.4 ) 70.7 19.0 354.8 1 100.0 1 . 2.3. Energy Use in Drying Information on a breakdown of non-feedstock energy consumption in the chemical industry in terms of the actual unit operations in which they are used is difficult to come by, given the tendency in industry to assemble information on energy usage on a lumped basis. At best, most industries tend to identify plant areas which are more energy intensive but hardly go the whole length of detailed logging to obtain specific data in relation to the individual major operations carried out. However, it is well recognised that of the unit operations, distillation, evaporation drying, size reduction, mixing and materials handling (compression, pumping and conveying) are among the most energy intensive. In addition to being a long established operation in the chemical industry, drying has also found very extensive application in a broad range of other industrial sectors. These include food processing, textiles, pulp/paper, ceramics 18 building materials (cement, timber, bricks) and ores/minerals dressing. Besides, there are also some other strictly non-industrial sectors. such as laundering and crop drying, in which drying may also be an important operation. TABLE 1-12 : WATER REMOVED IT DRYING IN SOME MJOR U. K. INDUSTRIES(1975)8 Material Water removed tonnes x 10 Paper and board 9.2 Bricks 2.1+ Milk, dried 1.85 Milk, condensed . 85 Gypsum . 74 Plaster and plasterboard 1.0 Textiles 4 China clay . 35 Fertilisers . 12 Timber, softwoods . 12 Timber, hardwoods . 05 Dyestuffs . 05 Vitrified china clay pipes . 11 Tiles, pottery and sanitary ware . 15 TOTAL 17.1 Reay presented data on the quantity of water removed from the products of some of these industries in U. K. (Tablet-12). Notably, the data covers most of the pulp/paper, ceramics, textiles and building materials industries, as well as sections of the chemical and food industries. However, it is estimated that the industries for which data is not available may account for about 1. of all the water evaporated in drying operations in U. K. industry'9 On this basis therefore, a total of about 29-30 million tonnes of water is removed. Using a latent heat of evaporation of water of 2350 KJ/kg, the annual energy requirement for drying will then be about 70 x 1012KJ. But for most 19 industrial dryers only a fraction of the total heat supplied is effectively used up in water removal. If a nominal efficiency of 591is assumed, which is rather optimistic since quite often it is as low as 25-3 0P '9', 0 annual energy consumption in drying would therefore be about 140 x 1012 KJ. To put this in perspective, it is noted that compared with the data published by U. R. Department of Energy2 for 1975, drying alone would account for about 2.1+, 5 of total energy demand and 6. q of energy consumption in all industry in U. K. 1.3 Project Objective Deeply rooted in the principles of material and energy balances, the chemical industry in general has always been a relatively efficient user of energy. However, the challenges posed by the energy crisis have inspired renewed interest and effort in even more efficient utilisation of the increasingly scarce energy resources. It has been estimated that the potential exists for savings up to 103 million tonnes coal equivalent in U. K. chemical industry by the end of this century if all potential energy saving measures are fully implemented 11 . These energy saving measures range from minor 'good housekeeping' measures, through modifications to existing plants or processes, to changes in design of new plants or devising of new process routes, in an increasing order of capital cost, implementation time and energy saving potential4. The minor 'housekeeping' measures applicable to most processes and 'unit operations, including drying, are generally well understood and identified and are increasingly being implemented. The present study is therefore geared toward possible energy saving measures in drying, either at the level of modification of existing installations, or involving innovation in the design of new plant. The scope of work includes: (i) Examination of pattern of energy consumption in drying and ident-fication of major sources of energy wastage. (ii)Investigation of application of more conventional heat recovery methods (iii)Investigation of more innovative methods, like heat pump application. In particular, development of a solids absorbent based heat pump cycle with possible application in heat recovery from dryers. CHAPTER TWO DRYING PROCESSES 20 2. DRYING PROC : USES 2.1 Scope of Application As used in this context, drying refers to the removal of liquid from a solid containing liquid in an unbound form. Thus drying in this sense includes liquid removal by mechanical means, such as pressing and centrifugation, and also by heat and mass transfer processes (diffusion and evaporation), but excludes the removal of moisture from gasses by dehumidification or adsorption processes. (However, liquid removal by mechanical means is relatively ineffective in extracting a substantial amount of the liquid and is therefore almost always followed by a heat and mass transfer process when used. Indeed. heat and mass transfer is by far the most common method of drying and the consideration of drying here will therefore be confined to this method. ) Though the above definition refers to any liquid, in practice evaporation of water is most frequently encountered. This will therefore provide the basis for much of the consideration. Besides, systems that do not involve evaporation of water almost invariably involve evaporation of volatile organic solvents which generally have much lower latent heats of vaporisation than water (see Table 2.1), with the result that they tend to be less energy intensive and therefore of less importance from the standpoint of energy utilisation in drying. TABLE 2.1. LATENT HEATS OF WATER AND ORGANIC SOLVENTS1 3 714 Solvent Latent Heat . KJ at 25 c (kg) Water 2,240 Benzene 381 Methonol 1)013 Ethonol 851 Acetone 161 Ethyl ether 293 Ethyl chloride 311+ Carbon disulfide 272 Carbon tetrachloride 182 Latent Heat of water Latent Heat of solvent 1.00 5.88 2.21 2.63 4.86 7.6tß. 7.13 8.2tß 12.31 21 The breadth and scope of application of drying in iniustry have already been outlined in Chapter 1. "The reasons for drying and its role in processing could be as diverse as the materials -, ýhich may be processed. Obviously however, products must be in the most suitable form for storage, for subsequent processing, for handling or to meet product specification for marketing purposes, and drying often provides one of the most convenient means of preparing them to meet these N requirements . 2.2 General Drying Behaviour 2.2.1 Equilibrium Relationships When a moist solid is exposed (for a fairly long period) to a constant draught of low relative humidity air, the moisture content of the solid will approach an equilibrium value. This equilibrium moisture content will depend on the prevailing conditions, such as temperature, nature of the solids and relative humidity of the air. For certain materials, under most conditions all the moisture held by the solid is vaporised on attainment of equilibrium and hence equilibrium moisture content is practically zero, while for other substances there may remain a residual moisture content at equilibrium. These are referred to as hygroscopic substances. Wet solids may contain moisture either in the bound or unbound form. (Wet sand, for instance, contains essentially unbound moisture; ) cotton, starch and leather on the other hand may contain both bound and unbound moisture. In general coarse granular solids tend to contain mostly unbound moisture and have equilibrium moisture values of zero, while others, especially fine, porous substances usually contain bound 15, ) moisture and therefore tend to be hygroscopic, 16 (The importance of equilibrium moisture content in drying lies in the fact that it represents the limiting moisture content for the given temperature and humidity at which the drying is carried out'' Fig. 2.1 below shows the equilibrium relationship between moisture content of solid and the corresponding moisture content of the air for hygroscopic and non-hygroscopic materials. LG Fig. 2.1. Drying Equilibrium Curves Moisture in solid. Residual Moisture Content. 2.2.2 Drying Periods. ýIf the moisture content of a drying solid is recorded with time, a general characteristic behaviour is obtained as shown in Fig. 2.2. which applies to most solids. l Though Fig. 2.2 indicates that the rate of drying (moisture evaporation) is subject to variation with time or moisture content, this variation can be better illustrated if the drying rate, obtained by differentiating the moisture content versus time curve, C is plotted against moisture content or time as shown in Figs. 2.3 and 2.4.. These rate curves clearly show that the process of drying is not a smooth continuous one but rather consists of several distinct stages. This phenomenom has long been recognised and has been extensively investigated by several workers. 17i18) Thus, the drying period can be formally divided into three periods: (i) an initial phase, represented by segment AB of the rate curves, during which the solid temperature adjusts to its steady-state value. It is usually so short that it is ordinarily ignored. (ii) Constant-rate period, represented by section BC, and (iii) Palling rate period, represented by section CDE. 2.2.21. Constant Rate Period In the constant-rate regime, drying proceeds at a fixed rate which represents the maximum possible rate attainable under the preva-4lin¬ conditions. The entire exposed surface of the dr_: , nC, solid is saturated and drying is thought to take place by diffusion of vapour from the 23 Fig. 2.2. Moisture Content vs Drying Time. Moisture content x XE tame e equilibrium moisture content. J Fig. 2.3. Drying Rate vs Moisture Content C B A Drying rate D N E/ XE Moisture content X Drying rate N time e 24 saturated surface across a stagnant air film into the environment. Moisture movement within the solid is rapid enough to maintain a constant saturated condition at the surface. The rate of dying is controlled by the rate of heat transfer to the evaporating surface and is completely independent of the nature of the solid. This is well- illustrated by the observation made by Sherwood and Comings19 in which they pointed out that under similar drying conditions many substances tend to dry at approximately the same rate in the constant- rate regime. TABLE 2.2 : MAXIMUM EVAPORATION RATES FROM MATERIALS IN A PAN WITH CROSSFLOW AIR AT 3 m/s and 500C. 19 Material Rate ofEvaporation kg/m' - hour Water 0.27 Whiting pigment 0.21 Brass filings 0.21 Brass turnings 0.21 Fine sand 0.20-0.21. Clay 0.23-0.27 The rate of mass transfer balances the rate of heat transfer in the regime and the temperature of the saturated surface remains constant. Thus the drying rate will be given in terms of mass transfer as: N= -dX =K A( P- P) - --------2.1 c dGO- 93 where Nc= constant rate of drying X= moisture content of solid Q= time A= area for evaporation K= mass transfer coefficient 9 p= partial pressure of vapour in air ps= vapour pressure of moisture at surface temperature of solid. 25 and in terms of heat transfer as: X Nc =-ä£= htA (T -T s) ---------2.2 a where ht = total heat transfer coefficient X= latent heat of vaporisation at surface temperature T air temperature Ts = solid Surface temperature. If heat is transferred solely by convection and in the absence of other heat effects, the surface temperature of the solid equals the prevailing wet-bulb temperature of the air under constant-rate drying conditions and ht is the convective heat transfer coefficient hc. If the heat is also supplied by radiation, as is sometimes the case in very high temperature drying, ht will be the sum he + hr, where hr is the radiation coefficient and Ts will be higher than the wet-bulb temperature. For conductive drying through hot surfaces, the solids temperature approaches the boiling point temperature rather than the wet-bulb temperature. ) During constant-rate drying, it is established that the rate of drying depends on the velocity, and hence flowrate, of air. For air flowing parallel to the solid surface in cross-circulation batch dryers, extensive experimental results have been satisfactorily correlated by the relationship 20,21, o. 8 ht - 0.0128G ------2.3 where G is airflowrate. For airflow perpendicular to the drying surface Molstad et a122 proposed the relationship ht =o. 37 G0.37 ----------2.4 Treyba]23 has suggested that equation 2.3 above can be reliable only within the mie 500 100°C) . Besides, at high temperatures further problems are encountered, as most. working fluids: (i) tend to be chemically unstable, (ii) readily dissolve i co. -pressor lubricants, (iii easily interact -v ith meta surfaces with which they are in contact. 103 Regarding the open-cycle steam recompression technique, it is obvious that it cannot be feasibly applied to drying below about 1)00C since the steam must be substantially superhea;, ed. On its part, practical application of AbHP to drying would also be beset by several severe problems, including: (i) many of the suitable solutions, notably the H2S04, KUH and NaOH are very highly corrosive and therefore create special handling problems; (ii) the solutions tend to be entrained or volatize, giving rise to liquid losses and problems of contamination in the dryer; (iii at high temperatures crystallisation boundaries may be approached leading to solidification problems in the circuit or unfavourably high fluid viscosities may be obtained. In view of the foregoing limitations an adsorption heat pump (AH? ) cycle is proposed here for heat recovery from dryers. The AHP, which is similar to the AbHP and CHP, consists of twin columns of solid adsorbent material, such that while moisture is being adsorbed on one from the dryer exhaust air the other is being desorbed. When the processes come to completion (i. e. the adsorbing bed is used up and the desorbing bed is stripped out), the columns are interchanged and the processes continue, with the stripped bed adsorbing and the used bed desorbing. The adsorption process which is exothermic replaces the absorber while the desorption process replaces the generator of an AbHP. The air is simultaneously heated by the heat generated in the adsorbing column as it is being dehumidified and is recycled to the dryer. CHAPTER FIVE ADSORPTION 1 04 ADSORPTION 5.1 Introduction 5.1.1 Nature of Adsorption Certain solids posses the unique property of preferentially concentrating onto their surfaces one component of a mixture from fluid solutions with which they are contacted. Adsorption is concerned with the exploitation of this property to effect separation of components of a liquid or gas mixture. The adsorptive process is specific and selective in nature. On contacting the solid and fluid phases one or more constituents of the fluid may be selectively removed from the stream while the rest pass through. while the material adsorbed is 5.1.2 Types of Adsorption The solid is referred to as adsorbent the adsorbate. It is necessary to distinguish between two types of adsorption phenomena: - - physical adsorption - chemical adsorption (chemisorption). (i) Physical Adsorption: - This results from intermolecular forces of attraction between molecules of the adsorbent and the adsorbate. The balance of these intermolecular forces is such that adsorbate is transferred from the fluid phase to the adsorbent phase. The process continues until the partial pressure of the adsorbed material equals that of the contacting fluid phase when equilibrium is deemed to be attained. By lowering the pressure of the fluid phase or by raising the temperature the process could be reversed and the adsorbent stripped of adsorbed material. (ii) Chemisorption :- Unlike physical adsorption, chemisorption results from chemical interaction between the adsorbent and adsorbate molecules and is frequently irreversible. The heat effects associated with this type of adsorption are usually large relative to those associated with the physical adsorption process. Chemisorption is generally unimportant in most industrial adsorption rocesses but often plays a significant role in catalysis. 105 5.1.3 Properties of kd sorbents Adsorbents are generally characterised by ahg 1y- porous solid structure, providing large surface areas per unit weight of material. The pores are usually very small, ranging from about .1 um to 50 X-- but their large numbers furnish enormous surface for adsorption. The performance of an adsorbent does not only depend on the nature of its constituent material, it also depends substantially on the method by, which it is manufactured as its crystalline structure and particle size are affected by the method of manufacture. Performance may also depen; on the previous history of an adsorbent since most are prone to deterioration with constant use. Adsorbents are made in three forms - powder, granules and beads. The powder has a very high specific surface area but gives rise to excessive pressure drops when packed in beds and is prone to dusting and entrainment. Though the beads on the other hand possess relatively low contact area, they are usually hard and rigid, enabling them to provide desired structural strength to withstand attrition. In general, the granular form possesses properties which lie between those of the powder and the beads. This commends it specially for widespread industrial use, in particular for operations involving gases. The main adsorbents in general use industrially include clays. bone char, activated carbon, activated alumina, silica-gel and molecular sieves (synthetic zeolite). Table 1 below shows the essential properties of three of these adsorbents (granular form). TABLE 5.1 : PROPERTIES Or AD3ORBEIý'T0 9' 50 ADSORBENTS PROPERTIES Activated Carbon Activated Alumina Silica Gel Specific surface area (m3/kg) x 10-3 1100 390 850 Pore volume (m3/kg) x 103 0.56 0.35 0.46 Average pore climate (A 0) 19 4-8 22 1 Bulk Density (kg/x'13) ý 1+80 800 . 20 Specific Heat (KJ/kg_0C) 0.84 0.92 0 '. 05 ,, o6 5.2 Adsor-ption Equilibrium Relations The capacity of an adsorbent for an adsorbate is limited by the static equilibrium relationship between the concentration of the aýsorbate in the fluid phase and in the adsorbed phase. Static e cuilibrium in the system can only be attained if contact between the two phases is sustained indefinitely. In practice, when fluid flows continuously through a fixed bed of adsorbent static equilibrium cannot be achieved but rather dynamic equilibrium is established. 5.2.1 Presentation of Equilibrium Relations The loading of adsorbate in the adsorbent at equilibrium is believed to be a function of the partial pressure or concentration of adsorbate in the fluid and the prevailing temperature only, i. e. W=f (p, T) (5.1) where W= loading p= partial pressure and T" temperature. When p is varied and T kept constant, a plot of 'ýý against p is called the adsorption isotherm (fig. 5.1), v i. e. W=f (p) T= constant (5.2) When p is constant and T is varied, the plot is the adsorption isobar (fig. 5.2) Wf (T) p= constant (5.3) The plot of p versus T at constant loading is the adsorption isostere (fig. 5.3). pf (T) VT = constant (5.4) The presentation of adsorption equilibrium data can be simplified by plotting the loading against the relative humidity of the adsoröate in the fluid (fig. 5.4), such that relative humidity is defined by the ratio x (5.5) Po where p0 is saturation vapour pressure of the adsorbate at the prevailing temperature. This, therefore, combines the effects of p ar. '=' on 7, '. 107 FIG. 5.1 - ADSORPTIOI; I SCý. N: S 'W p FIG. 5-3 : ADSORPTION I SO STERES P / w w FIG. 5.2 : S0Plf: Ol; ISOBERS FIG. . 1}: LOADING vs 108 5.2.2 Vonolayer Theory The monolayer theory is one of the theories that have been deduced to explain the mechanism of adsorption. It postulates that in the process of adsorption adsorbate molecules are retained as a surface layer on the adsorbent, one molecule thick. For very low concentrations of adsorbate in fluid, it is believed that the isotherm may be linear, similar to Henry's law equation. Then, W= Hy (5.6) where y= adsorbate concentration in fluid and H= Henry's constant. Following his empirical studies, Freundlich 52 suggested that for monolayer adsorption over a wide range of y, the isotherm is such that it is concave to the y- axis (fig. 5.5) and hence he proposed an equilibrium relationship of the form W= Kyn (5.7) where Kn are empirical constants. 'LIT f4 r, r- " PA7TT1TIIT. T(TFT1 e TS(VPT-t1 PM w Y 09 Consequently, the Henry's Law relationship can be re¬a ed as a special case of Freundlich' s exponential equation, when n=1, F reun:: =ich: ' s equation has been found to be more appropriate for correlating the equilibrium adsorption of liquids than vapours. From theoretical considerations Langmuir53 developed an alternative equation for this isotherm: WmKy 1+ Ky where Vm, K are constants. (5.8) On comparing experimental values for adsorption of various vapours with calculated values using Langmuir's and Freundlich's equations better agreement was obtained with the former. However, at higher pressures experimental values fail to agree so well with Langmuir's equation. 50 5.2.3 Multilager Theory It has been observed that in some cases the adsorption curve is S-shaped and increases to infinity as the partial pressure approaches the vapour pressure 55 This has been explained by the formation of a multimolecular layer of adsorbate molecules, of definite thickness. This type of equilibrium has been observed for the adsorption of water vapour on carbon black and of nitrogen on iron catalyst and is similar to the monolayer curve for low and intermediate partial pressures. Brunauer, Emmett and Teller 56 derived an isotherm equation for multimolecular layers that is similar to Langmuir's equation for unimolecular layers thus: - WmCy 11-(n+l)Yn+nyn+l] W= 1-Y . [1+(c_1)yCy1J where n= number of layers C is a constant given by C e(El -EL)/RT Eý is heat of adsorption EL is normal heat of condensation (5.9) R is gas constant 110 w When n= °° equation (5.9) reduced to W= Wm C% (5.10) (1-Y) 1+(c-1)Y In the other extreme, when n=I (only one layer adsorbed on surface) equation (5.9) reduces to Langmuir's equation (5.8). 5.2.4 Capillary Condensation Following his studies on the adsorption of water on silica-gel, Zsigmondy57 concluded that the adsorption process is a function of the capillary properties of silica-gel. This provided the basis for the theory of capillary condensation which has been used to explain certain adsorption phenomena. Provided that the fluid wets the adsorbent, leading to the formation of a concave meniscus, the vapour pressure pc over the condensate confined in a capillary is related to the saturation vapour pressure pv in free space, at the same temperature, by the Kelvin expression: lnPc =-20V (x. 1 Po r RT o is surface tension of condensate v= molecular volume in the liquid s;, &te r= radius of capillary PT r- r, ir. - WMI r"T), '('ýT. 7rT-T,, P A . 11 All the quantities of the right hand side of ecuation 5.11 are positive, and hence PC . 01) (5.31b) Dehler100, on his part, proposed that breakpoint capacity could be obtained as the mean of the equilibrium loadings corresponding to the inlet and outlet fluid streams. If adiabetic conditions obtain he suggested that allowance would be made for the temperature effect, by assuming that it modifies the relative humidity of the exit fluid. 5.8.2 Effects of Operating Variables on 'Aºb Several investigators have studied the effects of different operating variables on break-point capacity. Grayson 8, as well as Getty and Armstrong 98, working on adsorption of water vapour from air on bead silica gel and activated alumina respectively showed that increasing depth of adsorption bed increases breakpoint capacity while increasing air flowrate tends to reduce breakpoint capacity. Simpson and Cummings and Worthington 67, have also presented results that show an adverse effect of flowrate on Wb. On the other hand, Moss and MacLaughlin101. and Miller and Roberts 102, produced experimental data for adsorption on silica gel and alumina respectively, which indicate that breakpoint capacity would be independent of air flowrate. Worthington 67 investigated the influence of inlet humidity on the breakpoint capacity during non-isothermal adsorption of moisture from air on silica-gel. He concluded that higher humidities lead to reduced capacities. Grayson8 arrived at a similar concLusýon following his experiments using bead type silica-gel. In addition, he indicated that under certain conditions, such as large bed depths, the reverse effect might also be obtained. Miller and Roberts 102 on the other hand, suggested that breakpoint capacity would be independent of inlet absolute humidity. Defining the breakpoint bed relative humidity Xb as the ratio of adsorbate partial pressure in the feed to the saturation vapour pressure of the adsorbate at the mean tempera, ure of the bed Turing operation, it, has been suggested that a rely : io nship exis -ýz between and ý'; b For the water/ activated alur.. ina s;. -stem, l: i11er a nc .. ober-s02 1 29 FIG. 5.1 1: MILLER /RD BERT S DATA 11, -A % DT 4I MAL cow K 12 Wb SEMI-ISOTHERMAL COLUMN INLET DEW- POINT OF. "o 43 o"6! 0 96 " SO " 56 O 63 0 46 40 64 $0 100 I Cri FIG-5-15: WORTHINGTON'S DATA ýs R ýýýý -ý- rr 1 30 as well as Coulson and Richardson3 correlated Xb and 'º"b wi. a linear relationship. The data of Miller and Roberts are reproduced in fi¬. 67 5.14. ', '; orthington obtained data for water/silica-gel, reproduced in fig. 5.15, which indicate a non-linear relationship be ý' een Xb and Wb. 5.8.3. Teiperature Change Since the temperature wave is believed to precede the adsorption wave, it may be considered that while the bed attains mass transfer breakpoint, thermal equilibrium is established. This implies that heat transfer between the fluid and the adsorbent is substantially complete at breakpoint and thus the fluid and the solid attain about the the same average temperature Based on this, Miller and Roberts derived an equation for the temperature increase associated with non- isothermal adsorption, thus: Wb xA Ay ýT-ýP c+c bPP DY gs where Ay is change in adsorbate concentration. If the influent fluid is relatively dilute, Wb cp » cp Ay gs and hence, a Ay 0T=A Pg Equation 5.33 has been proposed by Carter65 (5.32) (5.33) and by Coulson and Richardson. Based on his experimental studies on adsorption of water on silica- gel, Worthingto suggested the following empirical equation for obtaining AT for the water/silica-gel system: 0 AT = 1000 where h 1000 is the moisture adsorbed in lbs/1000ft air and LT is temperature rise in °F. Similarly, Mantell50 suggested that the adsor. tion of 4.9 grains of water on silica-gel results in a temperature rise of about 1C °F. For the same system, thmell and Bateman l ottsined the folloring relation: X31 AT= 293oyi - 3.5 (5.35) provided that inlet, humidity yi lies in the range . 003-. 015 x&/kg and AT is expressed in 0C. For water/activated alumina, and within the same range of inlet humidity they obtained: T= 3500 Yj (5.36) 5.8.4 Regeneration Fre9 studied the conductive regeneration of silica-gel and alumina beds using embedded electrical heating coils extensively and suggested that the desorption requirements can be predicted from the movement of the desorption zone. He was able to correlate the speed of the desorption zone with the purge air flowrate for both types of adsorbent using linear relationships. For convective regeneration, Carter 65, Mantell50, as well as Hougen and Dodge 104 , have outlined closely related calculation procedures based on heat balances. The heat reauilements for regeneration are deemed to comprise essentially: (i) heat for vaporizing adsorbate at the temperature at which -most of the desorption takes place; (ii) sensible heat required to heat the bed to the regeneration temperature. In convective regeneration, these heat demands are evidently provided by the sensible heat loss of the desorbing air. Geankopolis105 has suggested that unlike in drying of solids in shallow beds, the temperature at which regeneration of thick beds takes place may be considerably higher than the air wet-bulb temperature. Rathmell and Bateman 1o6 indicated that this temperature, and hence the humidity of the effluent air, are directly related to the initial capacity of the used bed and the results presented by Worthington 67 for regeneration of silica-gel beds using hot, dry air in the temperature range 100-2500C agreed substantially with this. CHAPTER SIX THE ADSORPTION HEAT PUMP (AHP) 132 6. THE ý. DýORP IGI 3Ar, " BUT: (ziP) 6.1 . System Description The LHP may be operated either on a: (i) Temperature-Swing Cycle (TSC), or (iiý Pressure-Swing Cycle (PSCj. 6.1 .1. TSC Operating on a TSC, adsorption takes place at low temperature while desorption takes place at an elwated temperature, both columns being maintained at the same pressure, usually atmospheric. The dryer exhaust air is passed into the adsorption bed directly. A portion of the dehumidified air is returned to the dryer while the rest is heated to a high temperature and used to regenerate the other bed. A schematic diagram of the cycle is presented in fig. 6.1, while the successive steps involved in the cycle are illustrated on a p-T diagram in fig. 6.3. 6.1 . 2. PSC This is slightly more complicated than the TSC in that it incorporates a compresor, a cooler and an expansion valve. The dryer exhaust air is compressed and passed through the cooler, where some of the moisture is condensed and removed, before it goes to the adsorption column. After adsorption the air is expanded and while a fraction is returned to the dryer the rest is used to regenerate the other column at the reduced pressure (atmospheric). The need for air heating before desorption is thereby eliminated. Figs. 6.2 and 6.4 are a schematic and p-T representation of the system respectively. 33 TRe, YRe Desorbing Column TRi, YRi Ltion ýr FIG. 6.1 : TSC HEAT FIT DPY EP '1 34 Re ºesorbing column YRi FIG. 6.2.: PSC HEAT . PUS' DRYER. 135 5 P 3 \\ 6 FIG. 6.3 : p-T REPRESENTATION OF TSC 1-2-3: mixing of fresh air intake and recycle air to dryer. 2-4: heating air mixture in the dryer heater 4-5: drying 5-3" adsorption 3-6. heating regeneration air 6-7. regeneration ; 36 I T FIG. 6 .L .: p-T REPRESENTATION OF PSC HEAT. PU IT DRYER 1-2-3 : mixing of fresh air intake and recycle air to dryer. 2-4 : heating air mixture in the dryer heater 4-5 : drying 5-6 : compression 6-7 : cooling/condensing 7-8 : adsorption 8-3 : expansion 3-9 : regeneration 137 It is observed that there is a break between the air intake and air exhaust points I and 7 in fig. 6.3 and 1 and 9 in fig. 6.4, which may suggest that the cycle is not closed. This is not so. Basically, since the equivalents of the aisorption section and the generation section of an analogous closed-cycle absorption heat pump are present (adsorption and desorption respectively) the cycle can be considered to be closed. The moisture and carrier air stream is not brought back to the initial conditions at the end of the cycle however, as would be expected in a closed cycle, because there are no condenser and evaporator components. The peculiar nature of this application of the heat pump is such that the moisture from the dryer is already vaporised as it enters the heat pump and that removed in the desorption process is substantially degraded in energy, thereby obviating the need of an evaporator and a condenser. 6.2. Performance Evaluation The performance of the AHP in terms of its ability to upgrade the low quality energy in the air leaving the dryer is measured by the COP while the overall performance of the AHP dryer is measured by the cycle thermal efficiency. 6.2.1. COP Similar to an AbHP operating between temperatures TR, Th and To, the ideal COP of an AHP operating between the same temperature limits will be expressed as: Th (TR Tcý (6.1 COPm _( Th - Te ( TR where TR represents the temperature of regeneration Th represents the temperature of air leaving the adsorption bed, and Tc is temperature of air leaving the drying chamber. However, for a real cycle, COP must be expressed as: **Absorrtion Heat Pump X36 (6.2) COP = energy absorbed in the adsorber energy input for regeneration and hence, thermodynamic efficiency of the s. ystem be: COP OP (6.3) nT-C max 6.2.2. Thermal Efficient The overall thermal efficiency of the AHP dryer differs from equation 6.2 in that this refers to the evaporation of moisture from the dryer "rrith respect to the net energy input and hence will be expressed as: n_ heat used for evaporating moisture (6.4. ) net energy input From equation 3.21 the specific MER will be given by: n CE _A W 6.3 Equilibrium Model (6.5) In the simplest form, the AHP dryer can be considered to be such that: (i) there are no heat losses from the system and hence the processes are adiabatic; (ii) heat and mass transfer processes in the system go to completion and hence no consideration is given to 't-Q". v11 avv^. C equilibrium constraints. Thus, the system may be described by simple adiabatic heat and mass balances, enabling the maximum energy efficiency of the system to be determined. Consider the TSC dryer of fig. 6.1. such that the temperature and humidity condi ions are as s ov; r. in the diagram. '1 39 6.;. 1. Di er Let the amount of air recirculated to the dryer for 1 kg. air trrouE, h the dryer per unit time be rD. By overall moisture balance on the cycle. (1- r DJ (YRe Yo) = YDe i rD= (YRe ' Yo) -( De ' ID. ý (6.7) ( Re - Yo) moisture balance on dryer heater, _ (1-rn )Yo +rD Y (6.8) rD yDi '0 (6.9) yA yo e from 6.7 and 6.9 yD1-yo yRe - yo) - (yDe -"Di) (6.10) YAe-yo YRe Yo) yDe 3'D + MAe - 3rDi YRe-yo (6.11 i Ae yo The fractional saturation of air in the dryer f is expressed as: f= yDe ID. 1 (6.12) ysat - yD. i where Ysat - saturation humidity SD _ (1-f) YD +f Ysat ei from 6.11 and 6.13, YD + (YRe- Yo) (yhe- YDiý 1= y.. e-Yo ( 6.13) ysat ( 6.141 140 hence, YD. _y (Yr. - yo) - Yn e (YR e yo ) D. -sate (6. '7) f CYO _y) - CYO Y o-o ee saturation pressure in the dryer is obtained from: _ 291 y psat sat (6.16) 18 + 29 ysat where r= total pressure. hence, from Antoine's equation, _ 3883 - Tsat 1$. 4. - In p 230 6.17 sat Also, by simultaneous heat and mass balance on the dryer, operating under adiabatic saturation conditions: a w Tsat TD. C (Ysat YD. ) (6.18) 1 pg 1 a and TD e= TD l_cW (YD YD. ) (6.19) pge1 where AW is latent heat of evaporation and cpg is specific heat of air 6.3.2 Adsorption By heat and mass balance in the adsorbing column, Tie = TA. + _w (YA. _ Y. ) (6.20) c1e Pg but TA = TD ie and yA . YD ie hence , ýD A e-e+ ý- `= (D-'A) (6.21) Fý ee 141 6.3.3. Regeneration Suppose air is finally discharged 1OUlo saturated at a humidity YRe from the desorbing column, then the partial pressure of moisture in exhaust air will be given by: * 29 ii yR PRe e* (6.22) 18+29yRe and the temperature by : T* 3883 - 230 (6.23) Re - 18.4 - In pR e If all the heat given up by the hot air is used in desorbing moisture, by heat and mass balance on the regenerating column: T* - TR. -cW (yRe - YRi (6.210 Ip g Now if the air is actually discharged with an approach to saturation of SR %, then actual partial pressure of moisture in exhaust air will be: e PR pR SR eee (6.25) 100 and 18 ( PRe ) yRe = 29 ( II ) (6.26) PR e hence, TR _ TR. W (VRe- SRi) (6.27) eip ß 6.3.4. Cycle Performance Following the definition of equation 6.4 and noting that total heat input to the system consists of heat supplied to the dryer and regeneration heaters, the thermal efficiency of the cycle can be expressed as: n pý ( Di - De } (6.28) EM = cp (TD i -TA, ) + Cpg (1 - rD (T. -To }+ cpg(1- D) ( Ri-1 e} ge 142 n_ TD - TD EM ý, 11 (t 29 ) Di - Th . e+ (1-rD) (TR -oý 1 The ideal COP of the TSC as a heat pump v. ill be: COP _( TAe ) (T R. - TD max ) (TA _ TD (1e (6.30) eeTR. I and the actual COP will be: COP =TAe- TDe (6-1 r -T T 3) D R. A Ie 6.3.5. PSC Heat Pump Dryer The heat and mass balance equations derived for the TSC heat pump are also essentially applicable to the PSC heat pump. Ho.. ever the expression for thermal efficiency differs significantly from that of the TSC in view of the inclusion of a compressor and cooler and elimination of the regeneration heater in the PSC heat rump. In this regard therefore, the thermal efficiency of the PSC heat pump dryer will be defined by: cP (TDl-TDe ( D" _) (6.32) nEM Cp TTA+ Cp T (.: ý _r) (AT - o) +E -E gie9De comp re c where Feomp is energy input to compressor and Erec is energy recovered in the cooler. Similarly, the COP of the cycle will be expressed as: - COP =cps ý Tie TA i) EE (6.33) comp rec The ideal COP using the notation of fig. 6.2 will be : COP max (TAe )(TRi -TDe) (TA - TD 5( TR (6.3+) ee1 143 6.4 Lee-Cummings Procedure this model, the equations previously derived in res- For ect of the dryer in section 6.3.1. will also be applicable, noting hog: er that humidification of air may not take place along an adiabatic path. The equations for evaluation of performance, out ined in Section r'. 3.4. will also hold. 6.4.1. Adsorption The processes taking place in the adsorption column be described by an adaptation of the procedure proposed by Lee and Cumnings9 discussed in Chapter 5. The calculation procedure consists of the following steps: 1) Calculation of bed Reynold's Number NReb from: NReb DýGv 1 61-e Vý (6.35) where D is average adsorbent particle diameter P u is fluid viscosity is bed void fraction is particle shape factor and v is air mass flow velocity. 2) Calc ulation of Colburn factor jD : -0.51 j = b< 50) 0.91 (NR b) (NR (6.36a) D e e -0. x. 1 or jD = ý 50) 0.61(N b ) (NR ( 6.36b) e Reb 3) Calc ulation of gas phase mass trans'., ---r coefficient kga from: )2/3 (u Kea _ = jd v. a (D AB k (6.37) g where v is superficial air velocity DAB is gas phase . mass diffusivity is air density ß 4) Calculation of modified distribution ratio K. 144 K e= MRS ýý"3ý) ysat* ß ts is packed bed density M is slope of linear equilibrium isotherm in a clot of capacity vs percent humidity and ysat is saturation humidity evaluated at air irle , temDer--pure to the adsorber TA, Z 5) Calculation of dimensionless time ratio a a=1 .0+1 . 66 (1 ooo y)0.87 Cry <0.01) (6.39a) or, a=3.0 (Ay > 0.01) (6.39b) 6) Calculation of constants C1 and C2 given by: C1 = 1+. 66 a-2.30 (6.40) and C2=0.07 a+1.32-Kga gb (6.11) KE where Abis breakpoint time. Calculation of bed height modulus MZ MZ =2 (C1 2-2 C2) + C1 C1 2-4 C2 ( c. 42ý 8) Calculation of bed height Z Z_ MZ .v Ka (6. x'3) g This procedure enables the size of the bed and hence the breakpoint capacity of the bed, but condition of the outlet air TA e. Now, if it is thermal equilibrium is approximately attained bet end the solids in the adsorption bed, then: to be deter. -ned, not the temperature considered that Neer. the exit air AeT Sb where T is temperature of solids at bre-k-poi: ý,. s By a 1urp; ' heat ball. "ce cr. the column at breakpoint: 145 Ms cp s (TA e- TA Z)+c cpg(TA e- TA GG'y ?'A (6. `ý5 where Gy yA. - yA (c.; vý Ie and c is heat capacity of solids PS G is flowrate of air Ä is heat of adsorption 9b is breakpoint time Ts is initial bed temperature. I By moisture balance Ms "tW =G GbAy (6.47) where 0 'Iv = Wb - W1 (6-48) and Wi, Wb are loadings of moisture on adsorbent initially and at breakpoint respectively; MB is mass of the bed. If the bed is initially fully desorbed, then `, b MS G (6"ß+9) From 6.45 and 6.49 T_b yaA + ýp Ttii) + Z_Y CPS TSi Ae Qy Aw---c (6.50) -. 10 Ps+ pß This equation, which is similar to equation 5.32 proposed by Miller and Roberts102 can thus be used in conjunction , with Lee-Curlings procedure to predict the performance of the adsorption column. 146 6.4.2 Regeneration Fo11o';: ýnC the discussion of the nature of the tem_erature, humidity and loading changes and of equilibrium relations in desor: pion (chapter 5), the process can be described by a com'in tion of an appropriate equilibrium relationship and a lumped heat balance analysis. The initial loading of moisture or. the desorbing bed will be expected to be the breakpoint capacity obtained in t: e preceding adsorption and is considered to be in equilibrium with the exhaust air. If the desorption relative humidity is defined as a function of the exhaust air humidity and temperature yR and TR resnectively then any suitable system equilibrium relationsiip can to represented by: WWT2 (6-51) D where W is the initial loading on the desorbing bed. By assuming e thermal equilibrium between the desorbing bed and exhaust air, TRe _ TSD where T., is temperature of desorbing bed. JD A lumped heat balance on G6b Cp (TRl _ TRe) _ g hence, the bed therefore gives, a Ms Cps TRe - Ts. +A G8b yRe = yR. +G9b CP (TR1 - TRe) - (6.52) (YR YR. ý ei 1 1C p TRe-TS, s (6.53) (6.5k) Combination of equations 6.51 and 6.51 would enable the exhaust air conditions to be determined and hence performance of the system can then be readily evaluated. 117 6.5. Dehler Method The concept of length of unused bed (L3) am equilibrium section in isothermal adsorption enunciated by Collins95, has been 100 discussed in Chapter 5. 'výith Dehler's empirical deduction, which was also discussed in Chapter 5, this concept can be extended to non-isothermal conditions. He proposed that under non-isothermal conditions half of an adsorption bed may be deemed to be at equili. rium with the inlet air while the other half, comprising the LUE, v: ill be at equilibrium with the outlet air. This can therefore be alternatively used to represent the adsorption section of the iF under considera t ion. The relative humidity of inlet air to the adsorber can be expressed as: Xe. = PA. 1 (6.55) A. i where p, is partial pressure of moisture in inlet air, and pAo is saturation vapour pressure of moisture at air inlet temperature TAi" Thus, using an appropriate equilibrium equation, the loading of moisture at the inlet half of the bed may be given by: Wei =f (Xe. ý =f ýYA. TA. ) 1 1,1 (6.56) At the outlet end, relative humidity would be affected by the temperature effects associated with adsorption, which need to be k, --o,. -: n. The temperature rise may therefore be calculated usinc a procedure 67 such as the 'ý, orthington empirical equation A 70 h1000 v; he re h1000 is the moisture adsorbed in lbs/1000 ft' and AT is temperature rise across the bed in °F. in terms of actual humilities and in °C, then, AT 12.25 x 103 ^v TA A. + 273 I (5.34) air Lxsressing this (6.57) -148 Hence, TA = TA. + ý1 2.25 x 103 a y) ei TA. + 273 i (6.58) Thus, in a similar manner as the inlet end, relative humidity of exit air will be: x= PA ee e (6.59) and hence the loading at the outlet half of the bed will be expressed as: We =f (Xe )=f CYA ' TA eeee (6.60) Breakpoint capacity of the bed can then be obtained using Dehler's equation thus: VTb =2 (w +w ) ee ie (6.61) CHAPTER SEVEN EXPERDENTAL 149 EXPERIMENTAL 7.1 . Objectives An experimental programme was undertaken, with the following objectives: (1 Demonstration of the simultaneous dr-bring of humid air and temperature lifting (heat pumping) in an adsorption column, under conditions that simulate those that may be expected from a solids dryer. (2) Observation of the temperature and humidity profiles in the fluid and solid phases during the adsorption process and during the subsequent regeneration of the bed. ý3) Investigation of the effects of operating parameters, such as - air flowrate - air temperature - air humidity - bed height on the adsorption process, especially under conditions of high temperature and humidity. (4. ý Correlation of results to enable theoretical predictions from the models outlined in the previous chapter, which were derived from various proposals made by other investigators, to be tested against the experimental data obtained. 7.2. Design of Apparatus The experimental apparatus, which was purpose designed and erected for the study, consists essentially of three main items of equipment: - air blower - air heater - adsorption column. 15 The blower is a purchased item while the heater any the coiu. n were in-house designed and fabricated. Air is taken in by the blower., passed through the heater and then into the adsorption column. A steam junction is provided before the entrance to the column, to enable steam from the central mains to be mixed with the inlet air, in order to increase the air humidity, if desired. r_ schematic flow diagram of the apparatus is shown in fig. 7.1. 7.2.1. Air Blower The blower is an electrically driven, centrifugal device, with the following speed and power specifications: speed : 9500 rpm power : 0.75 hp (0.56 kw) The operating characteristics, flowrate and pressure head, are shown in the Appendix (C1) . 7.2.2. Air Heater This consists of a 1+0 x 40 x 1+8 cm. rectangular steel box in which 3 electrical heating elements, each of a power rating of I kw., are mounted. By connecting the elements in series, the temperature to which the air is heated can be readily controlled. The walls of the box are lined with 8 mm. thick asbestos sheets which are intended to serve as insulators, in order to curtail heat loss from the heater to the environment. 7.2.3. Adsorption Column The adsorption column consists basically of a 1.5 m. high, 23.5 cm. internal diameter and 10 mm. thick mild steel cylindrical shell made from a standard pipe section in which the solid adsorbent is packed. The solid is held on a perforated support grid which is mounted 25 cm. from the bottom of the column. The grid is supported on three cylindrical legs which rest on the bottom flanke, thereby providing the structural strength that enables it hold the weicht of the bed. 1: 7.5 cm. charge hole is provided on the to, flange for feeding in the adsorbent and two 4. cm. bore holes arc fitted on the shell, 15 cm. from the top and bottom flanges respectively, to serve as moist air inlet and dehu: id f ec sir outle ý. Five 6 rr:. bores are also fitted to the shell, at 25 CI T.. intervals, for ter. -Der-: eure mea,: rements along the bed. A 12 rr. diar--. for vent, ate ýo_ `_la ýe 151 W c* F. J. 0 cD (a g b N cf- (D m H ?1O -ý1 c+ 11 Of I a O CA 00 N" "O -J. CD P) c+ CD Co cD cD ct º1o x to CD H cD N" co C) r_ P pi C+ CD CD '-3 cD cD w C+ (D (D PD ca C 1 CD Ei (D c+ 1 e-4 i5* I-' O Ic cD E3 P c+ P. O 0 N- W g 0 m Pl N" (0 25 tu G to c-º a ýi f Ii Plate A: EXPERIMENTAL APPARATUS 152 Fig. 7.2 : Adsorption Column Dry air outlet m. therm coupl point Moist air inlet. 153 is designed for air venting in case of accidental pressurization of the column while a 12 mm. diameter drain line is located at the bottom flange to enable any accumulation of water to be drained away. A thin gauze sheet is also placed on top of the bed to prevent entrainment of the solid particles in the air stream. 7.3. Instrumentation and Procedure. 7.3.1. Flow Measurement The flowrate of air is measured with a Metrix X series, 65X tube size rotameter, with a type A duralumin float. A calibration chart provided by the manufacturer, which is shown in the Appendix (C2)ß is used to calibrate the rotameter. The flow of air is controlled by a damper which is situated at the suction side of the blower and by a globe valve fitted between the blower and the heater. _A valve is also installed between the heater and the steam inlet, which stops moisture from drifting back into the heater, thereby preventing possible electrical hazards. The admission of steam is detected by means of a pressure gauge located near the inlet point. A main steam inlet valve and a bleed valve enable the flow to be controlled. 7.3.2. Temperature Measurement Chrome-alumel thermocouple wires are fitted to the adsorption column by passing them through rubber bungs which are used to plug the temperature measurement tap points provided on the shell. Care is taken to place the ends of the thermocouples at the centre of the column axis and adsorbent is carefully packed around them. The other ends of the thermocouple wires are connected to a multipoint digital thermometer which in turn is connected to a switch selector unit. The thermocouples are thus used to measure the temperature of the bed at different positions along the length of the column during operation. The air inlet and outlet temperatures are also measured using thermo- couples fitted at the air inlet and outlet lines. 154. 7.3.3. Humidity Measurement Manny types of devices are available commercially for measuring the concentration of water vapour in gases. These include the dew point cup, resistance, capacitance, and electrolytic methods 124 A study of these instruments undertaken by Smith and Freeý5, showed that the use of most of the instruments is constrained by the difficulty of calibrating them, in addition to other problems. such as measurement range and time of response. They recuire gaseous p hcse standards for calibration, and these in themselves are difficult to produce and standardize. Free49 proposed an alternative gas-liquid chromatography method, using a carbowax 1500 (polyethylene glycol) column on a teflon support, a katharometer detector and hydrogen as the carrier gas. The instrument is readily calibrated by the injection of aqueous solutions of organic compounds, such as methanol and ethanol, concentrations of which can be readily determined. The instrument was however found suitable for measurement of very low water vapour concentrations and unsuitable for high humidity conditions, such as envisaged in the present study. Several modern electronic digital instruments which can be appropriately used for this purpose are now commercially available, but owing to cost constraints they could not be acquired for use in the study. It was therefore decided that most appropriately a wet bulb thermometer be used since it is cheap, does not require to be calibrated and if carefully used can give fairly accurate readings. The use of a wet-bulb thermometer for measuring humidities also requires a measurement of dry bulb temperature and combining these with an appropriate psychometric chart. To ensure that the wet and dry bulb temperature readings are directly comparable the two thermometers must be tested by putting them in a beaker of cold water and taking their readings simultaneously at regular time intervals, while the water is heated . If the readings are plotted, one on the absicca and the other on the ordinate, the pair that gives readings that lie on a straight line as close to the 4-5° line as : ossible is deemed to be most sui ;, able . 155 Fig. 7.3 : Wet/Dry Bulb Thermometer for Humidity Measurement. Dry Bulb Thermome to Wet Bulb Thermometer '1astic Tube. 'ater Cavity rubber Ring. Rubber Bung. rick 156 The instrument was made up by drilling two holes ir_ a rubber bung through which the two thermometers were passed. - short rigs :c tube was put around the stem of one of the thermometers, direct!;,, above the bung, to create a cavity in w. -. ich a water column can be held. The cavity was sealed at the base of the tube using a rubber ring. r_ thick, porous wick sheath was then slipped around the bulb o`_' the thermometer. Two such instruments were made up and used to measure the humidity of the air at the inlet and the outlet, by inserting the bung holding the instruments in holes drilled along the arrrorriate conduit lines. 7.3.4.. Procedure For each experimental run, the heater is switched on to give the appropriate level of heating desired. The blower is turned on and the flow of air is adjusted to the required level. Steam is then irrmeýiately let in by turning and adjusting the steam valves. It is al; ays ensured that the flow of air precedes the injection of steam in order to avoid the electrical hazards that may be associated with a backfrard drift of steam into the heater when there is no air flow. The folloving readings are taken at time intervals as the run proceeds: - air inlet and outlet temperatures, - air inlet and outlet wet-bulb temperatures, - bed temperatures at different heights. From time to time the air flowrate is read to ensure that a steady flow is maintained. Given the general practical difficulty of direct measurement of steam flow, the humidity of the inlet air is set by a careful and intuitive manipulation of the steam valves in co-ordination with the readings of the wet-bulb thermometer as well as the indication of the pressure gauge. Under conditions of high air flow rate, low humidity and high temperature, the removal of water from the wick proceeds very rapidly, exceeding the rate at which it is replaced by capillary floe: from the water cavity above. It becomes necessary therefore to remove the instrument from position, wetting the v. - therefore and reinsert--n¬ the instrument before a reading is taken. I57 At the end of each adsorptive run the adsorbent is regenerated by a flow of dry hot air. The procedure for the regenerative runs is similar to that of the adsorptive runs except that there is no steam injection to increase air humidity in this case. Air at ambient humidity is used. 7.3.5" Materials Granular type silica-gel adsorbent was used for the study. The properties of the adsorbent, as furnished by the suppliers are: Average Particle Diameter : Surface Area Pore Volume Bulk Density Specific Heat 7. L.. Results 3.5 mm. 0.85 m2/kg. 0.35 x 10 3m 3/kg. 1200 kg/m3 1.05 KJ/kg -0C. The series of experimental runs was carried out over the following range of operating conditions: air flowrate air inlet temperature air inlet humidity . 0.5 - 2.0 kgs/min. : 30 - 105°C. : 0.005 - 0.05 kg, /kg. bed height : 0.6 - 1.0 m. Typical temperature and humidity profiles for a run are presented in fig. 7.4. Fig. 7.4a. shows bed temperature at different times; Fig. 7.1+b is a plot of air inlet and outlet temperatures against time, as well as a plot of inlet and outlet humidities versus time. These observations confirm that the adsorj+4n front controls the breakpoint and is evidently preceded by the heat transfer front. These are also fairly typical of most of the runs. The double breakthrough suggested by Leavitt96 was not observed. This indicates therefore that the double 1,: 'ý phenomenom does not occur in the adsorption of water vapour on sil-yca- ý; el, at least in thi ai ; Far-tus. 158 I Staýýrýturo (°c) 7 f i Fip.. 7.4a : Adsorption Bed Temperature Profilaa (Run 37) Timo A (tins) it inlet it inlet it inlet 159 00 Fig. 7.41p : ndsorD Lion Air Temperature and Humidity Profiles (Run 37) 90 Te 80. 70 60' emp ra- ture T (00 -50 4.0 , 30 20 10 0 . 028 . 021 . 020 humidity (kg%kg) . o16 . C12 . 008 . 004 J0 7c 0 10 20 30 4-0 j0 6U 14 wins 159 IOOw Fir -. b: Adsorption Air Temperature and Humidity Profile$ (Fun 37) 90 70 ?t ra- '. uro I. Cj"50 40 30 20 10 0 0 1o T . 0iß . o24 ýý_ a: eý . ý' '' =ý: ýý ý, ý °' ýý s, ý _ý ýý.,. ý. ý- ýý ý ý ýý . 020 humidity Ö7G . 012 . 008 . 004 0 } 70 20 30 40 50 by Time 0 (gains) 160 The dehumidifying effect is illustrated in fig. 7.4. b. by ty which is the difference between the humidities of. the inlet and outlet air streams. The heat pumping effect is also demonstrated by the temperature increase of the air AT, as shown in fig. 7.4b. Fig 7.4a. shows clearly that the temperature in the bed varies with distance from the air inlet at any time. Evidently the layers close to the inlet attain a maximum temperature faster and fall off earlier. The temperature of the exit air at any time lies between the maximum and minimum temperatures in the bed at that time. The obtained results are tabulated in Table 7.1, where G T. i T e yi and y e is is is is is is bre i air air air air air 3kthrough time flowrate inlet temperature outlet temperature inlet humidity outlet humidity. 161 TkBLE 7.1 : EXP~RIA NU RESULTS RUN IAGT. Te yi ye No. ( bns) (kg/min) (oC) (°c) (kg/kg) (kg/kg) 01 90 1.25 70 71 . 0065 . 0005 03 55 1.25 35 75 . 0140 . 0010 05 105 2.00 55 58 . 0060 . 0005 07 195 0.65 78 68 . 0065 . 0005 09 410 0.65 37 4.3 . 0070 . 0005 11 135 2.00 34 4.7 . oo6o . 0005 13 50 2.00 7tß. 71 . 0060 . 0005 15 225 1.25 35 4.3 . 0060 . 0005 17 14.1.25 74.119 . 028 . 0030 19 75 1.25 30 68 . 0160 . 0010 21 260 1.65 30 35 . 0050 . 0005 23 60 1.65 80 75 . 0060 . 0005 25 4.5 2.00 87 76 . 0065 . 0005 27 110 1.25 69 68 . 0055 . 0005 29 75 1.25 92 77 . 0060 . 0005 31 4.5 2.00 32 70 0.016o . 0010 33 25 2.00 65 92 . 0240 . 0015 35 30 2.00 32 80 . 0230 . 0010 37 25 2.00 59 85 . 0220 . 0015 39 82,00 55 112 . 0210 . 0030 4.1 37 2.00 60 1 78 . 0180 . 0015 1+3 21 2.00 61 95 . 0350 . 0020 45 35 2.00 32 75 . 0200 . 0010 4.7 20 2.00 72 95 . 0350 . 0025 49 24.2.00 72 91 . 0260 . 0015 51 27 2.00 71 89 . 023 . 0015 53 25 1.25 104 97 . 0360 . 0025 0015 55 33 1 . 25 88 95 . 0210 57 24.1.25 85 101 . 0350 . 0025 59 20 1.25 85 101 . 0350 . 0025 61 140 1.25 69 67 . 0060 . 0005 63 70 1.25 72 1 75 . 0060 . 0005 65 24 .o 60 82 . of 6o . 0015 67 -50 2.0 62 72 . 0160 . 0010 69 16 2.0 75 101 . 0+20 . 0025 71 40 03 62 77 . 0130 . 0010 73 ý3 2.0 83 . 0,1 50 . 0. ' 0 162 7.5. Calculation Procedure If the average temperature in the bed and of the exit air is Te, then the temperature rise across the column up to breakpoint will be given by: T= Te ' Ti ý7.1 Humidity drop across the column will be: Ay= y1 - ye (7.2) The amount of moisture extracted, up to breakpoint will then be: Mb =G Ab Ay (7.3) and hence the adsorptive capacity of the bed at breakpoint will be wb = Mb (7-4) M s where M is mass of solid adsorbent held in the bed. s The partial pressure of moisture in inlet air can be obtained from the humidity, thus: _ Pi 29 11 Yi - (7.5) 18 + 29 yi where 11 is the total pressure. Saturation vapour pressure of moisture evaluated at average bed temperature can be expressed by the Antoine equation: P= exp ( 18.1 (7.6 Ob Te +230 Hence, from its definition bed relative humidity at breakpoint X will be: Pi b Pob (7.7) ýsample calculation, starting from raw experimental data, is presented in the Appendix and the calculate- results are shown in Table 7.2. X63 TABLE 7.2 : CALCULATED RESULTS RUN AyT Wb %b No. (kg/kg) (oC) 01 . 006 0 1 i 2.88 3.21 03 . 0130 40 3.82 5.77 05 . 0055 3 4.. 94. 5.31 07 . 0060 -10 3.25 3.65 09 . 0065 6 7.40 12.96 1 11 . 0055 13 6.35 9.06 1 13 ý . 0055 3 2.35 2.96 15 . 0055 8 6.61 11-13 17 . 0250 4.5 1.87 2.50 19 . 0150 38 6.01 8.86 21 . 0045 5 8.25 11+. 27 23 . 0055 5 2.33 2.50 25 . 0060 ! -11 2.31 2.60 27 . 0055 -1 2.94. 3.10 29 . 0055 -15 2.06 2.30 31 . 0150 38 5.77 8.12 33 . 0025 27 4.81 4.97 35 . 0220 48 5.64 7.61 37 1 . 0205 26 4-38 5.98 39 . 0180 57 1.23 2.17 41 . 0165 18 5.22 6.55 4.3 . 0330 34. 5.92 6.38 4.5 . 0190 43 5.68 8.16 47 . 0325 23 5.56 6.38 4-9 . 024-5 19 5.03 5.57 51 . 0215 18 4+. 96 5.34. 53 . 0335 -7 4.48 6.09 55 . 0195 7 3.44 3.91 57 . 0325 16 4.17 7.14 59 . 01+15 19 1+. 4.3 5.73 61 . 0055 -6 3.20 1.22 63 . 0055 3 2.57 2.50 65 . 0145 22 3.72 4.94. 67 , 0455 10 4+. 81 7.4.6 69 . 0395 26 5. L. 0 6. '0 71 . 01-, -, 15 . Al 0 l4. rL 9' L. 19 161. 7.6. Effects of Variables 7.6.1. Flowrate Several runs were carried out at an average air inlet humidity of 0.0060 kg/kg and at inlet temperature levels of about 35°C. and 0 75 C, while the flowrate of 'air was varied between 0.65 and 2.0 kgs/ min. The experimental observations showed that the breakpoint time and to a lesser extent the bed capacity were influenced by the air flowrate at both temperature levels. The slight fall in bed capacity as flow-rate is increased as illustrated in fig. 7.5. These data are fairly consistent with those of several earlier investigators 64,67,68, which indicated a measure of dependence of capacity on flowrate. They do not support the suggestions of Miller and Roberts 1029 and of Moss and McLaughlin 101 , that capacity is unaffected by flowrate. Since most earlier observations were made under low temperature conditions, it is of particular interest to note that the effect of flowrate on capacity at high temperature is similar to that for the lower temperature situation. The slight drop in bed capacity as flowrate is increased can be explained in terms of the reduced contact time between the adsorbent granules and the air, associated with a high flow. With reduced contact time, moisture is less able to permeate the pores of the adsorbent particles. The effect. of flowrate on bed relative humidity is similar to that on bed capacity and is also shown in fig. 7.5. 1 65 FiE 7,5 . Effect of f1o, r t; or. k- t. Cap city rr, ýeý ýur. 20_ 20 18 16 14 12 Bkpt. cap. Vt b (%ý IC E 4 L 0 8 ý6 4 12 Bed rel. humidity Xb 10 (i B 6 14 2 I0 T 2.0 25 .5'. o 15 flov, ra to G (kgs/mir. } 66 7.6.2. Temperature The adverse effect of temperature on the process of a sore tics is well recognised. However, little previous work has been done on the nature of this effect, especially with regard to inlet temperature of moist air. Experiments were conducted at atmospheric humidity and at two air flowrates, with inlet temperatures varying be t:: een 30 and 0 95 C. The predictable fall of adsorptive capacity and bed relative humidity as inlet temperature is increased is illustrated in fig. 7.6. A more interesting observation is the effective drop in ten era ;, ure of the air as it passes through the adsorption column (i. e. ý. T <0 under high temperature conditions. This phenomenon can be explained by considering the heat transfer processes that take place. If the released heat of adsorption is insufficient to raise the temperature of the solids to the inlet temperature of the air, then there could be a net transfer of heat from the gas phase to the solid phase, resulting in a decrease in temperature of the air. Obviously, high temperature conditions would be expected to be accompanied by shorter breakthrough times. This is even more significant if moisture is adsorbed on a hot bed from a hot air stream. Consequently, under such conditions, very loN breakpoint capacities are obtained. However, following the discussion of the nature of heat transfer taking place, high temperature increases in the air would be expected. This is best illustrated by considering Run 39. Though most of the runs were carried out with initial bed temperatures of 20- 30 o C, this was performed with an initial bed temperature of 105 ° C. was obtained, with a high temperature rise LT bed capacity of 1.2311- of 57°C, as compared with typical values of bed capacity of 3-81/-c and temperature rise less than 30°C, for lower bed temperatures. _ aw o ý«o : CD O N O 0 H 0 at Z JJJ 0N -P' N1OX "j x X 21 11 u vi m v r ýy o N' :r III o rr rs y pi. o 168 20,, FIG. 7.7 : Effect of inlet temperature on temperature chLni e 16 12 1 $k 4 Temperatur change 0 20 " In 14. 8 12 16 x1 . 25 kgs/r. L: o 2.00 kgs/can. 20 169 7.6.3. Humid it_y The manner in which the humidity of inlet air affects the break- point capacity is twofold. First, an increase in air humidity res',: Its in increased vapour pressure and hence higher adsorption capacity. However, since adsorption is intrinsically exothermic, the associated temperature increase tends to have an opposite effect. The dependence of capacity and bed relative humidity on air humidity is presented in fig. 7.8. At a low air inlet tear erature (350C), the adverse effect of temperature seems to be dominant and hence capacity decreases slightly as humidity is increased. This appears to be consistent with the observations of Grayson 68 and -. '; orthington67. On the other hand, at higher inlet temperatures, it was observed that the breakpoint capacity increased with humidity up to about 0.02 kg/kg and subsequently levelled off. This also agrees with the earlier suggestions of Grayson 8, that capacity may also increase with humidity under certain conditions, and of Miller and Roberts102, that capacity may be independent of absolute humidity, for adiabatic adsorption. The relationship between the temperature and humidity changes of the air across the column is plotted in fig. 7.10. The plots lend support to the general concept of a linear relationship between 6T and A y, up to aAy value of about 0.01 kg/kg. Such linear relationships have been proposed by Rathmell and Bateman 1o6 and by ':; orthington67' as noted jr. chapter 5. However, their linear equations do not take into account the possible effects of inlet temperature and flowrate on this relationship, which are clearly obvious from the plots- of fig. 7.10. It is noted also that these plots reinforce the previous observation that negative values of AT may occur under a combination of high temperature and low: humidity conditions. 170 Fig. 7.8 : Effect of inlet hunidity on breakpoint capacity. 10 8 6 Bkpt. cap Wb (%) 4 2 0 ý_ I-, - x 33ä0 C, 2.0 k¬s/nin. 0 62 C, 2.0 kgs/nin. 0 72 0 C) 2.0 kgs/unin v 88 °C, 1.25 kgs/min. . 01 . 02 . 03 . 0J+ Inlet humidity y, i (kg moisture/kg air). . 05 171 Fi .Q Effect of inlet hü:: i^ ty on be, - rel. hum. 10 8 6 Bed rel. hum Xb (0) If 2 0 X 33°C, 2.0 kgs/m! n. o 62°C, 2.0 kgs/min ° 72°C, 2.0 acts/min v 88°C, 1.25 kgs/min. 5 Inlet humidity y. Z (kg moisture/kg air 172 Fig. 7.10 : Temperature chance ix vs humidity change `yc A v 5 40 30 2C Temperature change AT ýCý .IC -I -2C 33°C, 2.0 kgs; 'z-n 2C 6- C, --. O 720C, 2.0 ký s ',. -n 88°C, ' . 2j i- s, / / zir. 73 7.6.4. Bed Height Experimental runs were carried out with bed he ¬his of 0.6, 0.75 and 1.0 m. under fairly similar conditions. Predictably, it was observed that higher breakpoint capacities and bed relative humidities resulted from increased bed heights (fig. 7.1 ). The influence of bed height can be attributed also tc contact time. A larger contact time would be associated v; ith increased bed height, which thus allows the adsorbed moisture to penetrate to the inner pores of the adsorbent particles, enabling adsorption of more moisture. Fig. 7.11 10 8 6' r Bkpt. cap. ,;. " b (%o) 2 0 174 Effect of bed height on bk-t. capacity and bed relative hu-ic, ty T10 x 2.0 kgs/min. 1.25 kgs/min. wb 8 --- Xb -, --0 1 Bed height Z (m) i i i =6 Bed rel. Xb (; =) 2 2 hum. 175 7.7. Correlation of Data The calculated breakpoint capacity is plotted against bed relative humidity for each run in fig. 7.12. These data show some similarity with the data of hiller and Roberts 02 for semi-isothermal adsorption of moisture on activated alumina (fig. 5.14) in that there is an indication of a relationship between breakpoint capacity and bed relative humidity. However, while their data depict a straight line relationship, these data do not. The data of fig. 7.12 are also comparable . ith those of Worthington 67, for non-isothermal adsorption of moisture on silica-del (fig. 5.15). It is noted however that a considerable difference exists between the two sets of data. 7, hile Worthington's plot appears to be convex to the X. b axis, the present data appear to be concave to the same axis. The shape of the plot suggests that the data can be correlated with a polynominal function of the form: n Wb =K Xb (7.8) A log-log plot of YTb against Xb is presented in fig. 7.13. Using the method of least squares, the data was correlated by the straight line equation: Ln Wb = 0.07 + 0.81 to Xb (7.9) hence, 0.81 Wb= 1.07Xb (7.10) Alternatively, the data was correlated by a pair of straight line equations, one suitable in the range 0 to 6 relative humidity and the other for relative humidities above 6ý. Thus: Wb = 0.8 Xb 0 6.0 (7.11b) It is noted however that the investigation was carried out under adiabatic and atmospheric conditions for which bed relative humidity values are low. For higher relative humidities, which may be attained under isothermal or high pressure conditions, equations 7.10 and 7.11 may therefore not be ap Jicable. 7, orthington suCgested that a rnaximur.. bed capacity o`_' azcut 2Cf, is attained t hir''n bed relative huriditiez for hier. recs', ire adsorption of wate: vapour on silica-gel. X76 Fig. 7.12 : Correlation or blot. caracity- "º; b and bed relative hu: --- ,4 ýý" xý I 14. 12ý 101 Bkpt. cap. W b (ý) 8 x x 6ý 2 x x x xx x fx x x` z xx x WA % it x K it x 0 02 1ý 68 10 12 11.1 9 Bed relative hum. Xb C 20 15 10 8 6 Bkpt. cap. Wb º77 Fig. 7.13 : Log-log correlation of bkpt capacity Wb and bed relative humidity X,. Li Bed rel. hum. Xb , 78 7.8. Comparison of Breakpoint Capacity and cuilibriurn Ca-n ac_ty The silica-gel / water vapour adsorption equii i-:, riuT has been discussed in Chapter 5. The concave shape of the plot (fig. 7.12 , and the nature of the corresponding equations (7.10 and 7.11) indicate that there is a resemblance between the breakpoint data ob, aine and the silica-gel / water vapour equilibrium relationship. In this regard therefore, it may be useful to compare the values of break- point capacity predicted from equations 7.10 and 7.11, for both isothermal and adiabatic operations, with equilibrium capacities obtained from some of the proposed equilibrium equations. A silica-gel adsorption bed is shown in fig. 7.14. Humid air is deemed to be passed into the bed from a convective dryer as shown and conditions are as indicated in the diagram. The calculation procedure consists of: () Dryer design calculations to obtain the outlet temperature and humidity conditions T2 and y2 and hence the inlet conditions of air to the adsorber. (2) Calculation of equilibrium relative humidity Xe, evaluated at inlet temperature and humidity, T3 and y3, to the adsorption bed. (3) Calculation of equilibrium bed capacity 'A'e using the following equations: (a) Simpson-Cummings 64 Equations - Toe = 1.57 + 0.58 Xe (0 < Xe < 50) (7.12a) We = 24.63 + 0.14 Xe (50 < Xe < 100) (7.12b) (b) Juty - Edwards 60 Correlation - If Xf is defined by Xf = Xe exp (1.8 T3/500) (7.13) where T is in 0 K, 179 Fig. 7.11 : Scheme for Evaluating Silica-Gel EcuilibriU and Breakpoint Capacities. Adsorption column Y4 ISO then, 2X 3.3 f 10.15X f( + 7.14' e1+ 13-33X f1+ 46.7Xf 3.3 (c) Mukesh-Rae equation - 0.781. 0.92Xf 0.92Xf (7.15) We -I +1.603Xf + -. 603 (4) Calculation of breakpoint bed capacity Wbi, for a bed operated isothermally, say by insertion of cooling coils in the bed, from equation 7.10 and 7.11, but using the equilibrium bed relative humidity Xe, rather than the breakpoint humidity Xb. This is necessita;, ed by the fact that the bed temperature and the air outlet temperature would remain at about air inlet temperature for isothermal operation. Thus: . 81 Wb .=1.07X e (7-16) or VI b. = 0.83Xe (0 6) (7.17b) (5) Calculation of air outlet temperature from the adsorption bed T4 and hence the breakpoint bed relative humidity Xb and breakpoint capacity Wb , for non-isothermal operation, using the bed heat a balance equation 6.50 and either equation 7.10 or 7.11 in combination. Using the procedure outlined above, a computer programme was set up and the results of the computation of the various capacities and relative humidities over a range of dryer opera ting conditions, and hence inlet conditions to adsorber, are presented in Table 7.3. { d d m cD cD f+" Fý" %. N N -' N - 1 1 1 y r 0- d Z P. - C N ( D Co C 0 1 ý 0 0 0 P 1 $ la E3 P) n m (D c c cf cf P. H to O O (D (D (D O " . F 0 t--i H H N" P" P. d d d F+" 1ý" >S '1 11 m Co H. H. p- 0 O G I G rf, : zr 5 (D cD o o c) bli d 'd >O 'I PP R) D lb I-' FJ n 0 0 w w a rj w sv v' d ) c+ c* A C) d d ý c 3 w zý x b O O N" N" A n w a a a" Q C) N" H" & c+ cD cD u C, 181 U IN N f`, N - ýD 0 ýln Oo - 0 -J - ýJ Ui N" == w lr+ . O\ CO IN Oý -ý Q1 o }< CD C. Z: CD O' '. O -' Q\ C'% 0 V7 Vi OD J N) C Fý" Ga (D F-J J C"~ V` W I-N "'"1 " L:. " N 1l 0 O' ý1 In Qn tz7 < ýp O\ N ' _ Q CD "ý g IN co N \171 -i N) - P P. G N O co tn c+ ?C L N) N) -' - 5 --' V' N) CID IN \. w -' CO 01 U' " o e, (D - N) "P N) -' 0'*, N) ýJ Ui vi 1J tu N" cY W N N) -' -' -a -` N) U' IN ON -, N) O CO U -F- \, m 1-1 F'- . " . " . . " . " " . o N 01\ N Q". \1.0 - o\ Ol\ 0 n 0 ef lTi W IN N -ý N -ý -ý r O a\ -' c r\ CD - o-\ J Un '. J " a N O -' -' U' -' N Oý W CO IN W ýJ N O Ui co OD O\ 4 W - Oý N N -` -` -s -"ý " O'N N) N) Q -J 0 VI ý N 0' lý .. M -1 . ý1 U' U' N N N" 0 0 1J -' H W ED o CD C+ P) ý31 PT, CD ö v' CO o\ w o o '. o v\ or\ 'n -p C+ " Q Q - -' UJ1 ýl P N Z. N IN U' N" N) 0 -4 ON CO N IN W N) C) P) O N" J O o, \ co v, Ln ýJ lT U' - U+ ý ä o e cc) C) W `i CC) - -, j -ý -ý r v 0 1 C+ P- C) c m ý! ýn Ln Un Un . J C)o A. - W -J O OD -! Co W OD ýJ N) H s r r 0 C er C Z 0 F2J C r U1 C r H C7 a n H H H C=] En 182 For each set of entries in Table 7.3 Xe is seen to be larger than Xb. This arises because of the increase in air ýeý, _ e= pure as it leaves the adsorber. However, smaller Xe ;. han Xb could be obtained, if conditions are such that there is a fall in air temperature across the bed. Among the three equilibrium equations consiiered, the Mu esh Rae equation evidently predicts a much higher equal ibrium capacit`- than the other two, which agree to a reasonable extert. A. closer examination however indicates that Jury-Edwards correlation would generally predict slightly lower values than Sir: psor-Cummings equation. For both isothermal and adiabatic adsorption, the results show that the two empirical equations obtained for breakpoint capacity, equations 7.10 and 7.11, agree fairly reasonably, especially for low and intermediate capacities. Overall, eauatior. 7.11 tends to give slightly lower predictions up to bed capacities of about 6; -' and slightly higher predictions for bed capacities greater than 6`,. Overall, predicted equilibrium capacity values are larger than isothermal breakpoint capacities, especially in relation to the use of Mukesh-Rao equilibrium equation, and to a less extent the Simpson-Cummings equation. Theoretically, since a fraction of the adsorbent in the MTZ is unused at breakpoint, it would be expected that breakpoint capacity would be less than the equilibrium capacity. The difference between We and Wb would however be expected to diminish under high temyerature, and hence low bed relative humidity conditions, since the heat and adsorption fronts would tend to travel faster and the breakthrough would be sharper. Evidently therefore, equations 7.10 and 7.11 will predict break- through capacities fairly satisfactorily. Following the relationship between Xe an-- Xb, it is obvious that adiabatic breakpoint capacity will be smiler than both isothermal breakpoint capacity and equilibrium capacity. ov: ev r, if conditions are such that Xe `X b, then the reverse situation will apparently hold. 183 7.9. Regeneration Various flows in the region 1.0 - 1.5 kgs/min. and tempera . gyres of 110-14.0°C were used for the regeneration of the uses bed. 1v ical gas phase and solid phase temperature and humidity profiles for a regeneration run are presented in figs. 7.15 and 7.16. These evidently support the wave theory of sorption. Clearly the profi; es show that desorption at any level in the bed takes place at a fairly steady temperature which subsequently rises when desorption is complete at that level. It is also noted that the desorption temperature at the air inlet end of the bed is nearest to the air inlet wet-bulb temperature and increases as the wave moves further along the bed. The initial temperature of the used bed was observed to exert a considerable influence on the mode of desorption. The adsorption run is usually stopped when the bed is saturated at a high temperature. If the bed is allowed to cool, the equilibrium is shifted and the bed loading is less than saturation. Thus, initially the flow of hot, relatively dry air may result in adsorption. Subsequently, the temperature of the bed rises rapidly. The equilibrium is then re- established and desorption then commences. If regeneration is commenced while the bed is hot, there is no initial period of adsorption since the equilibrium favours desorption from the start. Fig 7.15 represents regeneration of a hot bed while fig. 7.16 represents the regeneration of a cold bed. O or N O O O*N 0 0-1*-3 co N" 8 . ýO O O N 0 J 0 crl\ 0 CA O O O 1b4 cD o acs c' lTi ýS dý ýn r. y CD m i ý: ýý I ýa 1ý ;C Ic C' C_ C º. (~ N" .. J CD I c+ Iý. ö CD 1 i Iý N ^5 Ob c7 Ho ri PD ýn Oý ýJ Wmp 0000O00O 0 NN ÖC N" a iT ýý O rv ý' O . 0 4- O 1ßb ý.. ý Pý 9 s uH 4 1ý 8 I 0 L .. "" ". NyN s"" V1 ýJ1 V1 ti N to is t3 tql tr fq' w P.; f9 "g ä. % ?ýM 4- 0 5 .. f4 iý o 0-3 C'J A a p0 0 Q` 0 44 W w 0 r- 1, , cl* CHAPTER EIGHT PREDICTION OF PERFORMANCE 188 8. PR:: DICTIOKN- OF PERFORMANCE 8.1. Conditions of Air from Conventional Dryer Before proceeding to evaluate the performance of the AHP dryer using the various models outlined, in conjunction with the results of the experimental investigation, it is important that the nature and state of moist air delivered from a dryer, under various operating conditions, are clearly stated. By considering a convective dryer operated with air inlet temperature TDi in the range 50-200°C and fractional saturation f of 0.2 - 0.8, the outlet temperature and humidity from the dryer (TDe and YDe respectively) and hence delivery conditions to an AEP, T. i and YAi have been calculated. Ambient conditions are assumed to be 100C and 0.005 kg/kg humidity. The data are presented in Table 8.1 and in fig. 8.1. Clearly at constant dryer air inlet temperature as fractional saturation is increased, outlet temperature TD e falls while outlet humidity YDe increases. Hence, the more moisture laiden the air, the lower its temperature. On the other hand, operating at constant fractional saturation, TDe and YDe increase simultaneously with the inlet temperature since a higher inlet temperature would naturally imply a higher outlet temperature and the moisture holding capacity of air increases with temperature. These relationships are clearly illustrated in fig. 8.1. 189 ; ABLE 8.1 : :d CO DITO S FROM CONVEI IOIt L DRYER T f e T YDe D. 1 fractional o ( ýý (°c) saturation) (kg/kg) 50 .2 44.39 . 0074 .4 38.83 . 0097 .6 33.31 . 0121 .8 27.81+ . 01144 100 .2 86.26 . 0107 .,. 72.79 . 0161+ .6 59.57 . 0220 .8 46.59 . 0277 150 .2 126.91 . 0144 .4 104.53 . 0237 .6 82.83 . 0331 .8 61.77 . 0424 200 .2 166.76 . 0182 .4 134.93 . 0314 .6 104.4+3 . 0246 .8 75.18 . 0578 Fig. 8.1 : t-. ir Con' Conditions from Convective Dryer. 18Q 16 I 1 Exit air temp. T 10 De (0c) 60 40 20 oý 0 I/ ýýýi ý, \ ý4 = 50°C Di . 01 . 02 . 03 . 01+ Exit air humidity Y De (kg Hc0/kg air) TDi is d r; -er air inlet ter; ereture . f is air fractional s turation. . 05 i 00 °c ý. ý .v 191 8.2. Lauilibrium Model. The Equilibrium Model, which has been previously outýýreý (Sec ion 6,3), was used to predict the performance of the rP dryer, oaerating in the TSC mode (Fig. 6.1). The results are plotted in Figs. 8.2 and 8.3. 8.2.1. Drying conditions It is observed that the drying conditions, air inlet temperature TDi and fractional saturation f, have virtually no effect on the thermal efficiency of the TSC dryer. Operating at a higher drying temperature leads to a high outlet temperature TDe and consequently high influent temperature to the adsorber TAe, at any fractional saturation. rlso this gives a higher dryer outlet humidity YDe and hence a lower rec`ýcie rD. Erom equation 6.29, r1 _ TDi TDe (6.29) EM TDi TAe) + (1-rD TRi-To) it is evident that increased TDi and TAe would improve the cycle thermal efficiency r'EM while increased TDe and reduced rD would tend to lower it. Thus, the net effect is that TiEM is almost independent of TDi* The reduction in TDe and increase in YDe associated with an increase in f, for a given temperature, have opposing and counterbalancing effects on TAe. The lower TDe and the attendant reduction in rD consequently keep n EM constant. 8.2.2. Regeneration Temperature Examination of equation 6.29 suggests that lo-.,: er regeneration temperatures TRi will give higher efficiencies. However, it is noted that a low TRi implies also a low exhaust humidity Re and hence increased demand for regeneration air. Consequently, hot air recirculation to the dryer rD falls and falls too. This effect 0 is illustrated by the plots for regeneration at 150C and 200 C in fig. 8.3. AP Fig. 8.2 `-hermal Efficierov of S'- -'ever (F. auiiibrium: Yodel 100 70 60 Efficiency n (%) 50 40 30 20 10 90 TDi is dryer air inlet temperature x efficiency of conventional dryer, fl a efficiency of TSC dryer, 7- 80 TDi _ 500 n 20 C 'COIIV 200°C /o°C 110 = 50°C Di // 0! 4 11 .6 0 .4 fractional saturation f. ý. o 193 Fig. 8. . Effects of Regener- ;, ior. r , -a exhaust condition on _herE. a Efficiency of TSC Dryer E. uilibriua. Yodel) 106 EM is efficiency of '30 dryer ccnv is efficiency of conventional dryer 90 Ri is regeneration temperature SRe is exhaust sa . ura t or. 80 t 70 60 50 Efficiency 11 40 30 20 10 0 0 C SRe - 1 OCo 0C, SRe = 8C 0 C, S. C, :E 6 = 0. ý . 1+ .6 fr, ctional 19L1 8.2.3. Exhaust Conditions The closer to saturation the air is when finally exhausted, he higher the cycle efficiency, as shown in the plots for 6Cf., 80; - and 10O approaches to saturation in fig. 8.3. Though this is not directly reflected in equation 6.29, it arises indirectly through the increased recirculation of hot air wich accompanies exhaustion at higher saturation levels. 8.2.4.. Comparison with Conventional Dr-,, er. It may be instructive to relate the performance o4l' the TSC dryer to that of a comparable non-recycle conventional dryer which is considered to operate with the same temperature and fractional saturation. The results indicate that considerable advantage may be derived by TSC drying, especially if the air is exhausted at a sufficiently high level of saturation. This advantage is even more distinct when fractional saturation is low since the thermal efficiency of the comparable conventional dryer n is strongly dependent on f cony (see fig. 8.2). Hence, it may be inferred that ordinarily, operating a dryer with as high a fractional saturation as possible provides the simplest way to improve the thermal efficiency. However, if for other constraints, such as effect on drying rates, it is desirable to operate with low fractional saturation, considerable scope may exist for improving the thermal performance of the dryer by combining it with an AHP operating on the TSC mode. 195 8.2.5. Pressure Swing Cycle (PSC) Similar computations have been done for a PSO dryer (fig. t. 2) and the results are presented in figs. 8.4. an4- 8.5. The enerýý- consumption in the compressor Ecomp was obtained in the -,. ay previous-ly outlined in section 4.3, while the energy- recovery in the cooler E rec was calculated by heat balance on the cooler. Vrnereas the efficiency of the TSC dryer is virtually independent of the TDi and f. efficiency of the PSC dryer is found to increase with both parameters as shown in fig. 8.! . As the compression ratio is increased, the energy recuirenent for compression increases. Tough the heat available for recovery at the condenser increases simultaneously, this is insufficient to offset the increased energy requirement. This is because the efficiency of converting primary fuels to electricity is low, which makes E co mp very large. The net effect is a decrease in the efficiency of the PSC dryer at high compression ratios. With the predominance of compression energy in the total energy demand, it is also observed that lo. vering the cooler temperature TAi, thereby recovering more heat at the cooler, has only a limited effect on cycle efficiency. 8.2.6. Comparison of TSC and PSC Like the TSC dryer, the highest cycle efficiency of the PSC dryer would be expected when air is exhausted at the highest possible saturation level. However, in relation to TSC dzr, -in¬ under identical conditions, PSC drying would clearly provide a less favourable thermal efficiency. While efficiencies of up to 6Cf, ' may be expected using a TSC, the PSC dryer may not attain efficiencies greater than 25;. Similarly, relative to comparable conventional ýrjing, PSC ý: --r. "ýng offers a poor thermal performance, as is shown in figs. 8.4. and 6.5, over the range of conditions considered. The indicator. , -:. ref ore is that the PSC will not be suitable for enhancing the thermal performance of dryers. 196 Fig. 8.4 : Thermal Efficiency of PSC 'ý r (E _ 10Jk L'ý"je ? uilibr_ h: el) Pressure 90 TDi is dryer o efficiency 80.. x efficiency 70 60 5C efficiency n ýf: ý 4C 30 2C C 0- 0 ýr =JCrC . 4. .6 .6 ractiona_ s urn tc f' 6 atm air inlet te,:. _ era Lure. of PSC -e of convey. tioral dry; e r, TI cc :v. 200°C /50 4oC conv /T Di- 50°C Y0 Ti - 197 Fig. 8.5 Effect of Pressure or -f ^, e- ^ of 1 00 T (Eouilibriun: Model) 90 Dryer air inlet teEper_ .: -ýurýp .. ýDi=' C. 80.1 T, is pressure o efficiency of PSC dryer, r, Jh, X efficiency of convective dryer, 70 60 50 efficiency n 40 30 20 10 0 V. v-- fractional X98 8.3. Evaluation of Adsorption Models Consideration of the Equilibrium Model for ? 3C and P3C dr-rers is only useful as an indication of the kind of thermal efficiency that could be expected and for providing a simple basis for comparing the performances of the two cycles. To make a more realistic prediction it is necessary to perform calculations based on the other models outlined in Chapter 6 (Lee-Cummings Procedure and Dehler Method) which take into account the dynamic effects of the solid adsorbent ignored in the Equilibrium Model. Before proceeding to consider the overall system performance it is considered necessary that the Lee-Cummings and Dehler Models be applied to the adsorption bed, under identical conditions. This would enable the models to be compared on a common basis and to examine how they are affected by changes in operating conditions. The computation is based on the scheme illustrated in fig. 8.6. Moist air is passed into the adsorption bed from the dryer. The calculation steps are outlined below: (1) Dryer design calculation, to obtain the influent temperature and humidity conditions to the adsorber, Thi and Y Ai. (2) Calculation of gas and solid phase temperature and humidity changes in the adsorption bed using the Lee-Cummings Procedure. (3) Similar application of the Dehler Method to the adsorption colurn. 8.3.1. Lee-Cummings Procedure The data obtained from theoretical prediction of non-isothermal adsorption by the Lee-Cummings Procedure are presented in Table 8 2. Essentially, in consideration of an adsorption bed for heat pumping, the most important parameter is the increase in temperature of air as it passes through the bed AT, for any given inlet temperature and humidity conditions, since this reflects the temperature lif wing capacity of the system. In addition the breakpoint c&_r, acit., y of the adsorbert is also very important and enables the size of the bed reOL- - '-c be evaluated. A99 Fig 8.6 Scheme for 'valuation of ndsor`. 4 L: ode1s Tri. `ýAi Ac sorter ±e, yye TDi, Di 200 TABLE 8.2 : PREDICTIO:; OF ADSORPTIOT BY TDi (°C) f T JILV, (°c) wb (%) () T (°C) y k _/k 50 .2 52.66 2.38 8.41 8.27 . 0054 . 4. 52.58 3.05 11.10 13.75 "0077 .6 52.77 3.72 13.63 19.46 . 0101 .8 53.75 4.97 15.48 25.91 . 0124 100 .2 55.11 0.44 10.63 -30.85 . 0087 .4 69.98 1.56 8.34 -2.81 . 0144 .6 80.25 3.54 7.25 20.68 . 0202 .8 86.45 5.93 7.08 39.86 . 0257 150 .6 74.17 1.68 13.75 - 8.66 . 0311 .8 103.62 5.21 5.63 41.85 . 0404 200 .8 105.20 4.05 7.10 30.02 . 0558 201 Fig. 8.7. shows the prediction of AT over a range of drying conditions. Clearly AT improves as f is increased, with a stronger dependence at higher temperature levels. Therefore at relatively low values of f, low temperature operation would give a larger GT but at higher values of f the situation would tend to be reversed. The increase in AT with f obviously stems from the attendant increase in humidity as fractional saturation becomes larger and hence a larger heat release on adsorption. From fig. 8.8, the relationship between breakpoint capacity '. '. b and fractional saturation is similar to that shown between . iT and f. Thus, Wb would be expected to increase with saturation but to show more sensitivity at higher temperature levels than at lower temperatures. From its definition the breakpoint bed relative humidity Xb can be expressed as: _ PA. p A (8.1) where PA i is pressure of moisture in inlet air and Po is saturation vapour pressure of moisture at outlet temperature. Now, PAi is a direct function of inlet humidity YAi while PÄe is a direct function of outlet temperature TAe" By operating with a higher fractional saturation it is therefore possible to either increase or decrease X. From figs. 8.7 and 8.8, when the drying temperature is low (TDi = 50 C) both 1T and 'b increase gradually with f. Consequently, the humidity effect on Xb dominates and Xb increases with f. Hence Wb and the corresponding Xb increase in the same direction. However with higher drying temperatures (100oC and 150oC) dT rises more rapidly with f thanW b does. Xb becomes more dependent on the temperature effect and thus decreases with f. The net result is that Xb decreases while i; 'b is increasing. These effects are clearly illustrated in fig. 8.9. 2 02 Fig. 8.7 : Temperature lift vs fraction saturation (Lee Cumr.. ings Procedure) / T Di , is dryer air inlet temperature. / // 40 30 20 10 temp. lift AT °c) C -1c -20-- -30-- -40 203 Fig. 8.8 : Bkpt. capacity vs fractional Lee-Cummings Procedure 12 10 8 bkpt. cap. W b (%) 6 14. 2 0 0 fractional saturation f. 12 10 bkpt. cap. Wb 6 C ý) 204 Fig. 8.9 : Bkpt. capacity vs bed rel. hum. (Lee-Cummings Procedure) T bed rel. hum. Xb (°) 2 05 Fig. 8.10 Temp. lift vs humid4 c anpe (Lee Churs --n - Procedure 40 3Q 1 20 10 temp. lift AT 0 (o C) -10 -20 -30 -40 -50 TDiis dryer air inlet temperature. -60 r L 06 The variation of temperature change G. l 'n amount of mo_s : ure removed in the adsorber Ly is presented in fig. 6.1C. 6 evicen. iy rises very rapidly with Ay o, -ing to increased heat release as more moisture is adsorbed. It is noteworthy also that values of -1-T. less than zero (0T< 0) are predicted for concitions of high tempera Lure and low humidity (see fig. 8.7 also). The data presented in table 8.2 and figs. 8.7 - 8.1C were obtained by solving the Lee-Cummings Model for initial bed te.: erature Tsi = 30°C and superficial air velocity V= 1250 i /hr. To illustrate the effect of initial bed temperature further calculations have been done for Tsi = 60°C and the results are shown in Table 8.3. The Lee- Cummings procedure seeks to correlate breakpoint capacity for adiabatic adsorption through the amount of moisture adsorbed (i. e. humidity) and does not take into direct consideration the air outlet temperature and the bed initial and final temperatures. Consequently the values of l'; b predicted by this method are completely independent of T. si However since AT is obtained from the heat balance relation T_IVb(AYXA+ cp TAl) +'Lye ps Tsl Ae fir (6.50) EY p+Mb cp g it will show sensitivity to T si If the bed is initially hot, less heat is demanded to raise the temperature of the solids and hence tT increases. This is consequently attended by a reduction in the value of bed relative humidity X. b. Predictions for velocities of 750,1250 and 1800 r. /hr. are shown in Table 84. Evidently there is a slight increase in 7ýb vr*th velocity. Also there is a slight rise in AT, and hence a fall in Xb with velocity. 207 TABLE 8.3 : EFFECT OF I1 ITL L BAD TEIFEATURE ON ADSORPTION (Lee-Cummings Procedure) T. =30°C si T. =60°c si pT V1b Xb GT 97b Xb (°c) (%) (/Z) (°c) U) (% ) 8.27 2.38 8.1+1 13.96 2.38 6.41 13.75 3.05 11.10 20.00 3.05 8.25 19.46 3.72 13.63 26.05 3.72 9.97 25.91 4.97 15.1+8 32.10 4.97 11.55 -30.85 0.41+ 10.63 -10.70 0.44 4+. 34 - 2.81 1.56 8.34. 11.87 1.56 4.56 20.68 3.54 7.25 31.80 3.51+ 4.70 39.68 5.93 7.08 49.18 5.93 4.98 -8.66 1.68 13.75 11.07 1.68 6.32 41.85 5.21 5.63 55.24. . 21 3.59 30.02 4.05 7.10 47.69 4.05 3.; 7 208 TABLE 8.4. : EFFECT 0r VELOCITY 0:: t. DSOPp: Io (Lee Cummings Procedure) v= 750 W/h - v= 250 m/hr. v=1800 DT 111'b Xb DT Yirb Xb T b Xb (°c) (i ( (0C) \i°ý () (°c 7.80 2.11+ 8.60 8.27 2.38 8.1+1 8.56 2.55 8.29 13.09 2.67 11.47 13.75 3.05 11.10 14.16 3.34 10.88 18.58 3.15 14.22 19.46 3.72 13.63 20.00 4.17 13.27 24.96 4.11. 16.20 25.91 4.97 15.4.8 26.49 5.62 15.05 -31.54 0.42 10.99 -30.85 0.4. 10.63 -30.40 0.45 10-41 - 4.28 1.44 8.89 - 2.81 1.56 8.34 -1.87 1.63 8.00 18.61 3.17 7.89 20.68 3.54 7.25 22.00 3.80 E. 88 37.32 5.15 7.81 39.86 5.93 7.08 41.4.7 6.51 6.65 -10.31 1.59 14.73 -8.66 1.68 13.75 -7.59 1 "74 " 3.1 4 38.28 4.68 6.39 41.85 5.21 5.63 44.16 5.59 5.20 26.33 3.72 8.08 30.02 4.05 7.10 32.45 L.. 2E 6.7 209 8.3.2. Dehler Method In outlining the Dehler Method of describing adsorption, it was proposed that temperature rise in the bed will be obtained from Worthington's equation: AT= 70 h1000 (5.34) The predicted values of AT are plotted against fractional saturation f in fig. 8.11, for various dryer temperatures. Essentially AT would increase with f as was the case with the Lee-Cummings Model. However these results differ considerably from those obtained from the Lee-Cummings Model in that over the entire range of fractional saturation larger temperature lifts ( AT) would be obtained by operating at a higher temperature than at a lower temperature. From the nature of equation 5.31, it is obvious also that only positive values of AT would be predicted. Overall, it is observed that much larger values of AT are obtained here than with the Lee-Cummings Model. Furthermore, from equation 5.31 it might have been expected that the plot of temperature rise AT against humidity change ley will be a unique characteristic curve. However, since the humidity difference term 1000 is defined in terms of air volume, it is slightly sensitive to temperature and hence separate curves would be obtained for different temperature levels. (fig. 8.14). 210 TABLE 8.5 : PREDICTION OF LDSORPTIO1, BY Dý'ý E-P TDi f T Ae Wb Xb T y (0c) o ( C) (i) (; -) o ( c) (kg/kg) 50 .2 58.87 5.10 6.26 14.48 . 0051+ . 4. 60.06 7.69 7.79 21.23 . 0077 .6 61.53 12.91 9.02 28.22 . 0101 .8 63.30 16.85 9.91 35-46 . 0124 100 .2 106.98 0.90 1.33 20.72 . 0087 .4 108-40 2.61 1.92 35.61 -C144 .6 111.2lß 5.24. 2.33 51.67 . 0202 .8 115.6o 13.00 2.51 69. o1 . 0257 150 .2 153-41 0.12 0.44 26.50 . 0121 . 4: 153.86 0.81 0.71 49.33 . 0217 .6 157.73 3.03 0.88 7+. 90 . 0311 .8 165.36 8.69 0.91 103.59 . c404 200 .2 198.31+ 0.02 0.19 31.58 . 0162 .4 196.73 0.22 0.33 61.8o . 0294 .6 201.21 1.62 0.42 96.80 . 0426 .8 212.6C 5.92 C-43 137.42 -0/58 CA i 114-0 Fig. 8.11 : Temp. lift vs fractional Defiler ih: e thod /0C°C TDiis dryer air inlet temp. -J I I /ý temp. li DT (0 c) 0 fraction a saturator: f 212 was used to obtain the equilibrium loadings at the inlet and exit conditions. Thus, from equations 6.55 and 6.59, at the inlet, Xe. - PA. i A. i and at the exit, Defining Xf by The Jury-Edwards equation for silica-gel/water vapour equilibrium X= Ae ee p (6.55) (6.59) then, w_ 2X fi ei I+ 13.33 Xfi and Xf= Xe exp (1.8 TA/500) (8.2) 3.3 10.15 Xfi 1 +1.6.7Xfi" 10.15 Xfe3.3 1+ 46.7X, 3.3 (8.3) (8-4) W= ee 2X fe I+ I3.33XFe Wb was then obtained from Dehler's equation (6.61) thus : 1 Wb 2 (wei + wee) (6.61) Over the range of conditions considered the prediction of this model is that larger breakpoint capacities 17b would be obtained at low temperatures or high fractional saturations (fig. 8.12). In relation to the Lee- Cummings Model, the values of Wb obtained here are considerably larger. With the prediction of large temperature rises also small values of bed relative humidity Xb are obtained and the plots of Wb against JCb would therefore tend to be very steep (fig. 8.13). Generally, ? rb would be expected to increase as X. b increases at all temperatures, unlike the case with Lee-Cummings Model. 213 18 16 14 12 bkpt. cap. wb 10 (i) 8 Fig. 8.12 . l, " caoacit , vs frac. `_or ýalu^ý ,. ýýr (Dehler Method). 6 4. 2 C fractional 2 , 4. .6 .8 A1.0 r 1 16 14 12- bkpt. cap Wb (i%) 10 8 6 If 2. 0 214 Fig. 8.1 Bk: t. capacity vs bed rel. huý`d ýý. (Dehler Method) TDi is dryer aiz 150 2C 00 b bed rel. hur.. Xb (% lý ý` 217 Fig. 8.1.: Temperature lift vs hurr-- ytv charge. 'C Temp. (0c) hur_ý it :- chance (Dehler Method) v 216 8.3.3. Modified Dehler Method Following the apparent over-estimation of *m using ; 'gor t_4 equation, it is considered that the Dehler Model could be codifies by using the heat balance equation (6.50) in conjunction with Dehier's proposal. This would enable LT and "1b to be estimated simultaneously by an iterative procedure. The results predicted on this basis are shown in . able 8.6 and figs. 8.15-8.18. In general, lower temperature lifts and hence hicr: r bed relative humidities Xb are obtained but Wb is virtually unaffected. With the high temperature and low humidity conditions at the exit end, W ee would tend to be very small. From Dehler's equation above (C. 61), Wb would then be virtually determined by Wei which obviously depends on the outlet conditions from the dryer. Since the modification of the model is in no way concerned with the conditions in the dryer, it is understandable that Wb remains unaffected. 217 TABLE 8.6 PREDICTION 0F r. S)0 pTIOI; BY KC I: FIE: D ýý"^ -T-ý` T Di f TAe 7' b y C0 C) (0c) C) (%) (0 C) (kg/kß) 50 .2 55 5.30 7.51 10.61 . 0054 . 4. 55 3 7.92 9.74. 16.47 . 0077 .6 55.6 13.21 11.89 22.29 . 0101 .8 55.9 17.17 13.96 28.06 . 0124 100 .2 100.1 0.92 1.64. 14.64. . 0087 . 4. 101 2.62 2.48 28.21 . 0144 .6 101.2 3.26 3.29 411 . 63 . 0202 .8 101.6 13.02 4.04 55.01 . 0257 150 .2 135.8 0.12 0.71 8.89 . 0121 .4 142 0.81 0.98 37.47 . 0217 .6 14.2.8 3.03 1.31 59.97 . 0311 .8 143 8.69 1.65 81.23 , or+04- 200 .2 125.1 0.02 1.21 -41.66 . 0162 . 4. 167.9 0.22 0.65 32.96 . 0294 .6 179 1.62 0.69 74.57 . 26 .8 179.6 5.93 0.87 104.42 . 0558 T 120 100 80 60 Temperature lift AT (0c) 1+4 20 0 - 20 - 40 rig 8.1 218 vs r" ýional Sýil: r'= mar ( 1, n=iý'io+ TiA}ýlaý^ '. '_1". ßr - 0 219 18 16 14 12 bkpt. cap. Wb ýic ý10 8 6 if 2 Fig. 8.16 BkDt. capacity vs `'r c`icr a (Yedlified Serie-- I-'e`:: oa 01- 0 .2 .4 .6 .8 o r, . h, 220 3. YT"t. Ilb ;) Fii. 8.17 : Bknt capacitv°va bed rel. hum. (Iodifto4 A. hler Ieth ) :n / ; 4: r&LUrö bad rol. hut. Xb (5) 221 Fin. 8.18 : Teminer Lure lift vs Lod ified Deh1e :e nod 120 100 80 60 temp. lift 40 LT (°c) 20 0 -2C -40 222 8. k. Bed Relative Humidity (BRH) Model The discussions of the previous section indicate trat the Lee- Cummings and Dehler Models give measurably disparate predictions of adsorption under the kind of conditions encountered. It is considered that perhaps these disparities arise because even though the ccnce-ts embodied in the model were developed for non-isothermal condi : io:: s the ; may not be appropriate for the unusually high temperatures and humidities that are tenable in the present study but :. hach scarcely arise in ordinary adsorption systems. Consequently, an alternative model t-s been derived using the experimental results presented in the last chapter. Since these were mostly obtained under high tempera ;, ure and humidity conditions which simulate those that would be expected from a dryer this model would provide a suitable basis for testing the other models. 223 8.1.1. Outline of Model Correlation of the data obtained from experiments led to the derivation of the following relationship between breakpoint capacity and breakpoint bed relative humidity for silica-gel/water system: 0.81 Wb =1 . 07Xb (7.10) The Bed Relative Humidity (BRH) Model, derived from the experimental data, is based essentially on the use of this equation in combination with the lumped heat balance equation for an adsorption column at breakpoint Wb (Ay +c TA. ) + Ay cT s bAýi ps 1 TA e- 0yc Vb c 6.50 () Ps Pg to obtain the outlet temperature of air and final loading of moisture on the adsorbent. 8.4.2. Evaluation of Model Predictions of adsorption temperature lift and breakpoint capacity using this model have been done and the results are presented in Table 8.7 and figs. 8.19 - 8.22, in the same manner as with the Lee-Cummings and Dehler Models. Fig. 8.19 shows the predicted temperature lift AT against fractional saturation f. Relating this to the predictions of the other models, it is clear that the BRH and Lee-Cummings Models show essentially similar characteristics in that low temperatures would give larger temperature lifts at low fractional saturation and smaller temperature lifts at high fractional saturation. Negative temperature lifts (i. e. temperature depressions) are obtained with high temperatures and low saturations. However, the data obtained from Lee-Cummings predictions show a stronger dependence on fractional saturation at high temperatures (> 100°C) with the result that the curves rise more steeply. The predicted breakpoint capacity Wb is presented as a function of fractional saturation and temperature in fig. 8.20. The decrease in Wb as temperature is increased is a direct reflection of the unfavorable effect of temperature on adsorption, which has been noted earlier in chapter 7 from the experimental results. The predictions of Lee- Cummings Model differ from this since larger or smaller breakpo--nt 22) T Bl. r 8.7 : P:; DIC IOi OF ADSORP: ICI: BY BED REL. 2IVE HUZ`1DITY YODEL TDi f T Ae Wb Xb ýT (o C) (0c) (irl) (iý-) (0C) (ks/k¬) 50 .2 54.60 5.59 7.70 10.21 . 0054 .4 54-30 7.08 10-30 15.47 . 0077 .6 54.10 8.49 12.90 20.79 . 0101 .8 54.00 9.83 15.46 26.16 . 0124 100 .2 85.50 2.52 2.88 -0.76 . 0087 .4 83.60 3.75 4.71 10.81 . 0144 .6 82.40 4.92 6.58 22.83 . 0202 .8 81.70 6.02 8.44 35.11 . 0257 150 .2 103.0 1.90 2.03 -23.91 . 0124 .4 99.8 3.08 3.69 - 4.73 . 0217 .6 98.0 4.20 5.41 15.17 . 0311 .8 96.8 5.26 7.14 35.03 . 04or+ 200 .2 114.2 1.68 1.74 -52.56 . 0162 .4 110.3 2.85 3.35 -24.63 . 0294 .6 108.0 3.97 5.04 3.57 . 01+26 .8 1 o6.6 5.01 6.72 31-42 . 0558 ý-, - Lý' 40 30 20 10 Temp. lift (oc°ý -10 -20 -30 -40 -50 -60 i1 ier. ýerý pure lif vs BIRH L-) 0 L26 14 12 10 bkpt. cap Vdb (%) 8 6 if 2 0 8.20 Bkpt. Ca-Dl c-*-, - vs o- V 0 fractionate -- atura tior. f. i6 14 12 10 )kpt . ap. 8 6 4 2 0 227 rid $ 21 B Kit capacity vs bed rel. hu: I ý, (r ) 02L,. 68ivIc "+ bed rel. hu.:.. KL `'F 0 op ... r. .. 0 Ni 4" Ch co "0N V` m a N -p- ON 0 ci " cc ä i i c. % N kp3. v : ew-Hwi<.. H, xa4^. `e [. . ýw Wyviý Por`., » n«. NM? rtw eo-pn. n 228 Fig. 6.22 Temperature lift vs hýýi? - C', - ^ý_ (3- ý ?, ' --- ) 40 30 20 10 Temp lift ö0 -10 -20 -30 -4o -50 -60 b 'ature M 229 capacities would be obtained with low temperatures, depending on the fractional saturation. To the extent that larger breakpoint capaci ties are predicted for low temperature operation over the entire range of fractional saturation, the predictions of the Dehler Model are similar to these. However, comparison of figs. 8.12 (or 8.16) and 8.20 shows that Wb predicted from Dehler Model shows much more sensitivity to fractional saturation. Since equation 7.10 is presumed to apply over a wide range of conditions the values of Wb and Xb predicted by this model understardably lie on a unique curve (fig. 8.21). In this regard therefore, the predictions of this model differ markedly from those of the other two models, especially the Lee-Cummings Model. The temperature lifts obtained experimentally were presented as a fraction of humidity and temperature in fig. 7.10. Similar data, predicted from the derived model, are plotted in fig. 8.22 and they also show similar characteristics. The effect of initial bed temperature Tsi on the prediction of breakpoint capacity and final temperature has been examined also. Results for an initial cold temperature (30oC) and an initial hot temperature (60°C) are presented in Table 8.8. Evidently, adsorbing on a hot bed gives higher temperature lifting capability, as has been noted previously in consideration of the Lee-Cummings Model. This is consequently attended however by a reduction in bed relative humidity Xb and hence, from equation 7.10, a reduction in breakpoint capacity Wb also. Though the predictions obtained from the Lee-Cummings and Del-der Models variously agree with the experimental deductions in certain respects, very significant discrepancies have been noted in marry other respects. In general the predictions of breakpoint capacity and temperature lift from Lee-Cummings Model tend to be lower than those obtained from the BRH Model. On the other hand the values predicted by the Dehler Model tend to be larger than those predicted by the BRH Model, especially at high fractional saturations. These observations therefore seem to indicate further that the Lee-Cummings and Dehler Models are inadequate for the severe temperature and humi; ice conditions considered. Besides, in correlating adiabatic adsorption by means of 4,30 TABLE 8.8 : EFFECT OF INITIAL BD Oi, " _= ;ý ýýý'" (BAD -D- REL TIVE HL ` _'_ Z' -ý ý- Tsi - 30°c. Tsi ' 60°C. ýT Wb Xb Fb x C) (%) C ý) C °C) O C`= ) 10.21 5.59 7.70 12.71 5.08 6.84 15.4.7 7.08 10.30 18.27 6.35 9.02 20.79 8.4.9 12.90 23.89 7.54. 11-13 26.16 9.83 15.4.6 29.46 8.65 13.21 -0.76 2.52 2.88 5.44 2.08 2.27 10.81 3.75 4.. 71 17.31 3. o6 3.66 22.83 4.92 6.58 29.63 3.97 5.05 35.11 6.02 8.44 42.01 24.84+ 6.44 -23.91 1.90 2.03 -16.51 1.54. 1.57 -4.73 3.08 3.69 2.97 2.48 2.82 15.17 4.20 5.4.1 22.97 3.36 1.11 35.03 5.26 7.14. 43.03 4.18 5.38 -52.56 1.68 1.74. -4. +.. 66 1.37 1.35 -24.63 2.85 3.35 -16.53 2.30 2.57 3.57 3.97 5.0+ 11-77 3.18 3.84 31.4.2 5.01 6.72 39.62 4.01 5.11 Wl] 231 humidity, Lee-Cummings Model vitally neglects the direct effects of the final temperature conditions in the bed. Clearly, both humidity and temperature constitute the most important parameters that affect adsorption. Concentration on correlation in terns of humidity while neglecting temperature therefore renders the procedure defective. Regarding the Dehler Model, it is evident that the conditions of the influent air to the adsorption bed influence the predictions rather too strongly. In addition, the assumption that half of the bed is unused would seem to be rather arbitrary and is unlikely to hold for all conditions. In view of these, since the BRH Model is derived from the experimental data obtained under more appropriate conditions with regard to heat pumping requirements and strongly reflects the effects of temperature and humidity, both initially and finally, it will provide the basis for further prediction of the overall performance of the AHP dryer. 8.5. Overall Cycle Prediction by Bed Relative Humidity Model 8.5.1 Thermal Efficiency. The prediction of thermal efficiency of the TSC heat pump dryer T by this model is illustrated in fig. 8.23 in relation to the thermal efficiency of a comparable conventional non-recycle dryer, fl cony. The comparable conventional dryer is considered to be such that it operates at the same temperature and fractional saturation as the dryer incorporated in the heat pump dryer. Both efficiencies are evidently dependent on temperature and fractional saturation. While the efficiency of the heat pump dryer shows more sensitivity to temperature the efficiency of the conventional dryer depends more strongly on fractional saturation. It can therefore be deduced that AHP drying would provide considerable advantage over conventional drying under conditions of low temperature and low fractional saturation. At high temperatures and fractional saturations conventional drying would however be preferable. Thus, for any temperature TDi, there exists a maximum fractional saturation f at which the efficiency of the BHP dryer would exactly equal that of the conventional dryer. If thermz-l ýJý OF P .. jRýK R1i ý ýý üF TSO : "ý1: TABLE ý. F? DICTIOIý (Br2ýi YODEL) TD 1 ý°(1 ) f "RH nco nv . COP max. (/) KJKJ COP (KJ/KJ) T s (k, - silica-re x _ir 50 .2 39.72 4.15 22.56 . 43 1.88 . 10 . 4. 47.88 23.03 11.08 . 54. 4.89 . 14 .6 54-. 64. 41-34 7.46 . 63 8.50 . 15 i .8 59.31 59.22 5.73 . 69 12.12 . 16 100 .2 25.81 11.12 -37.30 - 10 . 28 . 51 . 1+ 34.94 27.57 12.25 . 22 1.78 . 60 .6 42.00 44.57 5.32 . 42 7.80 . 63 .8 48.14 60.50 3.61 . 56 15.50 . 62 150 .2 22.08 13.4.5 -2.46 -. 87 35.35 1.13 . 4. 31.85 29.93 -12.71 -. 13 1.02 1.21 .6 39.19 1+6.57 6.71 . 22 3.33 1.22 .8 43.99 62.38 3.33 . 1+2 12.56 1.24 200 .2 20.97 14.60 -0.54. -1.74 32.14 1.75 .4 31.31+ 31.56 -2.18 - . 48 21.80 1.84 .6 38.19 47.88 31.81+ . 03 . 11 1.86 .8 4+3.33 64.4.2 3.37 -33_ t 9.67 1.85 ýJý ri6.23 Thermal Effficiercv of TSC Dryer (3Ri-? b: odel} 100 90 x efficiency of conven-c::: l ý--, -cr cc nv 80 70 60 Efficiency n (/0 D0 4C 3c 2C iC l 0 efficiency of -, SC ar:,. -er., T Di is air irret temperature to dryer. ýi i ýX >'TDi5O ? 0oß 1000 , 500 ° TDi=200C _': '-C ion- _ :r-: gor. r .C "34 efficiency is the only criterion for co: 7:, ar4n: - h two dryers, this may be deemed to be the efi iciency at 'hich the _ý? r-, 'er rast be replaced by a conventional dryer. The re; lacement efficiencies for various temperatures, deduced from the plots of fig. 8.23, are presented in Table 8.10. TABLE 8.10. REPL: ýCEýýT ý --CIETN: CIyS OF D_RY_ R BY CG I; i E' i0 r:? ,, - ý. "_y. _1- TDi oc f effcy. .' 50 0.90 61.5 100 0.56 41 150 0.4+5 31+ 200 0.1+1 31 From the table, it is clear that the i-IP dryer will be replaced t an efficiency of 61 . 5`ý, with a by a conventional dryer a corresponding fractional saturation of 0.9, for a drying temperature of 500C. But for 2000C, it will be replaced at only 311', - efficiency, and 0.4 fractional saturation. A more succinct illustration is-provided by plotting the loci of these replacement points in fig. 8.24. This enables values to be determined for iilzei nediate temperatures. For exam. le at 120 OCI tie replacement efficiency will be 3 -, with C-51 fractional saturation, as shown in the graph. =35 i 00 90t 80 Fig. 8.24 Loci of e lacem er` Ef icjenc- ^, _? ^= Saturn ;, i or of JC D'" -e r 11.0 x efficieno y, r .9 ` 04= sacturation, f i .8 70+ Efficiency n 60 50ý 1+0 30 20 10 C 7 fr--c tional 6 sý ýýr tion l' f 1 rtý "4 T .3 .. 2 r, r 1r -Il tý r_'- übe -- 236 8.5.2. Comparison with Mechanical Heat Pump Drying Following the discussion of Section 4.2, the maximum 12R that could be expected from a closed cycle mechanical heat pump dryer for the present maximum attainable temperature of 100°C is about 3-4 kg/KWH. Using an BIER value of 3.5 in equation 4.7 the thermal efficiency of the system would be about 74-z- Now, this can be related to the AH? dryer by referring to fig. 8.23. It is evident then that for a temperature of 100°C the AH? dryer on its part cannot be expected to exceed an efficiency of about 55,:. However, it is interesting to note that the two systems in consideration have opposite performance characteristics, in relation to temperature. While lower efficiencies are attainable with low temperature in mech heat pump drying, higher efficiencies are attained in the AHP dryer. For the previous maximum attainable temperature of about 50°C and the reported associated MER of about 2.5 kg/KWH, the equivalent efficiency would be 5 0. But from fig. 8.23, considerably higher efficiencies, up to 60-65, ý-., could be expected for the AHP dryer operating at the same temperature. Hence it can be inferred that for low temperatures AHP drying will likely provide more scope for heat recovery than mechanical heat pump drying but for high temperatures the reverse may hold. 8.5.3. Coefficient of Performance COP, as has been stated, is a measure of the heat pumping efficiency of a heat pump system. The theoretical maximum COP and the estimated actual COP of the TSC heat pump are presented in figs. 8.25 and 8.26. From equations 6.30, TAe TRi - TDe (6.30) COPýx CT _ýC Ae De TRi which was derived by analogy with a heat driven Carnot engine, it is clear that the theoretical COP may be very high when Tie and hence AT, is low and tends to infinity as -T approaches zero. 237 It has been shown that under certain conditions the inequality AT<0 may hold. Under such conditions then COP max may also become negative. Similarly, from equation 6.31. (6-31) COP = TAe - TDe 1-rD TRi -TA e) it is clear that the actual COP of the system may also become negative when AT is negative. This phenomenom is illustrated in figs. 8.25 and 8.26 by dashed strokes. However, it is noted that the magnitude of the negative values of COP max and COP do not have any practical significance except merely indicating situations of net energy input to the system rather than output. The predictions indicate that the actual cycle COP generally increases with fractional saturation but decreases with air inlet temperature, as shown in fig. 8.26. A comparison with fig. 8.23 shows that COP is favoured by the same conditions that favour the overall thermal efficiency of the cycle. However, even under the most favourable conditions the highest attainable predicted COP is less than 0.8. This implies that even at best more energy actually needs to be put into the heat pump than it delivers to the dryer. The values of theoretical COP and actual COP have been used to estimate the thermodynamic efficiency of the heat pump and the data are plotted in fig. 8.27. Evidently, since the conditions that produce negative values of COP max and COP are the same, from equation, 4.9 T= COP (4-9) max positive values of thermodynamic efficiency may be obtained when both COP and COP are negative. Since such values of COP can only be of max trivial significance, the corresponding values of nT are also equally trivial. Values of nT corresponding to negative COP's are also indicated by dashed strokes in fig. 8.27. Bearing this in mind, it is observed that generally log, values of effective n T, of the order 5-2C(,, are predicted. This poor performance can be interpreted tc mean a very severe degradation of the high quality- heat input to the desorbin6 column 238 Fig. 8.25 Theoretical C0 of NCO FL --- e : era : ure TDi is air inlet, ýc dr^, -er. COPý 4o, 3a 20 10 0L -lot -20 -30 fractional saturation f ý1 } 0 200°C 1 50°C J C 0C 500C 0° C 239 Fig. 8.26 r ctual COP of TSC Heat : um-, 2.0 T T 0C xo Týl_ i 00ý C 1. +I n1=150ýC 5Iv mom, Z=2000 c L 1.0 "5 COP KJ/KT 0 -5- -1.0 _1 ,5 L. o 0 240 Fig. 8.2% ihermod-, -na iic ^f f_cie:, cv o "TSC ; =ý: 50 .o 30 Thermodynamic efficiency nT 20 rC 10 0 x TDi 50°C Om= 107,0 `Di TDi l 50°C V Tai=20000 T, is air inlet temperature to dM, ýer. Dl fractional sa ;, ura tior_ f. 0 241 8.5.4. Adsorbent Bed Size The amount of adsorbent and hence the size of the adsorption column required for any given conditions of operation can be estimated from the predicted adsorbent bed capacity and the rate of moisture removal in the column. Thus: M GG b6y- s Wb (8.5) where Ms is weight of adsorbent required and 8b is the cycle time required for the bed to attain breakpoint before it is regenerated. For a cycle time of 2 hours and airflowrate of I kg/hr, the silica-gel adsorbent requirement has been predicted for the range of conditions considered. The results are presented in fig. 8.28. Evidently, Ms will be virtually independent of the air fractional saturation in the dryer f and very heavily dependent on dryer air inlet temperature TDi. Following the discussions of sections 8.1 and 8.4., both breakpoint capacity Wb and moisture removal in the bed Ay increase as f and will thus tend to produce counter-balancing effects on Ms, keeping it fairly constant as f varies. On the other hand, while 6y increases with TDi, Wb is adversely affected, and hence Ms will increase very considerably with TDi. Since the smallest bed size would imply least capital expenditure, low temperature conditions would therefore be considered most favourable in this regard. 8.5.5. Initial Bed Temperature Calculations have been done for a dryer operated under the two sets of conditions indicated in Table 8.11 and with initial bed temperatures (Ti siof 70°C and 960C respectively. The results are shown in the table against similar results for a bed initially cooled to 30°C. The analysis of section 8.4.. has showed that if the adsorbent bed is not allowed to cool down sufficiently after regeneration, the effect is a higher temperature lift but a lower breakpoint capacity. Thermal efficiency would therefore be expected to be improved by adsorbing on a hot bed, though ` severe penalty would be incurred by way of substantially increased bed size re iuire: cnt. L'1"L Fig. 8.28 ýcsorD2r: ` D°z size for SC -'r-e 2. Di =200 - A 1.6 1.2f O 150 1.2. Mass of solids m s 1.0 (kgsilica-el kg air .8 .6 .4 .2 0 _I- ; O: .i rL c. ion' 7' 6 :2 if 6 .ý'. 0 243 TABLE 8.11 : EFF--C"-' OF IT ITI L BED T7'1r7- O l\ y R1 V RY-:,:, E OF "'3- (BRH 1,: C DLL TDi=1000C, f=0.6 TDi=1500C, T = 300C . T =700C T =300C T . =900C. ' s i si 31 1 3 Xb( ) 6.40 4.55 3.04 1.69 V, b( v) 4.59 3. x+8 2.54. 1 . 58 M (kg) 0.63 0.84 1.21 1.95 S TI RH( ) 42.00 1+3.68 31.85 36.86 n (%) 44.57 44.56 29.93 29.91 Conv COP 5.32 3.83 -12.72 7.18 max COP 0.42 0.60 - 0.13 0.30 r}r 7.80 15.67 1.02 1.14 2i}-} For the latter set of conditions consi ereil ý= 150°C, ( it is observed that lis almost doubles e increases b`. - about rýi 1 j; _. The COP and thermodynamic efficiency of the heat pu ;, are also improved when TSi is high. 8.5.6. Regeneration Temrera Lure Table 8- 8.12 at shows a comparison b.. ": en re¬ , ion a,, SOG ., and 150°C for t,, -. -o sets of dryer operating conditions. spare: týý" 1ý' s is not affected by the temperature of 1 . _i efficiency would however be expected to diminish lo-:: er reE-ener: - ; or. temperature, in the same manner as has been discussed for the ^uil ýriun Model. Though n drops, COP and n imtrove with lower T : mis si. therefore, implies that despite the decline in overall energy performance there is less degradation of heat in the system when regeneration is carried out at lower temperature. TABLE 8.12 : EFFECT OF REGEL 'ION r7, EYPE : 2-JRE OI: PERFORMýC , OF TDi = 100°C, f=0.6. TDi= 150°C, f=0.4 TRi = 200°C. TRi=150C. TRi 200°C TRi=150°C Xb (%) 6.40 6.27 3. ()1+ 3.11 Wb (i) 4.59 4.61 2.54. 2.55 M (kg) 0.63 0.64 1.21 1.21. s TI RH (icl, ) 42.00 37.82 31.25 29.58 conv. (i) 44.57 41-. 1+7 29.93 30.61 COP 5.32 3.86 -ý 2.72 -6.94 max COP 0.42 0.46 -. 13 - . 16 ,, 7.80 11-79 1.02 2.3tß m 1 245 8.5.7. PSC Dryer Since the analysis of the PSC dryer based on the Heat and Mass Balance Model has clearly indicated that it will not improve the thermal efficiency of drying processes and that it is less efficient than the TSC dryer, further consideration on the basis of the BRH Model cannot be expected to provide more encouraging results. Nevertheless, application of the latter to the PSC dryer is worth considering, in view of the fact that it enables the solid adsorbent requirement to be estimated. Predictions for TSC and PSC drying systems, operating under the same drying conditions, are shown in Table 8.13. Though the data reinforce the superior thermal performance of TSC over PSC, the required bed size for the latter would clearly be much smaller. This arises directly from the higher bed relative humidity Xb and thus higher breakpoint capacity 71b which are associated with the high pressure condition in the adsorption bed. In this regard therefore the PSC dryer is seen to offer considerable advantage over the TSC dryer. Investigation of the impacts of the pressure 1I and condenser temperature TA. (Table 8.110 on bed size, also show that high pressure and low condenser temperature would tend to give larger Xb and Wb and hence lead to smaller bed size. The small bed size requirement generally associated with PSC drying may be deceitful however, considering that extra capital outlay will be incurred in incorporating a compressor and a cooler in the cycle. The cost of the cooler is partially offset though by the fact that the need for a regeneration heater is obviated. Increasing the pressure or reducing the condenser temperature, while leading to even smaller beds, will imply even more capital expenditure for a larger compressor or cooler. In addition, more capital cost will also be incurred in making the adsorption column into a pressure vessel. 26 TABLE 8.13 cG1ý_`ý sc1: OF OF 1 and P. 30 Jý ýi f=0.6, Ti -150°c TSC PSC Xb (%) 4.83 33.18 71b (i) 3.77 17.45 (kg) M 1.22 . 12 s "' (') RH 39.19 9.34 COP 6.71 1.0 max COP . 22 . 06 TI T(; ') 3.33 6.22 TABLE 8.11 : EFFECTS OF TOTAL PRESSURE AND COI, -DEJ; SER TE1. IPERATURE 0. ',, ' F RFORLL-:., CE OF PAC DRYER BRri MODEL f=0.6, TDi=150°C Ti =6 atms Tgi = 50°C ri = 10 atms TXi=50 °C n= 10 atms TAi= 75°C ' () Xcl'o 33.18 50.52 26.79 b Wb () 17-45 26.45 14.. 73 Ms (kg) . 12 . 05 . 27 RH("") 9.34 8.24. 7.32 COP 1.00 1.00 1.00 max COP . 06 . 02 . 06 n 6.22 2.47 5.73 CHAPTER NI CONCLUSIONS AND SUGGESTIONS FOR FURTHER WORK 247 9.1. 9. CONCLUSIONS AND SUGGESTIONS FOR FURTHER WORK Conclusions. In conclusion, the general deductions that may be drawn from this study are presented as follows: 1. Measurements taken on an industrial convective tumble dryer show that the dryer operates at a thermal efficiency less than 4. and that -öä of the heat supplied to the dryer is lost as sensible heat in the air exhausted from the dryer. This is considered typical of many industrial dryers and clearly indicates that the main scope for imtrovement of thermal efficiency of dryers is in recovering the heat lost in the exhaust air. 2. Two conventional methods of heat recovery have been considered, namely: (a) exhaust air recirculation, (b) heat exchange recovery. (a) Calculations show that air recirculation can affect the thermal efficiency of a drying process either favourably or adversely, depending on the mode of operation. Operating with a constant fractional saturation of air, the best thermal efficiency would be obtained with the highest possible recirculation ratio, whilst the best efficiency for operation with fixed wet-bulb temperature would be obtained with the minimum permissible recirculation. It was shown however that the conditions weich favour improved efficiency in either operational rno1e are also those that retard the rate of drying. A method of optimizing air recirculation in terms of thermal efficiency and dying rate has been suggested. (b) Drying with a heat exchange recuperator helps improve t; vermal efficiency, especially under low fractional saturation and high temperature drying conditions, but jr. vie,. of the very low heat .. r`n. 3fer coefficients associated with air-air heat exchange and the lo,: temperature driving force obtained, this necessitates ir: to1erabiý- lýrýe -ýe_t ercný : E, er surface areas. 248 j, Experiments on adsorption of water vapour from air on silica-gel adsorbent affirmed that temperature and humidity are the most important factors that affect the process. In particular, the results show that with a combination of high temperature and fairly low humidity conditions a depression of air temperature as it passes through the adsorption column would be obtained, rather than temperature lifting. Hence under such conditions air left the column colder than it entered, i. e. AT <0 1ý. By defining the bed relative humidity at breakpoint Xb as a function of air inlet humidity YAi and average bed temperature T Ae thus: f (Y Ai' TAe) the adsorption process was found to be described by a polynomial function, similar to the system equilibrium relation, of the form: n Wb = KXb where K and n are constants and Wb is the breakpoint loading of moisture on silica-gel. Correlation of the experimental data produced the following equation: 0.81 Wb=1.07Xb 5. The correlation was used to set up the Bed Relative Humidity (BRH) Model for predicting the performance of the Adsorption Heat Pump (AHP) Dryer. The predictions of this model enabled similar predictions obtained from two other models synthesized from various proposals put forward by other investigators to be tested against experimental results. These other models, Lee-Cummings and Dehler, variously agreed with the BRH Model in certain respects but in general they were found to be inadequate for the unusually high temperature and humidity conditions encountered. The BRH Model was therefore considered to be most appropriate for predicting the system performance. 2+9 6. Comparison of predictions of performance of the M dryer operated as temperature swing cycle (TSC) and as pressure swing cycle (PS-%') showed that owing to the large energy demand for compression, the TSC dryer would provide much higher cycle thermal efficiencies. The performance characteristics of the TSC dryer showed that the best thermal efficiencies, up to 60Z, would be expected if the dryer is operated with low temperature and high fractional saturation. However, in comparison with an equivalent non-recycle conventional dryer which is deemed to operate at the same temperature and fractional saturation, the TSC dryer would be expected to provide measurable advantage only at low temperatures (< 1000C) or low fractional saturations ( <0.5). This arises mainly because the efficiency of the conventional dryer was found to be much more sensitive to fractional saturation than that of the TSC dryer. 8. In relation to mechanical heat pump drying the AHP dryer was found to compare favourably only at low temperatures ( <1000C). 9. The predicted coefficient of performance (COP) of the heat pump is such that more heat would be put into the cycle to regenerate the used bed than would be extracted from the dryer effluent air. In addition, very low thermodynamic efficiencies, less than 20, are predicted, implying that the heat supplied for regeneration at high temperature would be substantially degraded in the system. 10. Very large adsorbent bed sizes are estimated for high temperature drying but smaller beds would be required for lower temperatures. 11. In view of both the poor thermal performance and the large bed size requirements, AHP drying will not be appropriate for the high temperature conditions for which it was initially conceived. Nevertheless, the cycle seems attractive for low temperature ( <100°C) application. 250 9.2 Suggestions for Further : '; ork 1. Following the above conclusions, though the origina. purpose of the JJ dryer is lost, it may be useful to investigate its applicability to low temperature drying further. In this regard, there seems to be sufficient justification for construction of a prototype AHP dryer for practical demonstration ar» tes :, inF- of the performance of the overall unit. 2. Data obtained from practical tests will enable a more precise comparison to be made with conventional drying and with mechanical heat pump drying. This will also enable the appropriateness of the present theoretical prediction Model to be tested. in particular the practical performance data will highlight other losses, including those associated with pressure drops across the adsorbent beds and along the conduits, which have not beer. s uff icier- . 1; ý considered yet. 3. There may be need for more rigorous modelling of the s: jstem, especially in the adsorption bed, if practical data are un::. atched by the present theoretical predictions. 4. The use of silica-gel adsorbent here was based mainly on its relative cost advantage. It may be interestinij to extent the consideration to other types of adsorbent, especially molecý-1ar sieves. 5. The i.. H. P. may have possible applications in domes tic air conditioning and heating and it may be useful to investigate this further. REFERENCES ý-r 251 REFERENCES 1. The British Petroleum Company Ltd. : Statistical Rev-e: ': of the World Oil Industry, (1965,1970,1975 arO- 1960). 2 U. K. Department of Energy : 'Digest of U. K. Energy. - Statistics', HASO, (1970 and 1980). 3. Kantyka, T. A. : Department of Industry Energy Seminar. 'Energy Conservation in the Chemical Industry, Newcastle, (1976) . 4.. Grant, C. D. : 'Energy Conservation in the Chemical and Process Industries', I. Chem. E, (1979). 5. Griffin, J. M., Steel H. : 'Energy Economics and Policy', Academic Press, New York, (1978). 6. United Nations : U. N. Statistical Year Book, (1977). 7. CEFIC : 'Consumption of Energy Sources of the European Chemical Industry for Feedstock and Fue1, l, (1980). 8. Reay, D. : Chem. Engr.; 311, (July/August 1976), 507. 9. Lawton, J. : Phys. Technol; S (1978), 214. 10. Hodgett, D. L. : Chem. Engr. 311, (July/August, 1976), 510. 11. I Chem. E. : 'Materials and Energy Resources', (1976). 12. U. K. Department of Energy : 'Energy Paper No. 11, Energy Research and Development in the U. K. ', HMSO, (1976). 13. Liley, P. E. : 'Physical and Chemical Data', in 'Chemical Engineers' Handbook', (ed. R. H. Perry and C. H. Chilton), 5th Edition, (1973). 14.. Gallant, R. w. : 'Physical Properties of Hydrocarbons, Vols. I and 2', Gulf Publishing Co. Houston, (1968). 15. Tsao, G. T. ; Wheelock, T. D. : Chem. Engr. Z., 13, (1967), 201. 16. McCormick, P. Y. : 'Solids Drying Fundamentals', in 'Chemical Engineers' Handbook (ed. R. H. Perry and C. H. Chilton), 5th Edition, (1973)- 17. Sherwood, T. K. : Trans. AIChE. , (1936), 150. 18. Cealgske, N. H. , Hougen, O. H. : Trans. AIChE, L, (1937), 283. 19. Sher-wood, T. K., Comings, E. ';; .: Trans. AIChE, 22, (1932), 118. 20. Seborg, C. O. : Ind. ä Engineering Chem; , 2ý, (1937), 169 252 21. Shepherd, C. B., ; Haddock, C. ; Brewer, R. C. Ina. 6_ Engineering Chem. ; Q, (1938), 389. 22. Molstad, M. C. ; Farevaag, P.; Farrel, J. A. : Ind. & Engineering Chem. ; ýQ, (1938).. 1131. 23. Treybal, R. E. : 'Mass Transfer Operations' 2nd. Ed. McGrav: Hill, (1968). 24.. Nissan, A. H. ; Kaye, V'W . A. ; Bell, J. R. : :: IChE. J. 1. (1959), 103. 25. Gilliland, E. R. : Ind. & Engrng. Chem. ; ., (1936), 506. 26. Wenzel, L. A. ; White, R. R. : Ind. & Engrng. Chem ;, (1951), 1829 27. Sherwood, T. K. : Ind. & Engrng. Chem. ; L1, (1929), 976. 28. Newman, A. B. : Trans. AIChE ; 26 (1931), 203. 29. Slitcher, C. S. : 'Theoretical Investigation of the Motion of Ground Waters' U. S. Geol. Survey, 19th Annual Report, Part 2. (1897-8), 301. 30. Wheat, J. A. ; MacLeod, D. A. : Can. J. Chem. Engrng. I, (A pril, 1959), 47 " 31. Keeys, R. B. : 'Introduction to Industrial Drying Operations', Pergamon Press, (1978). 32. Heertjes, P. M. ; Tuider, H. E. : De Ingenieur, 2A, 28, (1962), 53. 33. Bates, H. T. : Chem. Engrng. Prog. ; 56,7, (1960), 52 3L.. Peck, R. E. ; Kauh, J. Y. : AIChE J; j, 1, (1969), 85. 35. Oliver, D. R. ; Clarke, D. L.: Proc. Inst. Mech. Engrs; 187, (1973), 515. 36. Kisakurek, B .: Peck, R. D. ; Cakaloz, J. : Can. J. Chem. Engrng, , (F eb. 1975), 53. 37. Berger, D. ; Pei, D. C. J. : Int. J. Heat & Mass Transfer, 16, (1973), 293. 381 Luikov, A. V. : 'Heat and Mass Transfer in Capillary - Porous Bodies', j Pergamon Press, Oxford, (1966). 39. Foust, A. S. ; . 'ý'enzel, L. A. ; Clump, C. 7,. ; Maus, L.; and Anderson, L. B. : 'Principles of Unit Operations' , John Wiley, (1960). 40 Poersch, W. : 'Present State of Drying Technology' - I. Chem. E. Solids Drying Course, Birmingham, (1977). -w- 253 4'1. Friedmann, S. J. Marshall , R. : Chem. ngrng . Profi; 45,8, (1949), 482. 4.2. Saeman, -, '. '. C. , 11. itchell, T. R. : Chem. Engrng. Profi, Q, 9, (' 95i. ) , 4.67. 1.3. Brutton, C. F. ; Miller, C. O. ; Schnette, ":. H. : Trans. AIChE; 18, (1942), 123. 41+. Kolbusz, P. : 'Industrial Applications of Heat Pumps - Heat Pumps and their Contribution to Energy Conservation', Leyden, (1976), 201. 45. Hawtin, P.; Reay, D. : 'Drying - The Increasingly Expensive Art, ' I. Chem. E. Congress, London, (1974. ). 4.6. Ashworth, J. C. : 'Energy Management in Industrial Drying'. I Chem. E. Solids Drying Subject Group Tech. Leeting, Birmingham, (Dec. 1980). 1+7. Ashworth, J. C. : 'Energy Efficiency of Convective Solids Dryers', in 'Energy for Industry' (ed. P. T. C'Callaghen), (1979), 34.9. lß. 8. Ashworth, J. C.: 'Energy Performance of Drying Systems' , -Continu-ng Education Course, 'Industrial Solids Dr; ing', Birmingham, (1; 80). 49. Free, D. : 'The Design of Industrial Fixed Bed :. dsorbers', Ph. D. Thesis, CNAA, (1966). 50. Mantell, C. : 'Adsorption' , 2nd Ed. ; McGraw Hill, (1951 ) 51. Hougen, O. A. ; Watson, K. M. ; Ragatz R. A. : 'Chemical Process Principles, Part V, 2nd Ed.; John ; '; iley, (1954. ) . 52. Freundlich : cited in 'Adsorption' , C. L. ? antell, (1951) , p. 25. 53. Langmuir, J. : J. Am. Chem. Soc.; L8, (1916), 2267. 54.. Rozycki, J. : Chem. Engrng. Prog. ; 7,5, (15176), 69. 55. Brunauer, S. : 'The Adsorption of Gases and Vapours', Princeton University Press, (1945). 56. Brunauer, S. ; Emmett, P. H. ; Teller, E. : J. Am. Chem. Soc ; 60, (1938), 309. 57. Zsigmondy : cited in 'Adsorption', C. L. L: antell, (1951), p. 23. 58. Rabinoivitch, E. ; ':; ood, V. C. : Trans. Faraday, Soc ;, (1936) , 94+7. 254. 59. Carter, J. Z. ; Barrett, D. J. : Trans. I Chem.,:,.; ýl , (1973), 75. 60. Jury, S. H. ; Edwards, H. R. : Can. J. Che m. Engrng.; 49, (Cct. 1971), 663. 61. Hubard, S. S. : Ind. & Engrng. Chem. 46, 2, (1954), 356. 62. 1Iukesh, D. ; Rao, C. S.: Indian Chem. Engr. ; j, 4, (1975), 29. 63. Jury, S. H. ; Horng, J. S. : Can. J. Chem. Engrng. ;j (Feb 1973) , , , 131. 64. Simpson, E. A. ; Cummings, W. P. ; Chem. Engrng. Prog; 60,4, (19o64) 57. ) 65. Carter, J.. : Chem. & Proc. Engrng. ItZ, 9, (1966), 70. 66. Derr, R. B. : Ind. & Engrng. Chem. ; 30,4., (1938), 3&+" 67. Worthington, R. : Chem. & Ind.; (Dec. 26,1964), 2110. 68. Grayson, H. G. : Ind. & Engrng. Chem; 1+7,1, (1955), 41. 69. Zwiebel, I. ; Gariepy, R. L. ; Schnitzer, J. J. : AIChE. J; 18,6 (1972), 1139. 70. Thomas, H. C. : J. Am. Chem. Soc.; 62, (1940), 1583 66, (1944), 166tß. 71. Lightfoot, E. N.; Sanchez-Palma, R. J.; Edwards, D. O. 'Chromatography and Allied Fixed Bed Separation Processes', in 'New Chemical Engineering Separation Techniques', ed. H. M. Schoen, Interscience Publishers, (1962), 99. 72. Hougen, O. A.; Marshall, W. R. : Chem. Engrng. Prog. ;1,4., (191+7) , 197. 73. Kyte, T. S. : Chem. Engrng. Sci.; 28, (1973), 1853. . 74.. Tien, C. ; Thodos, G. : AIChE. J.; ý, (1959), 373. 75. Morton, E. L. Murril, P. º,.: AIChE. J.; , (1967), 965. 76. Antonson, C. R.; Dranoff, J. S. : C. E. P. Symp. Series, 96, (1969), 20. 77. Davis, K. G., Manchanda, K. D. : Chem. Engrng. ;LI, 19, (1971. ) , 102. 78. Glue ckauf , E. : Trans. Faraday So c; 21, (1951), 1540. 79. Garg, D. R.; Ruthven, D. M. : Chem. Engrng, Sci; 28, (1973), 791,799 29, (1974), 571,1961. 80. Carter, J. V..; Husain, J. : Trans. I. ChemE; ý0, (1972), 69. 81. Hiester, N. K.; Vermulen, T: Chem. Engrng. Prog.; 1±8,10, (1952), 505. 255 82. Vermulen, T. : , dvances in Chem. Engrng; 2, (1 X58), 11TH. 83. Masamune, S. ; Smith, J.?,:. : AIChE J; 11 , (1965), 34. 84. Rosen, J. B. : J. Chem. Phys. 20, (1952), 387. 85. Rosen, J. B. : Ind. & Engrng. Chem. ; _46, (1954), 1590. 86. Masamune, S. Smith, J. 2. :I&E. C. Fundamentals, 1, (1964)9 179. 87. Acrivos, A. : Ind. & Engrng. Chem. ; L-8,4, (1956) , 703. 88. Sanlaville, J. : Genie Chimique, Z, 4, (1957), 102. 89. Carter, J. W. : Trans. I Chem E. ;1, (1966) , 253. 90. Carter, J. 14. : Trans. I Chem E. ; 46, (1968), 213. 91. Meyer, O. A. ; 1; Veber, T. V. , ': AICh. E. J.; j, 3, (1967), 457. 92. Carter, J. W. : AIChE. J. ; 21 ,2, (1975) , 380 93. Michaels, A. S. : Ind. & Engrng. Chem ; ll, 8, (1952)q 1922 94+. Nutter, J. I.; Burnett, G. : AIChE. J.; 29 2, (1963), 202. 95. Collins, J. J.: C. E. P. Symp. Series, §j, 74., (1967), 31. 96. Leavitt, F. V. :: Chem. Engrng. Prog; ý8,8, (1962), 54. 97. Eagleton, L. C. ; Bliss, H. : Chem. Engrng. Prog. 49,10, (1953), 543. 98. Getty, R. J.; Armstrong, K. P. : I. & E. C. Proc. Des. & Dev. 1, (1964), 61. 99. Lee, H. : Cummings, W. P. : C. E. P. Symp. Series, , 74, (1967), 42. 100 Dehler : Cited in Adsorption', C. L. Mantell, (1951), p. 262. 101. Ross, M. McLaughlin, E. R. : Ref. En¬rng, (Feb. 1951), 167. 102. hiller, A. 7,7. ; Roberts, C. ';.. . Ind. Chemist, 34, (: arch, 1958), 141. 103. Coulson, J. Y. ; Richardson, J. F. : 'Chemical Engineering Vol. 3, 2nd. Ed. ; Pergamon Press, (1979). 1 UJ . Hougen, O. A. ; Dodge, F. ºT .: 'The Drying of Gases', Edwards and Arbor, Michigan, (1947). 105. Geankoplis, C. J. : 'Transport Processes and Unit Operations', Allyn and Bro w: n, Boston, '1978). 106. Rathmell, B. L.; Bateman, P. J. : Trans. Inst. Hea. t and 7 ent. Engrs ; 15. (1952), 471. 107. Noden, D. : Chem. & Proc. Engrng. ;. 4, ('1972), L4 E 108. Hanna, V1 . T.; Frederick, T. J. : Chem. -nrrn-. FroL. -74. ; (976,71. 109. Townsend, D .W.: Chem. Engr. ; 361 , (C ct. 19E-O-), 632 256 110. , Velty, J. R. ; 'Dicks, C. L. ; 7. ilson, R. E. : '-Fundamentals of Momentum, Heat and Mass Transfer', John ; ': iley, (A, 969). Ill. Holmar_, J. P. 'Heat Transfer' , 3rd Ed ; EcGra. Hill, (1972). 112. Perry, R. H. ; Chilton, C. H. : 'Chemical Engineers' -a- 'book' , 5th Edition, McGraw Hill, (1973). 113. Lawton, J.: 'Closed and Open Cycle Heat Pumps for Drying', I Chem E. N. -'17,. Branch Symposium, Salford, (. _pril, 1981) . 1114.. Cole, J. W. : Chem. Engr. ; 296, (Feb. 1975). 76. 115. Freshwater, D. C. : Trans. I. Chem . E, 29,1, (1951), 11+9 116. Freshwater, D. C. : British Chem. Engrng; 6,6, (1961), 388. 117. Baxter, W. R. S. Hodgett, D. L. ; Kolbusz, P. ; and Lawton, J: Proc. Conf. on Electricity for Materials Processing and Conservation - 'Drying by Heat Pumps' , London, (March, 1977). 118. Lyle, 0. : 'The Efficient Use of Steam' WISO, (1947). 119. ERG Open University and ERSU Rutherford Laboratory: 'A Chemical Heat Pump /Energy Storage Device and its Application' - Report on Feasibility Study, (Nov. 1978). 120. Weiner, A. L. : Chem. Engrn¬.; 81 , 19, (1974), 92. 121. Haselden, G. G.: 'Alternative Cycles for Heat Pumping' , I. Chem. E. N. '#T. Symposium, Salford, 1981) . 122. Offenhertz P. O. : 'Chemically Driven Heat Pumps for Solar Thermal Storage' , International Solar Energy Conference, Nev; Delhi, (Jan. 1978). 123. Fujii S.; Kaweyama, H.; Yoshida, K; and Kunii, D. : J. Chem. Engrng, Japan, 10,3, (1977), 224. 12lß.. Clevett, K. J. : 'Handbook of Process Stream . nalysis' , Ellis Horvrood, Chister, (1973). 125. Smith, C. R.; Free, D: Chem. Engr ; 202, (Oct. 1966), 236. 126. Redshaw, J. N. G. :I Chem E Symp. Series No. 50. P. 91. 127. Reid, R. C. ; Prausnitz, J. M.; Sherwood, T. K.: 'The Properties of Gases and Liquids', 3rd Ed. b: cGraww Hill, (1977). 128. Raldow, VI-. Y..; Wentworth, ' .. E. : Solar Energy, , (1979) , 75- 129. Linnhoff, B. ; Turner, J. A. : Chem. Engr. 361, (Oct. 1980), 621. 130. Dryden, J. G. C. : 'The Efficient Use of Energy', IPC/Dept. of Energy, (1975) , p. 167. APPENDIX 257 APPENDIX A THERMODYNAMIC AND TRANSPORT PROPERTIES. Al . Vapour Pressure Clayperon developed a mathematical relationship between vapour pressure of a pure substance and temperature: FT TV -TT Al. 1. where P= vapour pressure T= absolute temperature 7ý = heat of vaporisation at temperature T VG volume of gas and VL= volume of liquid. If the volume of liquid is negligible in relation to the gas volume and the ideal gas law is presumed to apply, the above relationship reduces to the Claus iu s-Clapeyron equation: p= a4a RT 2 A1.2 where R= gas law constant and X= molal heat of vaporisation. M Then, if the term x_ is assumed to be constant and independent of R temperature, Inp=A -T A1.3 A, B are constants. Obviously, equation A. 13 can be reliable only when the vapour pressure is relatively low, such that it obeys the ideal gas laws and when the liquid phase volume is very small. 258 Besides, the temperature range over which it is applicable will also be narrow in order for molal heat of vaporisation to remain constant. Antoine proposed a modification of this equation which enables it to be applied over a wider range of pressures and temperatures: . 1ý Inp=A B Al T+C where C is also a constant. It has been found that for most materials C is approximately - 430K, 51 and hence, Inp=A -T-B 43 A1.5 From data provided by Hougen, Watson and Ragatz51 , values of constants A and B for water would be 18.1 and 3883 respectively when vapour pressure is expressed in millimeters of mercury (mmHg) and temperature in °K. Hence vapour pressure of water may be expressed as: In p=18.1.3883 - T_3 A1.6 If T is in °C, then Inp=18.4- 3883 A1.7 T+230 This equation has been used to find the vapour pressure of moisture at different temperatures. 259 A2. Heat ofi Vaporisation The heat of vaporisation of a substance is a function of the temperature. As temperature is increased heat of vaporisation decreases. At the critical point there is no distinction between the liquid and vapour states, either in enthalpy or in physical properties, and heat of vaporisation therefore becomes zero. According to Trouton, the ratio of the molal heat of vaporisation Xm of a substance at its normal boiling point to the absolute temperature Tb is a constant. Thus. A Tm=KT A2.1 b KT is termed Trouton's ratio. For many substances, especially non- polar liquids, this ratio is approximately equal to 21, where the latent heat is expressed in cals/gram mole and temperature is in 0K. For polar liquids, values of this ratio tend to be muc: -i greater than 21. Watson extended the Trouton concept and produced a relationship 51 that would apply reasonably well to both polar and non-polar liquids. He found that the following empirical equation satisfactorily represents heats of vaporisation over a wide range of temperatures, . 38 ý2 ( X2.2 I- (1-Trl Try Tr2 are reduced temperatures at states I and 2 and A, 9 A2 are corresponding heats of vaporisation at states I and 2. If heat of vaporisation at a standard temperature, say normal boiling point, is known, then heat of vaporisation at any riven tem-erature can be readily obtained. 260 For water, with a normal boiling point of 100°C (373°K) and a corresponding heat of vaporisation of 2256 Y,. /Kg, as well as a 0 critical temperature of 647.3 K, then heat of vaporisation at any temperature would be: A . 38 = 2256 A2.3 This equation has been used to estimate the heat of vaporisation of moisture at different temperatures. A3. Heat of Adsorption Adsorption of a vapour on an adsorbent is accompanied by evolution of heat. If chemisorption takes place the heat evolved is usually of the order of the heat of reaction between the substances but if physical adsorption prevails, less heat, of the order of the heat of condensation, is evolved. In general however, the heat of physical adsorption is still measurably greater than that of normal condensation. This is believed to be mainly due to the surface interactions that take place during adsorption, resulting in the wetting of the adsorbent surface and the release of extra heat. The heat of wetting would depend on many interactive factors which include temperature, pressure, vapour phase concentration, adsorbent phase loading, and nature of adsorbate and adsorbent materials. Accurate generalized information on heats of adsorption is therefore very difficult to obtain and empirical deductions are often resorted to. Working on the adsorption of water vapour from air on activated alumina at a temperature of 25-300 C. Miller and Roberts 102 reported an average value of heat of adsorption of about 1.2 times the normal latent heat of condensation. Working at slightly higher temperatures, Free49 reported heats of adsorption in the range 2790-3370 KJ/kg (1200-1450 Btu/lb) for the same system. He also suggested that for adsorption of water vapour on silica-gel under comparable conditions the heat of adsorption is in the range 2326-2560 KJ/kg (1000-1100 Btu/lb) implying that the heat effect 261 associated with adsorption of water on activated alumina is greater than that associated with adsorption on silica-gel. He attributed this to the existence of greater attractive forces for water vapour in activated alumina, than in silica-gel. Free's findings agree with A the work of Rathmell and 106 , in which they proposed the following linear equations for correlating temperature effects: AT = 3500 oy for activated alumina. n3.1 OT = 2950 Ay-3.5 for silica-gel. A3.2 Clearly, this implies that under comparable conditions the heat evolved by adsorbing moisture on silica-gel would be less than that evolved by adsorbing on activated alumina. The average value of about 2150 KJ/kg suggested by Free for adsorption on silica-gel at about 40 0C, is about 1.03 times the heat of condensation at that temperature. Bearing in mind also the Miller and Roberts suggestion of 1.2 times for activated alumina, it was decided in this study to use a value of heat of adsorption of moisture on silica-gel 1.05 times the heat of condensation at any temperature. 7iatson's equation (A2.3) was used to evaluate the heat of condensation at the temperature. A. 1 Vibcosity From consideration of elementary kinetic theory of gases in momentum transfer, Reid et al 127 related the transport property (i. e. viscosity) to the molecular properties of a gas. Thus they derived the equation: U (constant) ` A4.1 0 262 For a rigid, non-interacting molecular sphere model the value of the constant was obtained thus: u= 26.69 TET A4.2 Q where u is viscosity in micropoise M is molecular weight T is temperature in °K a is hard sphere molecular diameter in 0 However, for most gases, intermolecular forces do exist between molecules. Chapman and Enskog127 considered this to arise mainly from collision between molecules and obtained a modified form of equation A4.2 thus : u= 26.69 rT .3 -q v where Qv is the collision integral. Qv will therefore be unity if the molecules do not attract each other and the Chapman-Enskog equation (A4.3) reduces to equation A4.2. In order to obtain the viscosity of any gas, 2v must therefore be known. At low pressures, Neufeld et al proposed an empirical equation which relates QV to a dimensionless temperature T* thus : AC +( E A4.3 y=(T ý) + (exp. DT*) exp. FT*) A, B, C, D, E, F are constants and T* is given by: T* _ ()T A4.4. The following values were suggested for the constants: A=1 . 16145 C=0.5 2+87 E=2.16178 B=0.14.874+ D=0.77320 F=2.43787 263 Equation A4.. 24. was however found to be suitable for only non- polar gases. For polar molecules Stockmayer127 suggested that the equation could be modified by addition of a polar parameter c" Then, 2 SZ = S2 0 . 26 V(P) V(np) + T* A4.5 where Q V(p) is collision integral for polar molecule and 0 V(np) is collision integral for non-polar molecule. For air, the following values obtain: Q=3.71 and 78.6 and for water vapour, Q=2.64.1 £= 809.1 k Also, for water, which is polar, 6=1 . 0. Using these values, viscosities of air and water vapour can then be obtained. The individual viscosities of air and water vapour were calculated by the above procedure. By considering moist air as a binary mixture of air and water vapour, Wilke' s approximation was then used to obtain its viscosity from the individual pure component viscosities. 6 u= y1 u1+ y2 u2 A4. m y1 + y2 ID 12 y2 + y1ID 21 264 where and A5. = viscosity of mixture (moist air) I'M VI , U2 = pure component viscosities (air. water vapour) y1 , y2 = mole fractions (air, water vapour) 1f (u ý112)2 (M MM ý42 1), 2; 12 2ý X21 X12 p2 1 ý1 ý2 Diffusivity A4.7 A4.8 The transport property associated with gaseous diffusion mass transfer, diffusivity, can be related to the molecular properties of a gas in a similar manner as viscosity. Thus, for binary diffusion of gases A and B, ,2 = 1.858 x 10-3 T3/2 [(MA + A1B) /MA MB] D AB A-5-1 Po AB where DAB = binary diffusion coefficient, cm 2/s T= temperature, °K P= pressure, atm a= hard sphere molecular diameter, A° QD = diffusion collision integral and Q AB aA +Q B 2 Defining B by, 1 CAB _ (A E B)2 and T* by k T* = (-----T :: b A5.2 A:. 3 265 then, SI D would be expressed as: Aý D= T* +e px DT* + exnFTý` + --G---E exp HT* where the constants are: A=1.06036 B=0.15610 C=0.19300 D=0.47635 E=1.03587 F=1.52996 G=1.76474. H=3.89411 15.4 With the same values of a and 9"k used in estimating viscosities of air and water vapour, the binary diffusion coefficient of the air-water vapour system was calculated by this procedure. 266 APPENDIX B St". 1ITLE Cl*-Lkdul---, --l-=i-. Oý Data Condensate flow = 20.5 1-iters/6 mins. Steam pressure = 65 psig. of wet fabrics = 117 lbs. Wt. of dry fabrics = 72 lbs. Ambient wet bulb temp. = 18°C. Ambient dry bult temp. = 25°C. Initial temperature of fabrics = 22°C Final temperature of fabrics = 1.0°C. TABLE B. 1 TLTILER =ýI III., .T ýü L 0L'T ,T Er F: RA: UPE S rains Tin C Tout C 0 90 50 1 104 60 2 101+ 67 3 104 71 4 104. 72.5 5 104 75 6 103 76 7 102 76.5 8 1c4 76.5 9 ýC1+ 77 10 104 76 11 104. 76.5 11.5 104. 76.5 12.5 106 78 13.5 l o6 77.5 ; 4.5 106 79 15.5 106 80 16.5 106 62.5 1ý 1o 85 267 Calculation Procedure The air inlet and outlet temperatures T. and Tout are plotted in 'ig. D1 and their mean values over the entire drying period graphically determined. Thus, T =1C4°C Tout= 75 °C. Moisture Extraction: - Moisture evaporated = 117-72 = 45 lbs. Drying time, 9= 17 rains. Moisture Extraction Rate = 45/17 lbs/min. = 1.20 kgs. /min. Steam Rate: - Condensate flow = 20.5 lits/6 wins. = 3.112 kgs/min. Flashing occurs as condensate expands through the valve on the outlet line. Therefore the condensate rate as measured is inaccurate and must be corrected by doing a flash calculation for the prevailing conditions. Y kg/min F k&// X kg/min Fig. B2 : Scheme for flash calc latior_ 268 0 00 O c-- C O CO \ID 00 J . -a o- 00 l* 1 t7-- nC+- CD ým ti 'S N" ct C- CTJ ('t CD 9 CD Pl C'* En 269 Let X= flowrate of condensate as measured Y. flash steam rate F actual condensate flov1rate. By mass balance, F=X+Y but X=3.1}2 B. 1 hence, F=3.42 +YB. 2 By heat balance, FHF= XHX + YH y B. 3 where H is enthalpy. At steam pressure 65 psig, HF = 272 Btu/lb. At atmospheric pressure, HX = 180 Btu/lb. and Hy = 1150.5 Btu/lb. therefore, F(272) = 3.4 2 (180) + Y(1150.5) B. 4 Solving B2 and B), F=3.76 kgs/min. = actual steam rate. Specific Steam Rate: - Steam rate = 3.78 kg/min. Moisture removal rate = 1.20 kg/min. hence, specific steam rate = 3.78/1.20 = 3.15 kg steanVkg moisture. Thermal Efficiency: - Later.. heat of condensation of steam A 2115 K /kg Latent heat of vaporisation of moisture m= 200 ý: 3/kg. Assuming complete transfer of heat on condensation of steam, rate of enera' consumF tion will be: - qý =aS F _ :. 7E x 2115 = 7995 }/min 270 Energy use for vaporisation of moisture 1: y<ä be: q 2m where M is moisture removal rate q2=2400x1.2 = 2880 KJ/min Therefore `ý thermal efficiency vril1 be `fiver. by 'n = q2 x 100 1 3 66c Air Flowrate :- Dry bulb temperature, Tdb = 250 C `diet bulb temperature Twb = 18°C For these conditions, Now, let air flowrate air specific water vapour air inlet humid i .: ý y in = 0.0120 = G1 kgs/min heat =C Pg specific heat =c Pv air humid heat will be given by ch = cr + Pv y B. 5 9 by heat balance on steam batteries, T _- o G Chin (Tin läß I Cli G=L. 6 m Chin L in_ Tdb ) 7995 1 . 0ý x 79 9- ý' r s/ min. 271 (iv) Loss Through Dryer Walls. This is given approximately by the difference between to ;, al energy consumption and the sum of the inputs into other components. Hence, loss through walls will be: 100-(36+59.5+1.2) =3.3% Ideal Efficiency: - The ideal thermal efficiency of the dryer will be given as: TI heat used for evaporation ideal heat used for evaporation + sensible heat lost in exit air. 36 x 100 "ideal 36 + 59.5 = 37.7 Alternatively, this can be obtained by the concepts of equations 2.48 and 2.54 in section 2.5. Mean inlet air temperature = 104. OC Mean outlet air temperature = 75 °C hence, 104-75 n ideal 1 01ý. -25 = 36. T% lower, r` .+ 16 j ýý xa" 7 Blower Performance (Typical figures) CZ 7 50 45 40 35 ý 30 moo 25 20 Gp 1S hnaA. 'vo I hýsrwýl v. ýrwý ice. hsarv VS- Fb- "& SOprM M 1SOOTm Il boo rpm - ; 20C 5 auster Exhauster Performance (Typical figures) v. u... vomm fw Nmn w... %» ww-o% N... fam. Ms is zs,, M%,. ý "srr is s 20 30 40 s0 6C 70 CfM $C 6c 10 v Note: - q, r F1 , Dv, f; aures oven are recresen. 3t' .t reýerreý tc n ake conoi :, -s a; ý. ýr-- ._ ýº- ra' , 'ý a: ^rf, sre. er Charac teris . ics 7-i s - ri Appendix C2 : Flo. - Calibration Chart ROTAMETER MANUFACTURING CO. LTD. CROYDON ENGLAND. PUBLICATION No. R P. 2427 3f 2 21 N W F_ H z W V Z W me w co 1 3 2 oý 0 - L T + I CALIBRATION CHART FOR AIR. METRIC X SERIES TUBE SIZE 65X ( DURALUMIN FLOAT TYPE A ( I I l SIZE 65 d= 65-DD u, = I4S"57 gm I pa 44 I y1 -1 ( I I ( I I A I I I d + 1 i l l A l l I I I -j I I I I I I I I I I I ( I I I I I I I I ( ( I ( ( I ý ý ý t soo 1000 1500 FLOW: LITRES/MINUTE AIR AT IST AND 760 mrr,. H`. Abs. W a F- H O J LL x W N vN ui D tu H W x O d' N W d W Vf x u ce b Z d 0 I- I V Z 0 I- Co u , 2-74 ýrT r PPElTi D -ý "-. ýý ^y . "^'1 ,. ý ý+ ýT S AMPL ;. TErP TxD I men Ita1 Data (Run 45) Air flowrate =1 .6 m3/min. Height of bed = 0.75 m. Dry bulb ambient. temperature = 1800C riet bulb ambient temperature = 11.5. °C TABLE Dl Ryý'ý D ^ýTr_ ( XP Pii 11.5 ) (mins) Ti (oc) ýiw (oc) T3 (°c) T4. (°c) 5 (°c) M`e (ový Tew (ocý 0 28.5 26 22 22 22 31 12 5 30 26 77.5 30 25.5 74+ 26 10 32.5 27.5 76.5 77.5 35 76.5 26.5 15 33.5 27 74.5 81 83.5 77.5 26.5 20 32 26.5 68.5 80.5 $1+ 77 26.5 25 32 27 63 79.5 83 77 26.5 30 31.5 26.5 60.5 77.5 82. 76.5 25.5 35 32 26.5 57.5 75 80 75 26.5 40 32 26.5 56 72 78.5 73.5 28.5 45 32 27 54.5 70 75.5 73.5 29.5 50 32 26 53 67.5 74. 71 31.5 60 33 26.51 50 63.5 70.5 68.5 33.5 70 33 27 4.6.5 61 68 64. 33.5 80 32.5 27 44 57.5 66.5 58.5 32.5 90 32.5 27 40 54.5 65 53 31 -' N 000 L (,, ) h Cr, 'El O CD c+ CL vi (7, -11 CD 0000 / H ý-3 H II !II! N -ý1 N V1 Lfl lJt It $-b O00 cD cD CD C* ct c+ ýT. R C c 275 0-3 OOOO , -. I V 0 01 y 0 N 0 O ý" a , -. - m O 1 0 CTI\ 0 0 Co O 0 0 N 0 276 CD o ºb W r - OO po pG0 Cý r ý, C C 0 cr\ 0 0 0 C- 277 Calculation The bed temperature profiles are plotted in fig. D1 , while the air inlet and outlet temperatures and humidities are plotted in fig. D2. From fig. D2, the average air inlet temperature and humidity T. and yi are : T. = 32°C yi = 0.020 kg/kg Also from fig D2, the average air exit temperature Te is: Te =7 5ýC Breakpoint was obtained for outlet humidity hence., AT =T e aYYi Breakpoint time 6b = 35 Air flowrate G=I . 6m ye = 0.0010 kg/kg T. =43°C I Ye = . 0190 kg/kg mins. nin =2 kgs/min. Hence, amount of moisture extracted up to breakpoint will then be: (ýQ b' =1 . 33 kgs. 720x fl x (. 235)2 x . 75 Mass of adsorbent M= 23.4 kg. Thus adsorptive capacity of bed at breakpoint will be: w", b= Mb s . 0566 kg/kg = 5.68; , 278 The partial pressure of moisture in inlet air ýs expressed as: 29= Y. P1 _ 18+29y. Now, II = 700 mm Hg hence , pi = 27.16 mm rig By Antoine equation, saturation vapour pressure of water' at the aver-_e bed temperature (or air exit temperature) will be: p= exp (18.1 - 3883 -) ob Te +230 = 289.70 Bed relative humidity at breakpoint Xb will therefore be: X Pi Pob = . 0819 = 8.19 279 DIX E 1. RE ý'C7103" : ý" ^- 1`'W -, The computer algorithm which r; as s, -, -,., ;ro.. -ýý v- - yý r_` ; he overall performance of the TSC heat pur..: dr er is ou , ire- beIo- : Start nput data Set exhaust humidity yRe. Calculate dryer conditions; saturation temperature and Call humidity Ts , ys Dryer inlet humidity y Design outlet tempera- ture and humidity TDe. yDe, recirculation ratio r D. Calculate adsorption conditions: Call bed relative humidity Xb Adsorption- breakpoint capacity J; b BRH outlet temperature T, e Mass of adsorbent M S. Call Desorption Calculate desorption conii ticr. s : equilibrium cap-city 'k'ýR exhaust tem:, erature T Re 280 ýFýý ý_ E cont' d. Call Evaluatic 281 SUBROUTI:, E DRYER DE, 3IGN and Calculate T De yDe from dryer air saturation equation. Calculate rD by moisture balance. .; e ýurr. ýUý`OU?.. T^ r' RF IO - BRH '-C L 283 SUBROUTINE DESORPTION Call Desorption Calculate T Re from heat and mass balance. Calculate VAR from equilibrium relations. Return 284 SUBROUTINE EVALUATION Call Evaluation Subroutine calculates the cycle thermal efficiency, heat pump maximum and actual COPs and thermo- dynamic efficiency. Re turn 285 ýF1 H, -t Late. Mý tý y-Y 1 l/ýf r A. ýýIr : '/r`iýii.! I ý1. ý ýý -r (! fff!! t'('? Fy(. - r(F i !_1 `º- V am' ill rll ki 1 ýºtjvi"i ýaý I1ýC ! ~, ýý ýý M" -ý i" j'z: ii ki, -0 ýý'ºi y t ý, - .fý: +. -ýJt" f, rý 14 1". '"`iß "! "i i. ' % ti' ! ii I ^ý+ KN! f1 t'"l" .Z? sýi_1 t. -- 0- LET ý J rte' 17* j7;, -ý äa i jýt I !-ii ý` iýll"1 }! j. l v fifI; 7 r- -i L' 0: T -T h L. ý I "! ! hr 1"*`i % Lt ~' 1 r'F . 1. i ' L ._ "" I. y i ý' ýýtýt ý- f: ! Lf ri-! '='ý' L`"5/l"±+ý. _ 1. fýJ ýfýr '" "yýh t. " i ' . L~ j" 3 r 'Y11 -- ir. ti'.. 't. " ý Ef'ý'iý "-ý-"I. 1 iý'titý tr` ý %"'i ý i% j'T NT' Y4 T ßö! ß _% 77 ` =E7 4 ' 0ý tl 'ß0 T I .. . 1 1 vR 1Nr . NT .{ 20 P 'I N Y3 = Y2 00 GEC VIL- ý''1 Lt ! -1 Lý 1, ß+i r' I" -I 1512 1514 l i5 rh r Ll . 152-2 "52.1 f f"+ c"ý 8 15(2 i133 1532 54 1542 11,544 X 546 +` 4(. ' 1ý5! 15 52 1 54 15Sä .: 1Gß 1 56 . . 5'G v74 1.; r 15 1X8'0 .., ZÖb _7 C, 610 T4 '. " ; "ý ; '. 0,1 P NT ýý ýFý `. : 11 ß'"C1 , o! 1 FF 7- , L- ;; - p-, ONi r ýI rTrp r*- "fý, R %f äs i ti; RF/ IS r. .-a rr+ {" ja4 ri "I L' rr r er.. i_ "... %! h' t_ u r. r, r- T ,t-1tt . -.. -; 7 RE! " T-?. '. `OLLO&II G S'JRSCFI I Fv FEN Et 1_F ? r. ýý iNC? Wr TNT: RE! '! 1 . 75 DRYER 'LT }! i 15 ! r REM - iý 4: ß. 1O RR INLET ?? E f'1 4 . "s AL'ti 0 ; 'RE R E%' jT REM 65 15 _N TN CONE' fN IN DRYER REM L" 15 RE'U-- T. i NL EJ REM 7 15 R EÜNTN EAUT REM Cl 15 AiR 5P HEAT REM C2 15 MRTER YAF'OR SP H RT REM C! 15 =1LICA GEL P HAT RE N 15 WATER, ' SP HEAT REM H 15 HUMID HEAT OF RIR RE! '! 50 15 RMST EN T F#, TN fn Ri 15 RECYCLE RATIO REP! Z" _" ARE RN I TE `AT ION i-'AR'I R±-;; 'L E REM J` _ARE RN I Tr RR TT ON f' /MMY FUNCTION REP! C; IS flR CYCLE CUP REM C6 IS P TUAL. CYCLE COP. REM M'' 15 THEFSttlfF`NAM %tý EFFCY REM NI 15 THERMAL t FFC Y RE IM1 ; 'i 15 MRS 5 OF ti 11 I CA GEL PEf' REN-1 L>3 15 INITIAL TEA1P OF Rt' URPTION PEP REM 1) 15 I NIT I PL TEMP OF DESORPTION Pr '. -P REP! E`5 ARE ENERGY TERMS IN ErFCY CALCULATION RE! '! L 15' LAtENT R PT REPM k9 15 CRITICAL TEMP OF ORTER REH TS I5 BT TEMP OF W TtR REM LS 15 LATENT HEAT PT TS R' }'ý IE VAP PRESS RT Pir_ ENT ý=E T URN it r ý'! % i7 '., C rrr 1 .6 287 2ü00 1! 10 020 20 30 2'35 ? Z! eC 2050 2080 2120 120 10 2 2160 3020 tiL h tiýfý+tt 3100 . `'?. ILS . 7120 : 140 7142,00 -s'. 4lýýýi3 4 145 4L5! 4ý 4 4L56 -A 4r 00 ý Ci 'r; º 1%1ý t'1-` ý`ý k r Ell .. ET Z2=0 V-25='TICI +Z2 Til-'ý'ýý'ttýalºý ýt: i-ýrý^ý%-ý; vFVr% Vr1-F'ý# i} 'ý'jCýcº"tlý-}'tTi-("ý'ý-ýýtt"ý =! ` jiP S'r) r '% / !. 'Z L` "ýý ý" i Lý ý ýý EL I& LJ 'F, t . . ý . 77 ýy: ý (' i`"N 1jr; 1º3 ý? . ý' ,"`: ý' ; ;. + -jý'j r' i Y" a M' j t"'; F'1 ') ; ` ý . IF J`. ýý JLý Tý l i1 i? 7' .1 ir=,, j4- tt? ýY'ýtý-ýý_S) 0- %I }x LET --, =2 . : _T4 Nk T" : '=K4 L4=rNf (K) TO S, 0 IF z+. ý: ' GOI +f=J4 ý'l=1rýý ýrºý cri V-7, - v45, ' j4 r'N Lit r :' VA ; -Ci *i'7.1- 7 'i > : il jil'. ,. Ii . 10 ., '.. ! -F I r! iI1 ; ''':: ý. : 7 288 5l tJ (? REM Lfj '1 ,, j7N~f _%: ti = . .1-, 0- ': = &d i-7. W ý- i- - w ig ~ *ý ý .^ý.. ' :L t"1 ': ýýr :_" ýý tJý' (" IT -?: l ti- rr .i lit j"] ýlr_ý .. ' y'F`'if"_`... 'ý`^ ... ' ý..:. " __ -_`" I " L. -`+=±: I. » -- ý. W" -s:; A' rý rý i p' 1I to 7ry C, sv 5: 110 tun = INK . a: 0 5200 jM `rid = i'"ý''ýtiL t` ý- Lt?: I

Energy recovery in drying by adsorption heat pumping

Abstract

Extracted data

Energy recovery in drying by adsorption heat pumping

Abstract

Extracted data

Related items

Related items