Zainuddin A. Manan - Modeling, Simulation, and Optimization of Supercritical and Subcritical Fluid Extraction Processes

1 Cover

2 Title Page

3 Copyright Page

4 Preface

5 Nomenclature

6 1 Fundamentals of Supercritical and Subcritical Fluid Extraction1.1 Introduction 1.2 Supercritical Fluid Properties 1.3 Subcritical Condition 1.4 Physical Properties of Subcritical Fluid 1.5 Principles of Sub‐ and Supercritical Extraction Process 1.6 Applications of SCF Extraction 1.7 Solubility of Solutes in SCFs 1.8 Solute–Solvent Compatibility 1.9 Solubility and Selectivity of Low‐Volatility Organic Compounds in SCFs 1.10 Method of Solubility Measurement 1.11 Determination of Solvent 1.12 Important Parameter Affecting Supercritical Extraction Process 1.13 Profile of Extraction Curves 1.14 Design and Scale Up

7 2 Modeling and Optimization Concept2.1 SFE Modeling 2.2 First Principle Modeling 2.3 Hybrid Modeling or Gray Box 2.4 ANN 2.5 Fuzzy Logic 2.6 Neuro Fuzzy 2.7 Optimization

8 3 Physical Properties of Palm Oil as Solute3.1 Introduction 3.2 Palm Oil Fruit 3.3 Palm Oil Physical and Chemical Properties 3.4 Vegetable Oil Refining 3.5 Conventional Palm Oil Refining Process 3.6 Conclusions

9 4 First Principle Supercritical and Subcritical Fluid Extraction Modeling4.1 Introduction 4.2 Phase Equilibrium Modeling 4.3 The Redlich–Kwong–Aspen Equation of State 4.4 Palm Oil System Characterization 4.5 Development of Aspen Plus ®Physical Property Database for Palm Oil Components 4.6 Binary Interaction Parameters Calculations 4.7 Supercritical Fluid Extraction Process Development Part II: Results and Discussion 4.8 Palm Oil Component Physical Properties 4.9 Regression of Interaction Parameters for the Palm Oil Components‐Supercritical CO 2Binary System 4.10 Phase Equilibrium Calculation for the Palm Oil –Supercritical CO 2System 4.11 Ternary System: CO 2– Triglycerides – Free Fatty Acids 4.12 Distribution Coefficients of Palm Oil Components 4.13 Separation Factor Between Palm Oil Components 4.14 Base Case Process Simulation 4.15 Conclusion

10 5 Application of Other Supercritical and Subcritical Modeling Techniques5.1 Mass Transfer, Correlation, ANN, and Neuro Fuzzy Modeling of Sub‐ and Supercritical Fluid Extraction Processes 5.2 Mass Transfer Model 5.3 ANN Modeling 5.4 Neuro Fuzzy Modeling 5.5 ANFIS and Gray‐box Modeling of Anise Seeds 5.6 White Box SFE Modeling of Anise 5.7 Results and Discussion 5.8 Introduction – Statistical versus ANN Modeling 5.9 Supercritical Carbon Dioxide Extraction of Q. infectoria Oil 5.10 Subcritical Ethanol Extraction of Java Tea Oil 5.11 SFE of Oil from Passion Fruit Seed

11 6 Experimental Design Concept and Notes on Sample Preparation and SFE Experiments6.1 Introduction 6.2 Experimental Design 6.3 Statistical Optimization 6.4 Optimization of Palm Oil Subcritical R134a Extraction 6.5 Comparison of Subcritical R134a and Supercritical CO 2Extraction of Palm Oil 6.6 Sample Pretreatment 6.7 New Trends in Pretreatment 6.8 Optimal Pretreatment

12 7 Supercritical and Subcritical Optimization7.1 Introduction 7.2 Evaluation of Separation Performance 7.3 Parameter Optimization of Palm Oil Deacidification Process 7.4 Proposed Flowsheet for Palm Oil Refining Process 7.5 Conclusions Part II: ANN, GA Statistical Optimization7.6 Introduction 7.7 Traditional Optimization 7.8 Nimbin Extraction Process Optimization 7.9 Genetic Algorithm for Mass Transfer Correlation Development 7.10 Optimizing Chamomile Extraction 7.11 Statistical and ANN Optimization 7.12 Conclusion

13 Appendix A: Calculation of the Composition for Palm Oil TG (Lim et al. 2003 )

14 Appendix B: Calculation of Distribution Coefficient and Separation Factor (Lim et al. 2003 )

15 Appendix C: Calculation of Palm Oil Solubility in Supercritical CO2 (Lim et al. 2003 )

16 References

17 Index

18 End User License Agreement

1 Chapter 1 Table 1.1 Comparison of the properties of supercritical CO 2and those of ord... Table 1.2 The advantages and disadvantages of solubility measurement using t... Table 1.3 Advantages and disadvantages of the most commonly used SCFs. Table 1.4 Various solvent as SFs with critical properties. Table 1.5 Physical properties of some common solvents (Corr 2002 ). Table 1.6 Dielectric constants and dipole moments for various solvents.

2 Chapter 2Table 2.1 List of algorithms.Table 2.2 Degree of membership for height example.

3 Chapter 3Table 3.1 Changes in fatty acid composition of lipid classes at different st...Table 3.2 Triglycerides composition (wt %) of Malaysian palm oil.Table 3.3 Typical fatty acid composition of Malaysian palm oil.Table 3.4 Contents of various components in the unsaponifiable fraction for ...Table 3.5 Methods of refining of crude vegetable oils.Table 3.6 Specifications for crude palm oil and refined palm oil.

4 Chapter 4Table 4.1 Composition of triglycerides in palm oil.Table 4.2 Phase equilibrium data of palm oil related component‐CO 2system.Table 4.3 Pure component parameters for the extended Antoine equation.Table 4.4 Predicted physical properties of pure components in palm oil.Table 4.5 Optimal parameters for the RKA model for the tripalmitin (1) – sup...Table 4.6 Optimal parameters for the RKA model for the triolein (1)– supercr...Table 4.7 Optimal parameters for the RKA model for the oleic acid (1) – supe...Table 4.8 Optimal parameters for the RKA model for the α‐tocopherol (1) – su...Table 4.9 Optimal parameters for the RKA model for the β‐carotene (1) – supe...Table 4.10 Temperature‐dependent polar factors and binary interaction parame...Table 4.11 Comparison of predicted phase equilibrium for a palm oil – CO 2sy...Table 4.12 Comparison of results obtained from simulation (this work) and ex...

5 Chapter 5Table 5.1 Operational condition for four experiments.Table 5.2 Available data of Re, Sc, and Sh numbers.Table 5.3 Statistical performance of a designed black box.Table 5.4 Statistical performance of all four models.Table 5.5 Levels of independent variables used in RSM design.Table 5.6 Box–Behnken experimental design and validation.Table 5.7 Statistical model parameters and validation.Table 5.8 ANOVA analysis for Eq. 5.14.Table 5.9 Parameters of the best ANN network during training process.Table 5.10 Comparison of the ANN prediction with RSM for training data.Table 5.11 Level of extraction process variables.Table 5.12 Box–Behnken design with coded/actual values and results.Table 5.13 Analysis of variance (ANOVA) of the response surface model.Table 5.14 Comparison of different ANN networks.Table 5.15 DOE for passion fruit seed oil extraction.Table 5.16 Levels of independent variables used in RSM design.Table 5.17 RSM model parameters and validation.Table 5.18 Parameters of the best network.Table 5.19 Comparison of the ANN with RSM for testing data.

6 Chapter 6Table 6.1 Treatment and coded levels of the independent variables for CCD.Table 6.2 Experimental design from RSM for the palm oil yield (g oil/g sampl...Table 6.3 Regression coefficient for palm oil yield.Table 6.4 ANOVA for quadratic model.Table 6.5 Characteristics of CO 2and R134a at several pressures and temperat...Table 6.6 Costs comparison between CO 2and R134a.

7 Chapter 7Table 7.1 Nonlinear inequality constraints for SFE palm oil refining.Table 7.2 Bounds for optimization variables.Table 7.3 Optimum operating conditions for simple countercurrent process.Table 7.4 Bounds for optimization variables.Table 7.5 Optimum operating conditions for countercurrent SFE with external ...Table 7.6 Composition of the product obtained from a countercurrent extracti...Table 7.7 Operating variables and constraints.Table 7.8 Optimum operating condition from sensitivity analysis.Table 7.9 Comparison of the ANN prediction with RSM for training data.Table 7.10 Experimental range and levels of independent variables.

8 Appendix BTable B.1 Multicomponent data transformed to a CO 2‐free basis.