Optimization in Chemical Engineering: Deterministic, Meta-Heuristic and Data-Driven Techniques

Description

Optimization is an area in constant evolution. The search for robust optimization techniques to deal with the highly non-convex models that represent the systems related to Chemical Engineering has led to important advances in the area. The need for developing economically feasible processes which are simultaneously environmentally friendly, safe, and controllable requires for adequate optimization strategies. Moreover, finding a global optimum is still a challenge for a diversity of cases. Thus, this book presents a compilation of classic and emerging optimization techniques, focusing on their application to systems related to the Chemical Engineering. The book shows the applications of classic mathematical programming, metaheuristic optimization methods and machine learning-based strategies. The analysis of the described techniques allows the reader identifying the advantages and disadvantages of each approach. Moreover, the book will discuss the perspectives for future developments on the area. Condenses the state-of-the-art in optimization methodologies. Discusses the challenges in the optimization of Chemical Engineering-related systems. Compiles the experience of highly recognized researchers in the area. Cover Half Title Also of interest Optimization in Chemical Engineering: Deterministic, Meta-Heuristic and Data-Driven Techniques Copyright Contents List of contributing authors 1. Optimization and its importance for chemical engineers: challenges, opportunities, and innovations Abstarct 1.1 Introduction 1.1.1 History of optimization 1.1.1.1 Pre-Christian times 1.1.1.2 Middle Ages 1.1.1.3 The Renaissance period 1.1.1.4 Seventeenth–nineteenth centuries 1.1.1.5 Twentieth century 1.1.1.6 Twenty-first century 1.2 Optimization techniques in chemical engineering 1.2.1 Deterministic optimization techniques 1.2.2 Metaheuristic optimization techniques 1.2.3 Problem definition and formulation 1.3 Relevance of optimization in chemical engineering 1.4 Challenges and opportunities 1.4.1 Challenges 1.4.2 Opportunities 1.5 Where are we going? 1.6 Conclusions References 2. Deterministic optimization of distillation processes Abstarct 2.1 Introduction 2.2 Superstructures 2.3 Deterministic optimization models for the design of distillation sequences 2.3.1 Distillation sequences with conventional columns 2.3.1.1 Distillation sequences with conventional columns: linear models 2.3.1.2 Distillation sequences with conventional columns: nonlinear models 2.3.2 Distillation sequences with nonconventional columns: thermally coupled distillation 2.3.3 Three-component systems 2.3.3.1 General thermally coupled sequences 2.4 Models for the rigorous optimization of distillation columns 2.5 Overview and final considerations References 3. Optimal design of process energy systems integrating sustainable considerations Abstarct 3.1 Introduction 3.2 Model approach 3.2.1 Heat exchanger network 3.2.2 Optimal selection of working fluids 3.2.3 Thermodynamic cycles and their interactions 3.2.4 Objective functions 3.3 Results and discussion 3.3.1 Case study 3.4 Conclusions Nomenclature Greek symbols Variables Sets Subscripts and superscripts References 4. Metaheuristics for the optimization of chemical processes Abstract 4.1 Metaheuristic optimization of chemical processes 4.1.1 Optimization in chemical engineering: an overview 4.1.2 Metaheuristic optimization: concepts and relevance 4.2 Theoretical foundations of metaheuristic optimization 4.2.1 Principles of metaheuristics: exploration and exploitation 4.2.2 Classification of metaheuristic algorithms 4.2.3 Mathematical formulation of metaheuristic techniques 4.2.4 Handling constraints in metaheuristic algorithms 4.2.5 Issues and challenges in metaheuristic algorithms 4.3 Overview of metaheuristic algorithms 4.3.1 Genetic algorithms 4.3.2 Particle swarm optimization 4.3.3 Simulated annealing 4.3.4 Ant colony optimization 4.3.5 Differential evolution 4.3.6 Tabu search 4.3.7 Firefly algorithm 4.4 Applications of metaheuristic optimization to chemical engineering 4.4.1 Parameter estimation 4.4.2 Reactor optimization 4.4.3 Separation process optimization 4.4.4 Heat exchanger networks 4.4.5 Optimization of chemical engineering systems with metaheuristic algorithms and neural networks 4.4.6 Metaheuristic algorithms in real-world problems 4.5 Conclusions References 5. Surrogate-based optimization techniques for process systems engineering Abstract 5.1 Introduction 5.1.1 Background 5.1.2 Unconstrained optimization formulation 5.1.3 Direct methods 5.1.4 Finite-difference methods 5.1.5 Model-based methods 5.1.6 Process systems engineering applications of model-based methods 5.2 Model-based derivative-free optimization 5.2.1 Workflow for model-based DFO 5.2.2 Local versus global surrogates 5.2.3 Bayesian optimization (BO) 5.2.3.1 BO: surrogate model 5.2.3.2 BO: surrogate optimization 5.2.3.3 BO algorithms 5.2.4 Ensemble Tree MOdel optimization tool (ENTMOOT) 5.2.4.1 ENTMOOT: surrogate model 5.2.4.2 ENTMOOT: surrogate optimization 5.2.5 Constrained optimization by linear approximation (COBYLA) 5.2.5.1 COBYLA: surrogate model 5.2.5.2 COBYLA: surrogate optimization 5.2.6 Constrained optimization by quadratic approximation (COBYQA) 5.2.6.1 COBYQA – surrogate model 5.2.6.2 COBYQA: surrogate optimization 5.2.7 Local search with quadratic models (LSQM) 5.2.7.1 LSQM: surrogate model 5.2.7.2 LSQM: surrogate optimization 5.2.8 Convex quadratic trust-region optimizer (CUATRO) 5.2.8.1 CUATRO: surrogate model 5.2.8.2 CUATRO: surrogate optimization 5.2.8.3 CUATRO: improving high-dimensional performance 5.2.9 Stable noisy optimization by branch and fit (SNOBFIT) 5.2.9.1 SNOBFIT: surrogate model 5.2.9.2 SNOBFIT: surrogate optimization 5.2.10 Dynamic coordinate search using response surface models (DYCORS) 5.2.10.1 DYCORS: surrogate model 5.2.10.2 DYCORS: surrogate optimization 5.3 Surrogate optimization methods performance assessment 5.3.1 Performance assessment procedure 5.3.2 Unconstrained performance assessments 5.3.2.1 Mathematical objective functions 5.3.2.2 Results: convergence plots 5.3.2.3 Results: trajectory plots 5.3.2.4 Results: tables 5.3.2.5 Results and discussion: mathematical unconstrained functions 5.4 Surrogate-based constrained optimization 5.4.1 Problem formulation for constrained surrogate-based optimization 5.4.2 Explicit constraint handling methods 5.4.3 Constrained surrogate optimization methods 5.4.3.1 Constraint handling in Bayesian optimization 5.4.4 Performance assessment of constrained surrogate-based optimization algorithms 5.4.4.1 Mathematical objective functions 5.4.4.2 Results – convergence plots 5.4.4.3 Results: trajectory plots 5.4.4.4 Results: tables 5.4.4.5 Results and discussion: mathematical constrained functions 5.5 Chemical engineering case studies 5.5.1 PID controller tuning problem 5.5.1.1 System definition 5.5.1.2 PID controller benchmarking results 5.5.2 Williams-Otto benchmark problem 5.5.2.1 Williams-Otto benchmarking results 5.6 Code 5.7 Concluding remarks 5.7.1 Performance comparison of algorithms 5.7.2 Benchmarking on mathematical functions 5.7.3 General observations and future work References 6. Data-driven techniques for optimal and sustainable process integration of chemical and manufacturing systems Abstract 6.1 Introduction 6.2 State of the art in process integration 6.2.1 Reviews on process integration 6.2.2 Data-driven techniques in PSE 6.3 Data-driven methods for optimal process integration 6.3.1 Integration of two decision levels 6.3.1.1 Integration of design and control 6.3.1.2 Integration of planning and scheduling 6.3.1.3 Integration of scheduling and control 6.3.2 Integration of three levels 6.4 Effectiveness of data-driven methods in addressing process integration challenges 6.5 Outlook References 7. Applications of Bayesian optimization in chemical engineering Abstract 7.1 Introduction 7.2 Bayesian optimization methodology 7.2.1 Main steps in Bayesian optimization 7.2.2 Limitations and challenges of BO 7.3 Potential applications in chemical engineering 7.4 Background and theory 7.4.1 Gaussian process models 7.4.1.1 Gaussian kernel function 7.4.1.2 Training a GP 7.4.1.3 Forecasting using a GP model 7.4.2 Acquisition function 7.4.2.1 Mathematical definition 7.4.2.2 EI computational procedure 7.4.2.3 Best next sampling value 7.5 Applications of BO in process system engineering 7.5.1 Binary distillation column 7.5.2 Petlyuk column 7.5.3 Methyl-methacrylate polymerization reactor dynamic-grade transition 7.6 General remarks 7.7 Conclusions References 8. Sensitivity assessment of multi-criteria decision-making methods in chemical engineering optimization applications Abstract 8.1 Introduction 8.2 Methodologies 8.2.1 MCDM methods 8.2.2 Weighting methods 8.2.3 DOM modifications 8.2.4 Sensitivity analysis metrics 8.3 Applications in chemical engineering 8.4 Results and discussion 8.4.1 Effect of modifications on weighting methods 8.4.2 Effect of modifications on rank 1 alternative 8.4.3 Effect of modifications on the top three alternatives 8.4.4 Effect of modifications on the ranking of all alternatives 8.4.5 Summary of effects of modifications 8.4.6 Further studies 8.5 Conclusions References 9. Hybrid optimization methodologies for the design of chemical processes Abstract 9.1 Introduction 9.2 Different strategies for seeking the global optimal solution 9.2.1 Gradient-based global optimization methods 9.2.2 Derivative-free optimization methods 9.2.2.1 Surrogate-based optimization strategies 9.2.2.2 Global search strategies (metaheuristics) 9.2.3 Summary of the different optimization methods 9.3 Hybrid and memetic optimization methods 9.4 Case studies 9.4.1 EA-based hybrid optimization of an entrainer-enhanced pressure-swing distillation process 9.4.2 Hybrid optimization of an energy-integrated extractive distillation with entrainer selection 9.4.3 Surrogate-assisted optimization of heat-integrated distillation sequence 9.5 Conclusion and outlook References 10. Optimization under uncertainty in process systems engineering Abstract 10.1 Introduction 10.2 Uncertainty analysis and propagation 10.2.1 Static uncertainties 10.2.2 Dynamic uncertainties 10.2.3 Stochastic modeling and sampling techniques 10.2.3.1 Monte Carlo sampling 10.2.3.2 Variance reduction techniques 10.2.3.3 Importance sampling 10.2.3.4 Stratified sampling 10.2.3.5 Quasi-Monte Carlo methods 10.2.3.6 Bayesian and adaptive methods 10.3 Optimization algorithms 10.3.1 Chance-constrained programming 10.3.2 Decomposition techniques 10.3.3 Sample approximation methods 10.3.4 Sampling accuracy and probabilistic methods 10.3.5 Parametric programming 10.3.6 Robust optimization 10.3.7 Stochastic optimal control 10.4 Applications 10.4.1 Chemical synthesis and computer-aided molecular design (CAMD) 10.4.2 Process synthesis and design 10.4.3 Energy and environmental problems 10.4.4 Planning, managing, and scheduling processes 10.4.5 Supply chain management 10.4.6 Reliability 10.4.7 Quality control 10.4.8 Optimal control 10.5 Conclusions References 11. Optimal control of batch processes in the continuous time domain Abstract 11.1 Introduction 11.2 Batch process control: challenges and formulations 11.2.1 Unconstrained optimal control problem 11.2.2 Constrained optimal control problem 11.3 Methods to solve optimal control problems 11.3.1 OCP-1: Fixed final time problem with final state fixed 11.3.2 OCP-2: Fixed final time problem with free final state 11.3.3 OCP-3: Free final time problem 11.3.4 Numerical solution 11.3.4.1 Indirect methods 11.3.4.2 Direct methods 11.4 Optimal control problems for linear systems 11.4.1 Linear quadratic regulator problem 11.4.2 Linear quadratic tracking problem 11.5 Applications of optimal control theory to batch processes 11.5.1 Batch transesterification control 11.5.2 Lignocellulosic biofuels and biochemicals 11.5.3 Hydrothermal liquefaction control 11.5.4 Novel applications in sustainability 11.6 Conclusion and emerging trends References 12. Supply chain optimization for chemical and biochemical processes Abstract 12.1 Introduction 12.2 Supply chain modeling 12.3 Defining the objective function 12.4 Data generation 12.5 Supply chain optimization 12.6 Case study: optimization of the supply chain for the valorizatio 12.7 Conclusions References 13. Future insights for optimization in chemical engineering Abstract 13.1 Introduction 13.2 Optimization algorithms 13.3 Data-based optimization 13.4 Optimization using novel hardware 13.4.1 Parallel computing 13.4.2 Graphical processing units 13.4.3 Quantum computing 13.5 Conclusions References Index