Applied AI Techniques in the Process Industry

From Molecular Design to Process Design and Optimization

Inbunden, Engelska, 2025

1 203 kr

Skickas . Fri frakt över 249 kr.

Beskrivning

Thorough discussion of data-driven and first principles models for energy-relevant systems and processes, approached through various in-depth case studies Applied AI Techniques in the Process Industry identifies and categorizes the various hybrid models available that integrate data-driven models for energy-relevant systems and processes with different forms of process knowledge and domain expertise. State-of-the-art techniques such as reduced-order modeling, sparse identification, and physics-informed neural networks are comprehensively summarized, along with their benefits, such as improved interpretability and predictive power. Numerous in-depth case studies regarding the covered models and methods for data-driven modeling, process optimization, and machine learning are presented, from screening high-performance ionic liquids and AI-assisted drug design to designing heat exchangers with physics-informed deep learning. Edited by two highly qualified academics and contributed to by a number of leading experts in the field, Applied AI Techniques in the Process Industry includes information on: Integration of observed data and reaction mechanisms in deep learning for designing sustainable glycolic acidMachine learning-aided rational screening of task-specific ionic liquids and AI for property modeling and solvent tailoringIntegration of incomplete prior knowledge into data-driven inferential sensor models under the variational Bayesian frameworkAI-aided high-throughput screening, optimistic design of MOF materials for adsorptive gas separation, and reduced-order modeling and optimization of cooling tower systemsSurrogate modeling for accelerating optimization of complex systems in chemical engineeringApplied AI Techniques in the Process Industry is an essential reference on the subject for process, chemical, and pharmaceutical engineers seeking to improve physical interpretability in data-driven models to enable usage that scales with a system and reduce inaccuracies and mismatch issues.

Produktinformation

Utgivningsdatum:2025-01-15
Mått:170 x 244 x 15 mm
Vikt:680 g
Format:Inbunden
Språk:Engelska
Antal sidor:336
Förlag:Wiley-VCH Verlag GmbH
ISBN:9783527353392

Utforska kategorier

Mer om författaren

Dr. Chang He is an Associate Professor in the School of Chemical Engineering and Technology, Sun Yat-Sen University. His research focuses on the multi-scale integration, design, optimization, and sustainability of the advanced energy systems. Dr. Jingzheng Ren is currently an Associate Professor at The Hong Kong Polytechnic University. He received the 2022 Asia-Pacific Economic Cooperation (APEC) Science Prize for Innovation, Research and Education (ASPIRE Prize).

Innehållsförteckning

Preface xi1 AI for Property Modeling, Solvent Tailoring, and Process Design 1Yuqiu Chen1.1 AI-Assisted Property Modeling 11.2 AI-assisted Solvent Tailoring 101.3 AI-Assisted Process Design 141.4 Conclusions 17References 192 Hunting for Better Aromatic Chemicals with AI Techniques 23Qilei Liu, Haitao Mao, Lu Wang, and Lei Zhang2.1 Introduction 232.2 Machine Learning-Based Odor Prediction Models 252.2.1 Odor Predictions for Pure Aromatic Chemicals Using Group-Based Machine Learning Method 252.2.1.1 Database Preparation 252.2.1.2 Molecular Representation 262.2.1.3 Model Architecture 272.2.1.4 Results and Discussions 272.2.2 Odor Prediction for Mixture Aromatic Chemicals Using σ-Profiles-Based Machine Learning Method 292.2.2.1 Database Preparation 292.2.2.2 Molecular Representation 312.2.2.3 Model Architecture 342.2.2.4 Results and Discussions 352.3 Computer-Aided Aroma Design (CAAD) Framework 362.3.1 CAAD for Pure Aromatic Chemicals 362.3.1.1 Identify Product Attributes 362.3.1.2 Convert Product Attributes to Properties and Their Constraints 382.3.1.3 Choose Property Prediction Model for Estimating Properties 382.3.1.4 Formulate MILP/MINLP Model 382.3.1.5 Solve the Model Using Decomposition-Based Algorithm 382.3.1.6 Verification 392.3.2 CAAD for Mixture Aromatic Chemicals 392.3.2.1 Identify Product Attributes 412.3.2.2 Convert Product Attributes into Properties and Corresponding Constraints 422.3.2.3 Establish Property Models 422.3.2.4 Ingredient Screening 422.3.2.5 Verification 432.4 Case Studies 442.4.1 Pure Aroma Design for Shampoo Additives 442.4.2 Pure Aroma Design for the Ingredient in Insect Repellent Spray 452.4.3 Mixture Aroma Design for Aroma Substitutes 482.4.4 Mixture Aroma Design for Odor Tuning 512.5 Conclusions 532.a The CAS Number of Molecules and the Selected Groups 542.b The Calculation Formula of Odor Score 562.c The Parameters and Results of the ANN Model 572.d The Designed Results of Molecules for Case Study 2 592.e Aroma Compounds for Ingredient Screening 61Acknowledgments 75References 753 Machine Learning-Aided Rational Screening of Task-Specific Ionic Liquids 79Ruofan Gu and Zhen Song3.1 Introduction 793.2 Molecule Representation of ILs 803.2.1 Groups or Fragment-Based Representation 813.2.2 COSMO-Derived Descriptors or Fingerprints 833.2.3 Machine-Learned Representations 833.3 Machine Learning-Based Structure–Property Models 843.3.1 Machine Learning for COSMO-Based Models 843.3.2 Machine Learning for UNIFAC Extensions 863.3.3 Pure Machine Learning Models 893.4 Applications 903.4.1 Computer-Aided IL Screening 903.4.2 Computer-Aided IL Design 933.4.3 Extensions to IL Mixtures 953.5 Conclusion and Perspectives 97References 994 Integration of Observed Data and Reaction Mechanisms in Deep Learning for Designing Sustainable Glycolic Acid 105Xin Zhou4.1 Introduction 1054.2 Methodology 1074.2.1 Database Generation 1094.2.2 Deep Learning 1114.2.2.1 Deep Neural Networks 1114.2.2.2 Deep Belief Networks 1124.2.2.3 Fully Connected Residual Networks 1144.2.2.4 Random Forest 1154.2.3 Optimization and Prediction 1164.2.4 Life Cycle Multidimensional Evaluation 1164.3 Results and Discussion 1174.3.1 Data Analysis and Statistics Before Modeling 1174.3.1.1 Analysis of Experimental Data 1174.3.1.2 Data Dependence Analysis 1174.3.2 Model Comparison and Feature Important Analysis 1184.3.2.1 Model Comparison 1184.3.2.2 Feature Important Analysis 1234.3.3 Performance and Feature Analysis of the Optimized FC-ResNet–GA Model 1254.3.4 Process Multi-objective Optimization and Experimental Verification 1264.3.5 LCSA Based on the Optimized Parameters 1284.3.5.1 Original Life Cycle Framework 1284.3.5.2 Life Cycle Inventory Analysis 1284.3.5.3 Life Cycle Sustainable Interpretation and Assessment 1294.4 Conclusion 1314.a Pareto Optimization Set 1324.b Experimental Data 1334.c Construction Method of Process Simulation Database Using Reaction Mechanism 1344.c.1 Elimination of the Diffusion Limitations 1364.c.2 Reaction Kinetics 138References 1395 Innovation of Gas Separation Processes: Integrating Computational MOF Design and Adsorption Process Optimization 145Xiang Zhang and Teng Zhou5.1 Introduction 1455.2 Step One: Descriptor Optimization 1475.2.1 Material–Property Relationship of MOFs 1485.2.1.1 MOF Representation 1485.2.1.2 Data-Driven Model for Single-Component Adsorption Isotherm 1495.2.1.3 Multicomponent Dual-Site Langmuir Isotherm Model 1505.2.2 Integrated Optimization of MOF Descriptors and PSA Operating Conditions 1515.2.2.1 Descriptor Design Space 1515.2.2.2 P/VSA Process Model 1525.2.2.3 Integrated Design Formulation 1535.2.3 Results 1545.2.3.1 Benchmark Process Using Cu-BTC 1545.2.3.2 Optimal MOF and Process from Integrated Design 1545.3 Step Two: MOF Matching 1575.3.1 Material–Property Relationship of MOFs 1575.3.1.1 Property–Performance Relationship for PE/PA Separation 1575.3.1.2 Validation with 471 CoRE MOFs 1595.3.2 From Computational MOF Design to Model-Based MOF Screening 1615.3.2.1 Identification of MOF Building Blocks 1615.3.2.2 In Silico Synthesis of Hypothetical MOFs 1625.3.2.3 MOF Screening via Validity and Feasibility Constraints 1645.3.2.4 MOF Screening via GCMC Simulations 1655.3.2.5 MOF Screening via PSA Process Optimization 1655.3.2.6 Optimal Results of SMOF- 1 1655.4 Conclusion 167References 1686 Reverse Design of Heat Exchange Systems Using Physics-Informed Machine Learning 173Chang He and Yunquan Chen6.1 Introduction 1736.2 PINN-Based Inverse Design Method 1766.2.1 Overview of Inverse Design 1766.2.1.1 Standard Physics-Informed Neural Networks 1776.2.1.2 Design Optimization and Decision-making Methods 1806.3 Example 1: Finned Heat Sink Model 1816.3.1 System Description and Objectives 1816.3.2 Improved PINN Structure 1856.3.3 Results 1856.4 Illustrative Example 2: Tubular Air Cooler Model 1916.4.1 System Description and Objectives 1916.4.2 Improved PINN Structure 1956.4.3 Transfer Learning 1966.4.4 Results 1996.5 Conclusion 204References 2057 Integrating Incomplete Prior Knowledge into Data-Driven Inferential Sensor Models Under Variational Bayesian Framework 211Zhichao Chen, Hao Wang, Yiran Ma, Cheng Qiu, Le Yao, Xinmin Zhang, and Zhihuan Song7.1 Introduction 2117.2 Literature Review 2137.2.1 Transport Process Scale 2137.2.2 Unit Operation Scale 2147.2.3 Overall Summary and Technical Gap 2147.3 Proposed Approach 2147.3.1 Loss Function Derivation 2157.3.2 Knowledge Representation 2167.3.2.1 Knowledge Description 2167.3.2.2 Knowledge Section via Self-Attention Mechanism 2167.3.2.3 Similarity of GCN and SAM 2177.3.2.4 Sampling from Posterior 2197.3.3 Model Expressions 2217.4 Experimental Results 2237.4.1 Evaluation Metrics 2247.4.2 Process Description 2247.4.3 Prior Knowledge Analysis 2257.4.4 Baseline Models 2277.4.5 Model Performance Comparisons 2287.4.6 Comparison with L1 and L2 Regularization Terms 2297.4.7 Sensitivity Analysis 2307.5 Conclusions 2307.A Experimental Settings 232References 2338 Data-Driven and Physics-Based Reduced-Order Modeling and Optimization of Cooling Tower Systems 239Chang He and Zhiqiang Wu8.1 Introduction 2398.2 Full-Scale Physical Model of Cooling Towers 2418.3 Bi-level Reduced-Order Models 2448.3.1 Design of Optimal Experiments 2458.3.2 Multi-sample CFD Simulations 2478.3.3 Model Reduction 2488.4 Thermodynamic Performance Indicators 2518.5 Optimization Model 2538.6 Illustrative Example 2548.7 Conclusion 261References 2619 AI-Aided High-Throughput Screening and Optimistic Design of MOF Materials for Adsorptive Gas Separation 265li Zhou, Min Cheng, Shihui Wang, and Xu Ji9.1 Introduction 2659.2 Methodology 2669.2.1 Molecular Level Simulation and Screening Driven by Rigorous Molecular Simulation and Machine Learning 2669.2.1.1 Molecular Characterization 2669.2.1.2 Structural/Chemical Analysis-Based Prescreening 2679.2.1.3 Diversity Analysis and Dataset Splitting 2689.2.1.4 Molecular Simulation and Performance Evaluation Metrics 2689.2.1.5 AI-Aided Quantitative Structure–Property Relationship Development and Rapid Screening 2699.2.1.6 Process Level Simulation 2709.2.1.7 Reverse Molecular Design 2709.3 Case Studies 2739.3.1 High-Throughput Screening of Metal–Organic Frameworks for Hydrogen Purification 2739.3.1.1 Prescreening 2739.3.1.2 Rapid Screening 2749.3.1.3 Rigorous Validation 2749.3.1.4 Structure–Property Relationship Analysis 2769.3.1.5 Investigation on Practical Factors 2779.4 Conclusions 283References 28310 Surrogate Modeling for Accelerating Optimization of Complex Systems in Chemical Engineering 287Jianzhao Zhou and Jingzheng Ren10.1 Introduction 28710.2 Surrogate Modeling Techniques 28910.2.1 Polynomial Regression (PR) 29010.2.2 Polynomial Chaos Expansion 29110.2.3 Kriging 29110.2.4 Radial Basis Functions (RBF) 29210.2.5 High-Dimensional Model Representation (HDMR) 29310.2.6 Decision Tree (DT) 29410.2.7 Support Vector Machine (SVM) 29510.2.8 Artificial Neural Network (ANN) 29610.3 Application of Surrogate Model in Optimization of Chemical Processes 29710.3.1 Reaction Engineering 29710.3.2 Separation Engineering 29910.3.3 Heat Exchange and Integration 30010.3.4 Process Design and Synthesis 30110.4 Conclusion 303Acknowledgment 303References 303Index 313