Application of Machine Learning for Anaerobic Digestion 1st Ed

TABLE OF CONTENTS

PREFACE ii
TABLE OF CONTENTS iii

Chapter 1. Anaerobic Digestion 1
1.1 Introduction 1
1.2 Biochemistry and Microbiology 3
1.2.1 Hydrolysis - 3
1.2.2 Acidogenesis - 4
1.2.3 Acetogenesis 4
1.2.4 Methanogenesis 5
1.3 Process Control of Anaerobic Digestion 6

Chapter 2. Statistical Methods in Anaerobic Digestion 11
2.1 Multiple regression 11
2.1.1 Fundamentals of multiple regression 12
2.1.2 Response surface methodology 15
2.1.2.1 Design of experiments in AD 16
2.1.2.2 Central composite design 17
2.1.2.3 Full factorial and fractional factorial designs 22
2.1.3 Application of RSM in AD processes 26
2.2 Multivariate analysis techniques 29
2.2.1 Overview of multivariate methods in AD 29
2.2.2 Principle component analysis (PCA) 32
2.2.2.1 Mathematical foundation 33
2.2.2.2 Application of PCA in AD 34
2.2.2.3 Interpretation of PCA results 35
2.2.2.4 Interpretation of key parameters 36
2.2.3 Redundancy analysis - 39
2.2.3.1 Linking environmental parameters to microbial communities 41
2.2.4 Principle Coordinate analysis 48
2.2.4.1 Applications in sample comparison and clustering 51
2.2.5 Non-metric multidimensional scaling 58
2.2.5.1 Concept and application in microbial community analysis 61
2.2.6 Canonical Correspondence Analysis 68
2.2.6.1 Correlating microbial communities and process performance 71

Chapter 3. Kinetics of Microbial Growth and Substrate Utilization 78
3.1 General concepts of biokinetics in AD 78
3.1.1 Definition and importance 82
3.1.2 Microbial growth curve 83
3.2 Importance of Biokinetics in AD Process Contro 89
3.3 Fundamental Kinetic Models in AD 96
3.3.1 Monod Model 97
3.3.2 Haldane Model 99
3.3.3 Contois Model 100
3.3.4 Lotka-Volterra Model 102
3.3.4.1 Case studies: Interaction modeling 108
3.4 Fundamental biokinetic equations 111
3.4.1 Mass balance equation development 113
3.4.2 Analytical solutions 115
3.4.2.1 Solution for batch reactors 115
3.4.2.2 Solution for CSTRs 117
3.4.3 Numerical solutions 119
3.4.3.1 4th order Runge-Kutta Method 119
3.4.3.2 Solution for batch reactors 121
3.4.3.3 Solution for CSTRs - 123
3.5 Kinetic parameter estimation 125
3.6 Applications of biokinetic modeling in performance prediction 128

Chapter 4. Artificial Intelligence (AI)-based methods in Anaerobic Digestion 131
4.1 General Introduction to AI 131
4.1.1 Basics of AI - 134
4.1.2 AI Approach in Anaerobic Digestion 136
4.1.3 Common AI Tasks in Anaerobic Digestion systems 138
4.1.3.1 Prediction 140
4.1.3.2 Classification 141
4.1.3.3 Clustering 141
4.1.3.4 Anomaly Detection 142
4.2 Description of Model Architectures 144
4.2.1 Machine Learning Algorithms 145
4.2.1.1 Linear Regression 147
4.2.1.2 Support Vector Machine (SVM) 151
4.2.1.3 Naie Bayes 156
4.2.1.4 Decision Trees 161
4.2.1.5 Ensemble Learning Models 166
4.2.2 Neural Network Architectures 174
4.2.2.1 Multi-layer Perceptrons (MLP) 175
4.2.2.2 Convolutional Neural Networks (CNN) 182
4.2.2.3 Recurrent Neural Networks (RNN) 194
4.2.2.4 Transformer 206
4.3 Training and Validation of Models 219
4.3.1 Data Splitting 222
4.3.2 Learning Objectives 223
4.3.3 Model Optimization - 225
4.3.4 Validation and Evaluation of Models 229

Chapter 5. Application of AI in Anaerobic Digestion research 234
5.1 Physico-Chemical Data 234
5.1.1 Reactor Configuration 235
5.1.2 Substrate Characteristics 236
5.1.3 Effluent Characteristics 238
5.1.4 Gas Phase Data 239
5.1.5 Applications in AI Modeling 241
5.2 Qualitative and Quantitative Microbial Data 243
5.2.1 Qualitative Microbial Data 244
5.2.1.1 Community Structure 244
5.2.1.2 Diversity Indices 245
5.2.1.3 Applications in AI Modeling 246
5.2.2 Quantitative Microbial Data - 248
5.2.2.1 Microbial Quantification using qPCR 249
5.2.2.2 Applications in AI Modeling 249
5.2.3 Integration of Qualitative and Quantitative Data 250
5.3 Image Data 252
5.3.1 Spectroscopy Image 254
5.3.1.1 Application in Influent Analysis 254
5.3.1.2 Application in Effluent Analysis 254
5.3.1.3 Applications in AI Modeling 255
5.3.2 Microscopy Image 257
5.3.2.1 Biofilm and Granule Analysis 258
5.3.2.2 Microbial Community Imaging 258
5.3.2.3 Applications in AI Modeling 259
5.4 Time-series Data - 261
5.4.1 Characteristics of Time-series Data 262
5.4.2 Preprocessing Requirements for AI Applications 263
5.4.3 Applications in AI Modeling - 264

REFERENCES 266