The authors' innovative research ideas in power plant control are presented in this book. This book focuses on 1) cognition and reconstruction of the temperature field; 2) intelligent setting and learning of power plants; 3) energy efficiency optimization and intelligent control for power plants, and so on, using historical power plant operation data and creative methods such as reconstruction of the combustion field, deep reinforcement learning, and networked collaborative control. It could help researchers, industrial engineers, and graduate students in the areas of signal detection, image processing, and control engineering.
Author(s): Chen Peng, Chuanliang Cheng, Ling Wang
Publisher: Springer
Year: 2022
Language: English
Pages: 210
City: Singapore
Preface
Acknowledgements
Contents
Acronyms
Part I Introduction and Preliminaries for Power Plant
1 Introduction
1.1 The Research Background
1.2 Research Status of Flame Detection System
1.3 Research Status of Flame Image Processing
1.4 Research Status of Temperature Field Reconstruction
1.5 Research Status of Optimal Control for the Coal-Fired Boiler-Turbine Power Plant
1.6 Main Contents of this Monograph
References
Part II Detection of Furnace Flame Image and Reconstruction of Temperature Field
2 Adaptive Mixed Edge Detection of Furnace Flame Image
2.1 Methods for Converting Color Image to Gray Image
2.1.1 Common Image Conversion Algorithms
2.1.2 New Gray Conversion Method
2.2 Image Preprocessing and Edge Computing
2.2.1 Preprocessing
2.2.2 Edge Computing
2.3 Adaptive Edge Selection Algorithm
2.4 Simulation and Results Analysis
2.4.1 Gray Image Conversion Experiment
2.4.2 Edge Detection Experiment
2.5 Conclusion
References
3 Intelligent Segmentation of Furnace Flame Image
3.1 Spatial Distribution Characteristics of Flame Image
3.2 Extraction Model of Flame Image
3.3 Optimal Segmentation Threshold
3.3.1 Flame Image Segmentation Threshold Expression
3.3.2 Optimal Segmentation Threshold Expression
3.3.3 Particle Swarm Optimization Algorithm
3.3.4 Improved PSO Algorithm
3.4 Simulation and Results Analysis
3.4.1 Verification of Improved PSO
3.4.2 Verification of Flame Identification
3.5 Conclusion
References
4 Reconstruction of Temperature Field Based on Limited Flame Image Information
4.1 Combustion Characteristics of Boiler System
4.2 Flame Temperature Measurement Algorithm Based on Digital Image Processing
4.2.1 Two-Color Method for Temperature Measurement
4.2.2 Single-Color Method for Temperature Measurement
4.2.3 Full-Color Method for Temperature Measurement
4.2.4 Two-Color Temperature Measurement Based on Digital Image Processing
4.3 Temperature Field Reconstruction Based on Least Square Method
4.4 Temperature Field Reconstruction Based on Intelligent Algorithm
4.5 Simulation and Results Analysis
4.5.1 Candle Flame Reconstruction
4.5.2 Furnace Flame Reconstruction
4.6 Conclusion
References
5 Furnace Temperature Prediction Based on Optimized Kernel Extreme Learning Machine
5.1 Prediction Model by Using Optimized Kernel Extreme Learning Machine
5.1.1 Optimized Kernel Extreme Learning Machine
5.1.2 Objective Function of the Optimized Kernel Extreme Learning Machine Prediction Model
5.2 Human Learning Optimization
5.2.1 Binary-Coded Human Learning Optimization Algorithm
5.2.2 Continuous Human Learning Optimization Algorithm
5.2.3 Hybrid-Coded Human Learning Optimization Algorithm with Reasoning Learning
5.3 Implementation of OKELM Based on HcHLORL Algorithm
5.4 Simulation and Results Analysis
5.5 Conclusion
References
Part III Modeling and Intelligent Control for Power Plant
6 Process Modeling of Power Plant
6.1 System Introduction
6.1.1 Operation Principle of Once-Through Boiler
6.1.2 Operating Principle and Characteristics of Intermediate Point Temperature
6.2 Particle Swarm Optimization Based Modeling Method
6.2.1 Introduction of PSO Algorithm
6.2.2 Mathematical Description of PSO Algorithm
6.2.3 PSO Algorithm with Inertia Weight Factor
6.3 Local Modelling Based on Improvement PSO Algorithm
6.3.1 Uniformization Distribution Mode in Initialization
6.3.2 Improvement of Inertia Weight Factor
6.3.3 An Improved Method for Solving Local Optimization
6.4 Model Fusion in Fuzzy PSO Algorithm
6.4.1 A K-Means Clustering Based Method to Reduce Nonlinearity
6.4.2 Fuzzy K-Means Network
6.5 Simulation and Results Analysis
6.5.1 Parameter Identification of IPSO
6.5.2 Global Modeling in Plant-Wide Operating Range
6.5.3 Dataset Testing
6.6 Conclusion
References
7 Fuzzy K-Means Network Based Generalized Predictive Control for Power Plant
7.1 System Description
7.2 Process Modeling Based on FKN
7.2.1 Framework of FKN
7.2.2 K-Means Clustering Network
7.2.3 Learning of FKN
7.3 FKN Based Generalized Predictive Control
7.3.1 Local GPC
7.3.2 Control Strategies of FKNGPC
7.4 Simulation and Results Analysis
7.4.1 Control Results over A Wide Range
7.4.2 Control Results Under Disturbance
7.5 Conclusion
References
8 Deep-Neural-Network Based Nonlinear Predictive Control for Power Plant
8.1 Long Short Term Memory Network
8.2 LSTM Based Internal Model Control of USC
8.2.1 Internal Model Control
8.2.2 Internal Model Control Based on LSTM
8.3 A Composite Weighted HLO Network Based GPC for USC Unit
8.3.1 Framework of CWHLO
8.3.2 Local Modeling
8.3.3 CWHLO Models Based GPC
8.4 Simulation and Results Analysis
8.4.1 Nonlinear Modeling by LSTM
8.4.2 GPC Based on CWHLO Models
8.5 Conclusion
References
9 Intelligent Virtual Reference Feedback Tuning Based Data Driven Control for Power Plant
9.1 System Description
9.2 Intelligent Virtual Reference Feedback Tuning
9.3 Design of Controller Based on IVRFT-AHLO
9.3.1 Adaptive Human Learning Optimization
9.3.2 Controller Based on IVRFT-AHLO
9.4 Simulation and Results Analysis
9.4.1 The Benchmark Problems
9.4.2 Control of The Pulverizing System
9.5 Conclusion
References
Index
