Details

Neural networks were proposed in the last century to start after years of development, neural networks have been prominent in a number of fields such as medicine, psychology, computer science, control and so on. The term modern engineering cybernetics was perfected by the American electrical engineering expert Norbert Wiener in 1948, and by the Chinese scientist Qian Xuesen (Tsien, H.S.) in 1954, and for the first time control was extended to the field of engineering and technology.And the concept of combining neural networks with engineering control can be traced back to the 1980s at the earliest. Prof Karl Johan Åström of the University of California, Berkeley, USA, and Prof Yutaka Yamamoto of the University of Tokyo, Japan, were among the early pioneers who developed the concept and carried out research in this area. This paper focuses on the analysis of the behaviour of the energy object, which is physically modelled as a steam boiler. Modelling is mainly divided into four parts: combustion chamber, steam bubble, heat conversion in the chamber furnace, valve. The simulation goal is how to quickly reach the preset stability value, energy simulation and analysis of the main Matlab simulink simulation, using the traditional Proportional, Integral Differential control, particle swarm optimisation Proportional, Integral Differential ( PID) control, comparing the results of Radial Basis Function Neural Network, Fuzzy Neural Network, Artificial Neural Network in sinmulink. Compare the results in terms of overshoot, amplitude, rise time, setting time, stable amount of steam produced.

• LIST OF ABBREVIATION
• INTRODUCTION
  • 1 Theoretical Basics
    • 1.1 Classification of steam boilers
    • 1.2 Physical Modelling of Steam Boilers
      • 1.2.1 Laplace variation transfer function
      • 1.2.2Model of combustion chamber
      • 1.2.3 Furnace heat transfer modeling
      • 1.2.4 Model of vapor system
      • 1.2.5 Turbine model
    • 1.3 Introduction to the software environment
  • 2 Traditional algorithms with simulink implementat
    • 2.1Conventional PID
    • 2.2Particle Swarm Optimization PID Algorithm
  • 3 Multiple neural network simulations
    • 3.1 Fuzzy logic based on Matlab fuzzy toolbox
      • 3.1.1 Fuzzy neural network tuning of PID parameter
      • 3.1.2 Structure of a fuzzy neural network
      • 3.1.3 Matlba fuzzy toolbox to build fuzzy rules
      • 3.1.4 Fuzzy neural network simulink implementation
    • 3.2 Radial basis function neural network
      • 3.2.1 Radial basis function
      • 3.2.3 RBF Neural Network Construction
      • 3.2.4 RBF neural network simulink
    • 3.3 Artificial neural network based BP optimizatio
      • 3.3.1 BP neural network forward excitation algorit
      • 3.3.2 Backward propagation to update the weight
      • 3.3.3 ANN Matlab implementation based on BP algori
• CONCLUTION
• REFERENCE LIST
• APPENDIX A
• (reference)
• PSO algorithm
• APPENDIX B
• (reference)
• Fuzzy logic rule
• APPENDIX C
• (reference)
• Radio basis s function of matlab