Network Control System For Aluminum Electrolytic Production Process Development And Research

With the development of control theory, computer and network technology, the control objects in the human's life and work turn gradually from simplicity to complexity, portion to integer, and one spot to many spots. Accordingly, the control system also is developing in the direction of network, digital and intelligence. Based on the study of Networked Control System (NCS) from theory and engineering angle, the thesis introduces a control system of a plant of aluminum electrolysis. At the same time, this thesis researches the neural network expert system combined with such a complex control object as aluminum reduction cells and introduces the material implementation of the neural network expert system in this control system.The thesis includes seven chapters. In the first chapter, the background of the project and NCS is introduced. The second Chapter researches the NCS from both aspects of the theory and implement of engineering. The emphasis is how to implement the NCS. Chapter three gives out the whole scheme of the NCS of aluminum electrolysis and the implement methods of the hardware. The forth chapter researches the control algorithm and its material implementation. Chapter five introduces the principle of the software system. In the next chapter, the problems in the debugging process of this system and how to resolve them are introduced. Chapter seven is the conclusion of this system.The main work and research fruits of the thesis include the following:(1) Gives out the definition, structure and characteristics of NCS. This will help us to understand NCS.(2) Analyzes the current study situation of several main problems that influence the stability of NCS: network delay, packet loss and multi-channels transmission and finds the merits and demerits of each method, and indicates the direction of further research.(3) Expounds on why Field-bus is more popular than Ethernet in control system, and emphasizes to analyze the CAN-bus in detail.(4) Researches neural network algorithm combined with this project,especially the BP algorithm with momentum item.(5) Designs the whole control system of a plant of aluminum electrolysis, including the structure of system, the type choice of hardware and the flat choice of software, etc.(6) Designs the communication software based on CAN-bus between monitor and controller.(7) In Delphi 6.0 environment, designs a management database system, based on Oracle database and BDE technology.(8) Designs a monitor interface with simple manipulation, convenient use, strong function and friendly man-machine communion etc.(9) Develops a NN expert system to diagnose the cell status. Previously, control effect is determined by the manipulator's experience, so different cell has different effect in the same series in the same plant. The neural network expert system will improve the actualities.The debugging and running results indicates that the system has reasonable control scheme, advanced control algorithm, good robust, high operating reliability, friendly man-machine interface and convenient manipulation. It entirely can satisfy the demand of client.