| This dissertation is aimed at a 200MW coal-fired boiler which have been collocated the radiative detection equipments. The super-heater steam temperature transfer function of the coal-fired boiler is the large inertia, large delaying and more disturbances. Moreover, the dynamic changed with the load, which has been a difficult problem that troubles the controllers.The conventional PID controller can not achieve effective control. Fuzzy-PID control has many good characteristics, such as easy to design, good tracking ability, powerful robustness and eliminate the unknown disturbance etc. However, neural network can bring good performances when the control led objects vary in a big range. This dissertation combines the two control methods and presents a new fuzzy-PID control based on neural network model in order to overcome the large delay and uncertainty of the main stream temperature object in coal-fired power station.According to the feature of the large delay, the large lag and the various disturbances for the super-heater steam temperature, in this dissertation, Adaline neural network and BP neural network discrimination and modeling principle was applied to set up mathematical model of super-heater steam temperature in different load of the unit by analyzing the data from the working boiler. Then I have put the parameters into an encapsulation which can be applied in the control emulator.Based on radiative energy signal and above model, the dissertation proposes the fuzzy logic of fuzzy PID controller and applies it to the control system of super-heater steam temperature and carries on the simulation of it. Compared it with a normal PID control system, it was shown that fuzzy-PID control system is more effective than the common PID control system in the disturbance of radiant energy signal.Finally, the summary of the overall work, some advice on further investigation and modified measurements are given. |
