|
Circulating fluidized bed boiler, which has advantages of high efficiency flow, less pollution, fuel adaptability and so on, is the main direction of clean coal combustion technology. However, It’s control is very difficult, which on the one hand need keep a rapid response to the changes in external load, on the other hand maintains the temperature in the ideal range, It contains a set of multi-variable, strong coupling, non-linear and large time-delay characteristics. This paper has done some research about it. The main concepts and contributions are lies below:1) To solve the large delay characteristics’ influence, we research the nature and characteristics of the delay, establish a state observer according to the quadratic optimization principle, and achieve optimal control of time-delay systems by state feedback. Due to measurement error, model mismatch and other problems exist, we use neural network to design system controller by using a reference model.2) The designing of controller need refer to a optimal control law which is the only one determined by the order of system. In this paper, we choose ITAE function as a system’ performance indicators, an use the chaotic particle swarm algorithm to calculate optimal control law of common system.3) To solve the charged-coupled multi-variable characteristics’ influence, we choose different strategies to different control subsystems. For the main steam pressure control system, it needs a better dynamic performance. So we choose feed-forward compensation to decouple, by using feed-forward compensation to weaken coupling influence and using state feedback to eliminate the remaining part. For bed temperature control system, which needs keeping the temperature in the ideal range, we use internal model control to decouple, which looks the coupled information as disturbance, through the internal model principle to eliminate the disturbance.
|