Identification Modeling And Predictive Control Of Coal-fired Boiler Combustion System

There are millions of coal-fired boilers in China, which consume great quantity of energy, because of low efficiency and unadvanced automation level. With the coming of energy crisis, the government of China strengthens the supervision on coal-fired boiler for higher efficiency and low emission pollution. To improve automation on boiler is imperative, which will make widely the use of control system, and will make the high performance demand on control system also. However, an appropriate mathematics model on dynamic characteristic of boiler is the basis ground to design or optimize control system.This paper is mainly about identification modeling and generalized predictive control of coal-fired boiler. In this paper a grain coal-fired boiler was used as test bench .The paper analyzed the combustion system of the grain coal-fired boiler and reached the combustion model of coal-fired boiler(three input and three output model ) .Then according the characteristics of the combustion system of coal-fired boiler which are multi-input multi-output systems, the mathematical model of the combustion system of coal-fired boiler have been built by using multi-input multi-output system identification .In the modeling process, analyzing the multi-input multi-output system identification principles and methods and simplifying the multi-input multi-output system into several multi-input one-output systems which greatly simplifies the thermodynamic system model and the identification calculation. The later, with PRBS as input signal, makes use of least square algorithm to identify model parameters. In this part, PRBS parameters selection,data pre-process,algorithm implementation,model verification and so on are described in detail. This method can capture high frequency easily, and make less effect on running boiler as testing, so it is valuable to be generalized. The paper introduces the base theory of generalized predictive control, using the identified model of the combustion system of coal-fired boiler as the predicted model, giving the multi-variable generalized predictive control algorithm, and doing simulation runs with the actual number of the coal-fired boiler. At last the paper gives the adaptive generalized predictive control algorithm for coal-fired boiler which has slow time-varying characteristics.The mathematics model, provided in this paper, has definite guiding meaning for the control system design. The method of system identification modeling and the generalized predicted control described in this paper also benefit to other thermal process's parameter identification modeling and system control.