Research On Inducing Radiant Energy Signal In Unit Load Control Based On Model Prediction(MPC)

Combustion system of power plant boiler is a system with non linear,time varying,and strong coupling.It is difficult to accurately describe by using a specific model.But the traditional thermal power unit commonly adopts the feedback control system which is composed of PID controller,this feedback control system needs better accuracy.Therefore,it is necessary to find a better combustion optimization control method for the current thermal power unit.In this work,a novel control idea which combines the rapidity of the radiant energy and the overall nature of the model predictive control method is put forward,and then applies to the load control system in power plant unit.For a 300 MW unit,the flame image is obtained by the flame detector arranged inside the furnace,and the radiant energy signal of the furnace is calculated according to digital image processing technology.For slagging,fouling and other problems in this unit,adopting the historical data during a period of time is to dynamically compensate and modify the current radiation signal.And then selecting the threshold for signal filtering is to improve the accuracy of the radiation signal.Based on the actual operating data,the load is divided into three load levels: high(90%),medium(75%)and low(60%).Then,the dynamic relationship between the radiant energy signal and the main operating parameters of the unit is obtained by segmentation identification.The stability and versatility of these models are analyzed,the results show that the model is validated.The stability of the system is verified by the unit step response curve by introducing the model predictive control into the load control system of the power plant unit.The radiant energy and the power are used as the input signal and the output signal,respectively.The original and new power are compared,and it demonstrates that the reliability.To illustrate the practicality of the system,we verify the system using the data of actual operating conditions.The results demonstrate the effect of introducing the model predictive control into the optimal control of radiant energy combustion.