Improved Time Series Model Predictive Control And Its Application In Thermal Process

In the current energy structure of our country,thermal power generation still accounts for a large proportion.With the continuous expansion of power plant scale,how to ensure that all links in the thermal process can operate with high quality,high economic efficiency,safety and reliability is the key to the healthy development of electric power industry.Control strategy is the focus of the research.PID is widely used in thermal process control,which has simple structure and high reliability.However,with the increasing complexity of industrial process,and during the process of power generation,the dynamic characteristics of controlled equipment would change with the load change.So,the system has nonlinear,time-varying,time-delay and other complex characteristics.It is difficult for traditional methods to achieve ideal control effect.In this thesis,the control difficulties of thermal process are fully considered,the predictive control algorithm of time series model based on Laguerre orthonormal basis optimization is proposed.Combining with the idea of algorithm fusion in the design,the advantages of traditional control strategy are used for algorithm improvement.The improved algorithms are applied to the typical thermal process control systems and the control effect is tested.The main work and innovation are shown as follows:(1)Combining the incremental fractional order PID algorithm with the multivariable Laguerre function model predictive control algorithm,a predictive control algorithm with the characteristics of fractional order PID is proposed.This algorithm improves the cost function of predictive control by referring to the fractional PID control structure.The proposed algorithm is applied to the simulation experiment of the load control of thermal power unit.The results show that compared with other algorithms,this method has better tracking performance and robustness.The overshoot is effectively suppressed and control quality of the system is improved.(2)The coefficients of autoregressive with exogenous input(ARX)model are expanded by Laguerre orthonormal basis and a predictive control method based on ARX-Laguerre model is proposed.PID and fractional order PID controllers are used to improve the cost function of receding horizon respectively.The improved algorithm is applied to the simulation experiment of superheated steam temperature control.The results show that two algorithms improve the control quality of the superheated steam temperature system in different degrees.(3)The autoregressive moving average with exogenous input(ARMAX)model is expanded by Laguerre orthonormal basis and a predictive control based on ARMAX-Laguerre model is proposed.Combined with PID control algorithm,the algorithm is applied to the simulation experiment of the main steam pressure control of CFB boiler,which realizes the accurate tracking of the target and improves the quickness and robustness of the system.For the problem of parameter selection,in order to solve the problems of heavy workload and inaccurate value caused by prior knowledge and artificial debugging.In this thesis,differential evolution algorithm is used to optimize the PID parameters,which makes the algorithm more intelligent in parameter setting.