The Research Of Generalized Predictive Control Applied In Typical Process Control Systems

This article is mainly to make a series of researches on some typical process control characteristics,which will has significance for the continuous development of complex industrial control strategy.These industrial processes involve the petroleum,chemical industry,metallurgy and so on in many fields of national economy.As we all know,the process control is always a significant field,which has the diversity of control objects including temperature,flow,pressure,component etc.and different strategies should be applied according to various controlled object.In addition,the factors such as nonlinear,time lag,and uncertainty and so on increase the difficulty to set up the system model precisely.Even if we can get the approximate mathematical model after linearization,which would lead to the contradiction intensified between the model accuracy and the traditional quality that depend on model precision.At the same time,the industrial control system is developing towards large scale,integration which results in the contradiction further growing.In the paper,generalized predictive control(GPC)algorithm is studied in the THJ-3 process control system device,which is able to simulate complex control system in industy,including the nonlinear high-order three-tank water system,boiler lagging coupling system.Previously,the basic principle of predictive model,roll optimization and feedback correction in the GPC will be given introduction detailedly.Besides,the effects of relevant parameters design for control systems and the simulation of the variables change will also be presented in the following content.The GPC algorithm based on T-S fuzzy model method is proposed in the three tank water system,in which the T-S fuzzy model serves as the predictive model and rolling optimization continuous operates online.By appling the fuzzy predictive model,GPC method can completely describe the dynamic characteristics of nonlinear plants precisely.In order to reduce the influence of control performance degradation due to model deviation,real-time measurement data is used for feedback correction.In the fuzzy model,the premise parameters are identified by subtractive fuzzy C-means clustering(SFCM)and orthogonal least squares method is utilized to determine the consequent parameters.The advantages of subtractive clustering and fuzzy C-means clustering(FCM)are combined to automatically determine the initial number of clusters in FCM.When the algorithm is used in three tank system the simulation shows that it can restrain the overshoot and resistant interference effectively and have good robustness.In the boiler coupling system,a structure of feedforward decoupling control based on GPC was proposed,but before that it is necessary to eliminate the impact of the dead band time delay in the multivariable time-delay system.The Smith predictive strategy is used to eliminate the hysteresis quality in the characteristic equation by introducing state variables feedback loop.The design of the Smith predictor closely related to the controlled object transfer function and feedback channel function,which will not affect the design of decoupling controller.Finally combining with feedforward decoupling structure and diagonal matrix decoupling theory,the generalized predictive feedforward decoupling algorithm is put forward.On this basis to make corresponding improvement will modify the decoupling effect.The simulation shows that the strategy can get satisfactory control performance.