Predictive Function Control System Based On Temperature Of Electric Heating Furnace

Temperature control system is industrial process with a large range of working conditions,large time delay and nonlinear process,but the traditional controller is difficult to achieve the requirements of control,so we need to study new methods to solve industrial process with problems.Multi-model structure is constructed by a series of local models which can be reflected by a series of local characteristics,it can be used to solve the nonlinear problem caused by the large operating range.Predictive control is a new control algorithm,it according to the model to predict the trend of system changes in the future.It make the appropriate changes to the future trend,small calculated amount and that this control is suitable in the field of real-time control,robust performance and strong interference can well overcome the system,at the same time for the time lag link also showed good control performance.If we combine the advantages of both,it can promote the development of industrial control.This article mainly aims at the electric heating furnace temperature how fast and stable arrive to the set point of work,so we do the following aspects of research.First,the characteristics of large working range and nonlinear in electric heating furnace,the scope of work are divided a series of local models according to the temperature region,a local linear model based on the local scope of work,so as to convert a nonlinear system of local linear systems.According to the divided local linear model design the predictive functional controller work within the scope of the local linear model,and the Smith predictor is introduced to the controller design to reduce the time lag problem,finally when the temperature across the interval when different ideas for switching between controllers using multiple models.By using this method in the control of electric heating furnace,a better control effect is obtained.In the second point,when the local model is modeled,the model can be reflected the local characteristics,so a fractional order model is proposed.Through the Oustaloup numerical method for the fractional conversion of fractional order model for integer order model of higher order,thus the design of local controller model through the prediction function,finally determine the weighting coefficient of each model of the current error of each model,in order to get the control volume model of the.By using this method in electric heating furnace,analysis the experimental data,it shows that the multi model predictive function control can be very good to deal withthis kind of problem.