Optimal Design Of Flatness Control System Based On Improved Cloud Reasoning Network And Simulation Of DSP

Iron and steel products are widely used in building,electrical appliances,ship,aerospace and other industries.Plate strip is one of the important iron and steel products.With the prosperous development of China's economy,many industries become strict on the quality of plate strip,and the high quality of the steel strip has a large market.Flatness as the key quality index of board strip production,which makes flatness control technology has become an important technology need to be solved at present.In recent years,the mature of artificial intelligence theory has open a new route for shape recognition,control,and it gradually applied to industrial rolling engineering.In this paper,it takes the improved cloud reasoning network in flatness control as the research subject,and the improved cloud reasoning network flatness control system optimized by cloud genetic algorithm(CGA)is designed.Finally,the DSP simulation implementation of flatness control system is completed.First of all,on the basis of the traditional cloud reasoning neural network,the improved cloud reasoning neural network is designed by introducing correction network.Taking a 900 HC reversible cold rolling mill as the application background,improved cloud reasoning network flatness recognition model is designed.The improved flatness recognition model has higher identification accuracy through contrast with traditional cloud reasoning network flatness recognition model.CGA is introduced based on genetic algorithm(GA).The flatness recognition model optimized by CGA has better recognition effect through a simulation comparison,at the same time the DSP simulation implementation of flatness recognition model is completed.Secondly,improved cloud reasoning network flatness predictive model is designed based on CGA.Taking a 900 HC reversible cold rolling mill as research object,the flatness in the first,third,fifth pass is predicted by improved cloud reasoning network flatness predictive model.The simulation results show that the prediction model can predict flatness output well,so the validity of the prediction model is verified.Finally,the controller is designed based on improved cloud reasoning network flatness predictive model and flatness recognition model,and then a complete improvedcloud reasoning network flatness control system is constructed.The parameter of flatness control system is optimized by CGA and GA separately.The simulation results show that the flatness control system optimized by CGA overcome the shortcoming of GA,such as easy to fall into local extrema,and has better control effect.The exported flatness more level off,and so it is an effective flatness control scheme.In order to connect with practical engineering better,the DSP simulation implementation of the control system is completed.The results show that the control system can run in the DSP,and the control results are consistent with simulation results.