Intelligent Controller Design And Experiment Study On Apc System Of Cold Strip Mill

In order to improve the accuracy of cold-rolled strip thickness, it is urgent to develophigh-performance cold rolling hydraulic automatic position control (APC) system. Withthe maturity of the rolling theory and the development of control technology, the artificialintelligence is applied to the mill thickness control system to improve productionefficiency which has important significance. Based on the deep understanding of coldrolling process, the paper selects hydraulic automatic position control system whichimpacts strip thickness heavily as the research object. Aiming at the question thatconventional PID controllers lack adaptability by human manual tuning controlparameters, the artificial intelligence control technology is introduced into APC system,then making current widely used PID controllers modified and having an in-depthsimulation and experimental study.First, introducing the principles and components of cold-rolled hydraulic APC systemin detail, then based on the basic equations of servo valves and hydraulic cylinders, andcombined with the transfer function of each part, nonlinear and simplified linemathematical models which are comprehensive and realistic of APC system areestablished respectively and a linear model analog circuit is also built.Secondly, starting from the linear APC system, a RBF neural network self-tuningPID controller is designed, and also the RBF learning algorithm which is introduced inmemory factor is improved, and the controller is used to make sumulation study. Thenthrough the software of FB Generator, normal and improved RBFNN both are packagedinto the adaptive CFC modules, and making experimental research on SIEMENS FM458platform.Again, aiming at the impact that the parameters and structure of RBF neural networkon its control performance, the improved shuffled frog leaping algorithm is used tooptimize both parameters and structure of RBF neural network fully, and the welloptimized network is applied for simulation and PLC experiment like above.Finally, starting from the nonlinear APC system, due to the PID controller is difficult to deal with complex nonlinear systems, a hierarchical fuzzy controller based on leastsquares support vector machine identification is designed. This controller can realizereal-time identification and on-line control, and it dose not rely on the controlled model.Then making it compare with the conventional fuzzy controller and having a simulationstudy.