Research On Combination Intelligent Model Of Hot Rolling Strip Width Based On Improved Fuzzy Particle Swarm Optimization Neural Network

In complex industrial processes, modeling is a necessary precondition and fun for industrial control and optimization, and an accurate model is very good for the control optimization of actual production process, therefore, the research on modeling of complex industrial processes has great significance for improving product quality and reducing production costs. Due to the nonlinear, time-varying, random disturbances of complex industrial process, uncertainties of model parameters, and other factors, the establishment of simple mechanism model which relies on traditional mathematical tools can not meet the modern complex industrial control requirements, a combination of mechanism model with intelligent algorithms has become the development trend of the modeling techniques.Supported by Natural Science Foundation of Shanxi Province (No.:2010011022-3), an algorithm based on improved fuzzy particle swarm optimization neural network was studied in this paper. Taking the typical complex industrial process of hot-rolling strip width control as the research subject, a simplified mechanism model of the strip width was established.Then, the combination intelligent model of hot-rolling strip width was established by using the improved particle swarm optimization neural network algorithm to determine the mechanism model parameters which are difficult to obtain. It will make a good foundation for the next the step width control.The main research work in this paper is as follows:(1)The background and significance of research projects were clarified.The currently main modeling methods of complex industrial processes were summarized by analysis of the research status of complex industrial process modeling.(2)Aiming to the easily "premature" shortcoming of particle swarm optimization, an improved fuzzy particle swarm optimization algorithm was proposed on the basis of further research on particle swarm optimization algorithm principle. The superiority of the algorithm was verified by using Matlab simulation to search the optimal extreme point of test function.(3)After the analysis of BP neural network topology and the training process, the optimization process of neural network based on improved fuzzy particle swarm was given which substitutes improved fuzzy particle swarm optimization algorithm for gradient descent to optimize the training process of BP network. The validity of the algorithm was verified by using Matlab simulation to approximate nonlinear functions.(4)After in-depth research on a hot-rolling strip production site, learning and mastering the principle of strip width control process, the mechanism model of strip width control process was established. The hybrid intelligent model of hot-rolling strip width was established by using the improved particle swarm optimization neural network algorithm to determine the three mechanism model parameters, such as finishing width spread, the vertical roll pressure correction factor, and the width correction factor, which are difficult to obtain. The feasibility and validity of using the combination of strip width mechanism model with intelligent algorithm to establish the strip width combination intelligent model were verified by Matlab simulation platform.