Application Research Of Artificial Neural Network And Genetic Algorithms In Medium Plate Mill

With the rapid development of the national economy, the production of medium plate growth steadily, it is an increasingly important role in economic development necessary to become an important steel materials, but also led the development of rolling technology. The process control of modern medium plate mill is very complex, with multiple variables, nonlinear and strong coupling features, so that the traditional control model can't adapt the needs of modern rolling technology. Along with the artificial neural network and genetic algorithm artificial intelligence method applying in rolling fields, it is a new method to solve the problem by using of artificial neural networks and genetic algorithms for the prediction of rolling parameters.The influencing factors of rolling parameters in medium plate rolling process are analyzed. Combined with mathematical models, the control models of 4200 medium plate rolling process are researched by using artificial neural networks and genetic algorithms artificial intelligence methods. The influencing factors of rolling temperature in medium plate rolling process are analyzed. The GRNN neural network prediction model of rolling temperature is established by combining of GRNN and mathematical models. And compared with prediction results of traditional BP network, the result indicates that GRNN network is simple, high accuracy and generalization capacity etc to be used for prediction of medium plate rolling temperature. Against to defects of BP network which is easy to fall into local minimum, slow learning and the learning process easy to cause oscillation, neural networks and genetic algorithms are fused by analyzing their characteristics. The genetic neural network prediction model of medium plate carbon tool steel flow stress is established according to influencing factors of flow stress in medium plate rolling process. Through compared with prediction results of traditional BP and GRNN network, the result indicates that genetic neural network is the best performance to further improve prediction accuracy and generalization ability. The genetic neural network model can be used for prediction of medium plate medium plate carbon tool steel flow stress and providing basis for the development of rolling schedule.The Studies showed that the method of combined with artificial neural networks and genetic algorithm for forecast medium plate rolling process control models can further improve models prediction accuracy, and has broad application prospects.