Study On The Rolling Schedule Optimization And Model Self-Learning Of Five Tandem Cold Rolling Mills

In order to improve the strip quality ,there are two aspects must be paid attention: one is the reasonable settings of rolling parameters before rolling, which is called rolling schedule design, the other is the on-line control in the rolling process, which is called model adaptation and self-learning. As the rolling math models need adaptation and self-learning in the rolling process in order to get better ones which suit the rolling practice. The application of artificial intelligence in process control model research is a very effective way to promote the process control technology.Firstly, after studying the processing model and controlling model of the tandem cold rolling practice, the main math models for designing rolling schedule are analyzed. Several views to optimize rolling distribution strategies are researched, the ways'shortcomings to solve the rolling strategies are introduced.Secondly, different objective functions and constraint conditions are introduced for different requirements after the study of the characters of the cold rolling schedule. Further more, two optimization methods are proposed for optimizing the tandem cold rolling distribution. As a result, we optimized the two rolling schedule using adaptive genetic algorithm with different optimized objective functions , the experiment result showed adaptive genetic algorithm can overcome the limitation.Thirdly, considering the math models in tandem cold rolling are changing with the rolling time and rolling condition, the methods of math model adaptation and self-learning are proposed. After studying the two model adaptation and self-learning theory, the resistance of deformation and friction coefficient self-learning are researched, because they are the most important factors to the rolling force. Lastly, one kind of self-learning is used to improve the setting accuracy rolling force model.