KDD-based rolling schedule model for plate mill

Zhichun Mu, Dangxian Ma, Shaobo Wang, and Changsong Wang Information Engineering School, University of Science and Technology Beijing, Beijing 100083, China Mechanical Engineering School, University of Science and Technology Beijing, Beijing 100083, China

摘 要：<正>A novel method for generating a rolling schedule is presented, which is fundamentally different from the existing ones. KDD (knowledge discovery in database) techniques are applied for discovering association rules between rolling parameters in a large database of rolling operation, and based on these rules, the schedule for the crucial last six finishing passes is generated. Operational evaluation shows that the schedule generated by the new method outperforms that generated by existing methods. It also shows how in this application the human’s domain knowledge is applied to speed up the KDD process and to ensure the validity of the knowledge discovered.

关键词：