Quality analysis method for hot strip based on data mining

The hot rolling of strip is a multi-stage production process. There are about a hundred of control parameters and target parameters which are related to the quality of products directly in the multi-channel processes. It is the main development orientation for both the literal research and engineering practice to improve the quality of hot rolling products,by finding the information between the control parameters and target parameters,and making use of the information. It is investigated that combining data mining with the industrial features of hot rolling productions,the potentially,useful,ultimately understandable process knowledge can be extracted.The correspondence relationship between quality defects and control state can be got. Through a comprehensive analysis of control parameters weight vector and the set of high-risk relationships,the key control parameters can be find to improve. This method can reduce the economic losses,improve the production efficiency,and provide the scientific and accurate idea for the quality analysis of hot rolling strip production.