Case-based Reasoning In The Looper System Fault Diagnosis

Diagnosis technology is the necessary result of the development of manufacturing and technology. It can ensure the line's safe running, promote the product equality, avoid unnecessary repair, standard management etc. So it's important to research for large scale rolling line and even for all metallurgy industries.Looper system is an complex electro mechanic and hydraulic sub-system. Because of the faults' coupling, it's very difficult to set up the fault model. The general AI technology doesn't fit to this diagnosis job. Case-Based Reasoning is an analogism, and it needn't the fault model but the history case. It reasons based the past cases. It just fits to looper system. In CBR system, retrieve algorithm is a key to reasoning. The system uses the most nearest neighbor algorithm, it's efficient. This article also sets up a neural net to resolve the problem, and it gives the optimized arithmetic.Based the live data, it does some research on looper system and its faults. It classifies the faults and gets the relation between faults and data characters. It's the some base to diagnosis these faults. The rolling roll likes anamorphic because the high temperature environment. It gives the program process to judge the fault from data characters.CBR is a new AI technology, it is used in more and more regions. The diagnosis system uses this technology that fits the faults characters. In this system, some aspects need more research work, such as data collection, data analysis, the faults' mechanism and fault prediction. And this system is summered and prospected in the paper.