Exception Diagnosis Expert System Of The Sintering Process & Key Technologies

Expert system is a developing research field in artificial intelligence. Although many expert systems have been developed, play important roles and produce great economic and social benefits in all kinds of fields, there still have many problems to be solved both in theory and construction technology. New ideas and new technology are needed to improve the construction of expert system. Based on learning from successful cases of expert systems and studying the basic principles and methods of abnormal diagnosis, this paper presents the construction of an online diagnosing expert system SEDES (Sintering Exception Diagnosis Expert System) which can make accurate diagnosis of abnormal situations in sintering process in time and overcome the disadvantage of lacking experienced engineers in sintering production site.This paper discusses the research and development strategies of SEDES including the system organization model, knowledge acquisition method, exception diagnosis and system architecture. And then the paper presents the system design including function structure, technical architecture and information flow model. The paper mainly focuses on the discussions of the four key problems including the representation of abnormal diagnosis knowledge, organization and management of knowledge, mechanism of fuzzy chain reasoning and implementation of online expert system. To solve these problems, improved production rule is used to represent the knowledge and fuzzy operation and matching are used to make diagnosis and reasoning to deal with the uncertainty of abnormal diagnosis knowledge and insufficient exposure of abnormal situation. Different with traditional methods, SEDES adopts relational database to manage rules and make it convenient to organize and manage knowledge. JESS reasoning engine combined with struts and spring framework simplifies the development of the online expert system. Finally, a case study is described and the knowledge maintenance and fault diagnosis process are introduced in detail.SEDES has been successfully applied in monitoring and diagnosing of sintering process in 360m2 sintering machine in SGIS Songshan Co,.Ltd and its effectiveness has been proved through the comparison with the on-site diagnosis result of experts. The application shows the system has good operability and satisfying rate of accuracy.