Research On Fuzzy Multiple Attribute Decision Making Method For Process Industry Fault Diagnosis

With the rapid development of information technology in industrial enterprises,it is difficult to avoid failures in the process industry.If there is a failure,whether it is personal safety,economic loss or environmental pollution,It will cause irreparable situations.Compared with Western countries,China's fault diagnosis technology may be later,although the equipment technology diagnosis symposium has been held one after another,but at present,the research on fuzzy multi-attribute decision-making methods for process industry fault diagnosis is still relatively small.In terms of theory and method research,or in applied research,they have not received the attention they deserve.With the development of artificial intelligence today,fault diagnosis technology has entered a new era.Fuzzy multi-attribute decision-making research and fault diagnosis direction is also a new topic,which has strong exploratory and cutting-edge.This paper mainly studies the fuzzy multi-attribute decision-making method for process industry fault diagnosis..Based on the turbine fault diagnosis and the process fault diagnosis of Tennessee Eastman(TE),this paper explores a new method for fuzzy multi-attribute decision-making in process industry fault diagnosis.The work of this paper is as follows:(1)The large number of fault characteristics obtained by modern measurement technology mainly include incomplete information,uncertain information or inconsistent information.Among them,according to the actual situation,the fault frequency characteristics have incomplete and uncertain information.Therefore,in order to consider the importance of each fault frequency element,based on fuzzy mathematics and pattern recognition theory,the traditional fuzzy closeness method is improved.And a fuzzy closeness method based on single-valued wisdom is proposed,which is applied to the fault diagnosis of steam turbine.(2)The traditional K-nearest neighbor(KNN)algorithm does not consider the relative relationship between sample features.It is necessary to calculate the distance to determine a neighbor,so that the classification speed is slow and the computational complexity is high.Aiming at the above problems,this paper introduces the related concepts in fuzzy mathematics—fuzzy membership degree and closeness method,andproposes a fuzzy K-nearest neighbor algorithm(CSFKNN)based on weighted chi-square distance metric,which improves the classification speed and accuracy of the algorithm.Get a better classification effect.In order to make this method better applied,the CSFKNN algorithm is applied to the fault diagnosis of the process industry.(3)In the fault diagnosis process of the process industry,since the fault sample data has multiple attributes,some attributes have no interrelated effect,which causes a multi-attribute group problem of the fault sample data.In order to solve this problem,this paper combines the VIKOR method and the fuzzy K-nearest neighbor method,and proposes a fault diagnosis method based on VIKOR method and fuzzy K-nearest neighbor method(VCSFKNN)to transform the multi-attribute group problem of fault sample data into fuzzy.The multi-attribute decision problem is applied to the field of fault diagnosis of the Tennessee Eastman processes.