Aid Decision Making Of Equipment Fault Diagnosis Based On Data Mining

The purpose of this paper is to establish the aid decision system of fault diagnosisthrough applying the data mining technology into equipment fault diagnosis which canimprove the speed of fault diagnosis and make the fault diagnosis system more practical.In this paper, I improved the traditional algorithm of association rules mining and designa aid decision-making system of fault diagnosis based on the new algorithm which canrapidly diagnosis equipment faults and early warn faults to stop the more serious faultand provide decision support for equipment maintenance to guarantee safety operation ofequipments and reduce the equipment downtime rate. According to the characteristics ofthe research, I used methods of the statistical analysis, the model analysis and theempirical analysis. We mainly researched from the following aspects.Firstly, I read a large number of documents and expounded the background and thesignificance of this research based on the reading. I analyzed the present situation oftheoretical in the fault diagnosis and the related technologies in data mining. I alsointroduced the basic knowledge of the fault diagnosis technology and the data miningtechnology which settle the theoretical foundation of this paper.Secondly, after I analyzed the association rules algorithm of data mining, I foundthat there are two shortcomings of the classical algorithm. When links, pruning andcalculating support, it need to scan the database constantly and the classical algorithmhas a high requirement for computer memory. But, it requires rapidly and accuratelylocated fault positions and found fault reasons. Besides, enterprises as non-profitorganizations have a demand of low cost. Appling the classical algorithm into faultdiagnosis can not meet the requirements of enterprises which need a new kind of systemwhich is fast mining and lower demand of computer memory.Thirdly, this paper proposed the Bitmap-base Association Rule Optimizationalgorithm (BARO) which is aiming at the shortcomings of the classical algorithm. Thenew algorithm transforms the database into a bitmap matrix to reduce the time ofscanning the database. Because the structure of the bitmap matrix is simple, the demand of the computer memory is reduced. These make the data mining technology moresuitable for practical application.Fourthly, after requirement analysis, module design and system implementation, anaid decision making system of fault diagnosis is established based on the Bitmap-baseAssociation Rule Optimization algorithm through demand analysis, module designed andsystem established. The system cans quickly mining the information related to equipmentfaults which can provide support for the equipment maintenance decision. This systemcan meet the needs of enterprises.