Application And Research Of Data Mining In The Fault Diagnosis Of The Gear-box

In accordance with the widely application of databases and data warehouse in faultdiagnosis field, and many companies have installed the on-line monitoring system for someimportant equipments, and large scale data warehouses that store the operation status datahave formed. How mine out decision and knowledge that are useful to fault diagnosis fromthese abundant, disorderly and unsystematic data become the research emphases of this paper.This paper studied the data mining technology, on-line analysis processing technology andon-line analysis mining technology theory, combined the multi-dimensional data model-datacube with the association rules mining method, established the association rules miningmethod which based on the OLAP technology This paper also discussed OLAP technology,established the multidimensional data model-data cube, and combined it with the associatedrules mining method, applied the association rules model in gear-box fault diagnosis field tofind the rules exist in fault data, and offered decisive conclusions for diagnosis.Then researched the primary contents, introduced the distilled and analyzed method of thevibration character of familiar Gear-Box fault, put forward fault diagnosis ideology that basedon the data mining technology, introduced the structure, the working principle and process,researched the compose components of the mining model with the fault diagnosis ofGear-Box.Under the environment of the Windows XP operating system, a data mining model ofassociation rules is developed by using SQL Server 2005. Used the vibration test-bed's faultsimulation data to train and check the model, the results indicated that the rules which beproduced by associated rules model can offer decision for the fault diagnosis.