Study On The Construction Of The Features Database Of Rotor Vibration Signals

In the development of the mechanical diagnosis and condition prediction, the role of the knowledge processing technology has been of more and more importance. Being with the progress monitoring tools and technology, the acquisition of adequate conditon data of the equipment or devices has become easier to achieve. However, it still seems to be in the infancy stage to process and analysis the quantities of monitoring data effectively and to discover the fault diagnosis rules reasonably. Thus, it has attracted more and more attention on use of database construction and the knowledge acquisition through monitoring data processing.A method is proposed in this paper for database construction of coupling the data mining method and database technology for fault diagnosis system, and study was conducted on its realization and application for fault knowledge discovery and preservation. Main work of this paper as follows:(1) The overall system design was designed based on the needs analysis of database construction for the rotor bearing system fault diagnosis and knowledge aquisition. Development envoirnment, operation platform and data mining method were selected to construct the system. Interactive access between the system and Labview program was designed and accomplished.(2) The data mining module and its functions were designed and completed in dedtail in Business Intelligence Development Studio(BIDS). The decision tree algorithm for data mining provided by SQL Server 2005 was adopted to intergrate with database system effectively, which built a data processing procedure for obtain fault diagnosis knoeledge automatically.(3) The database management sub-system was designed and realized to insure the smooth operation of the system. The main work includes the design of user permission management, data management, database preservation and recoverary, etc. ADO.NET was employed to achieve the users'frequent access effectively.(4) Experiments of faults simulation were designed and conducted on the self-built rotor bearing system test rig, data were collected and processed to extract the features for the system, which is used to verify the system. The results showed that the database system designed in this paper can realize the functions of datamining effectively and knowledge preservation. .Conclusions were made that the construction structure of the database system can promote the efficiency of data mining and the accuracy of fault diagnosis to achieve the goal of rules aquisition and protection. The method provided a new approach for the diagnosis knowledge acquisition.