FAILURE DIAGNOSTICS WITH SVM IN MACHINE MAINTENANCE ENGINEERING

• Main Authors: DEÁK Krisztián, KOCSIS Imre, VÁMOSI Attila, KEVICZKI Zoltán
• Format: Article
• Language: English
• Published: Editura Universităţii din Oradea 2014-05-01
• Series: Annals of the Oradea University: Fascicle Management and Technological Engineering
• Subjects: Keywords:Bearingtest-rig; failurediagnostics; machinefault; machinelearning; supportvectormachine
• Online Access: http://imtuoradea.ro/auo.fmte/files-2014-v1/Deak Krisztian-FAILURE DIAGNOSTICS WITH SVM IN MACHINE MAINTENANCE ENGINEERING.pdf

Failure diagnostics as a part of condition monitoring (CM) technique is inevitable in modern industrial practice. Condition Based Maintenance (CBM) identifies all problems that cause further failures and suggests maintenance periods. Reducing maintenance costs and enhancing system availability are largely depends on information provided by precise and accurate failure diagnostics. The approach can be used widely in the several field of the industry. Data acquisition is related to measurement then data processing, feature extraction is needed, finally failure identification. In this paper Support Vector Machine (SVM) is discussed how to be used for diagnosing machines and machine elements. The aim of using SVM is to diagnose the system at a certain moment or predict its actual state in the future. SVM is progressing rapidly several new advances are revealed as the part of machine learning techniques. Due to experiments SVM efficiency could be approximately 90% or even higher.