Data Management and Synthetic System for Oil Monitoring

Yao Zhigang,Wang Weigang,Wu Tonghai,Mao Junhong,Xie Youbai

DOI: https://doi.org/10.3969/j.issn.0254-0150.2009.07.025

2009-01-01

Abstract:A two-stage data administration and intelligent analysis system,composed of on-line monitoring and off-line analysis modules,was established for equipment group life-cycle condition-based maintenance.On-line monitoring module was designed to store on-line data by a local database via internet and administrate the data automatically.A real-time trend was predicated from the on-line data using a GM(1,1) model.An intelligent wear state classification model was established using Support Vector Machine (SVM).On-line character indexes were input into this model to diagnose the abnormal and give an off-line analysis advice.The off-line data administration module was designed as an interface of input and maintenance of off-line analysis indexes (including physiochemical,spectrum,analytical and directing reading ferrograph indexes),and the oil quality can be evaluated by referring to a criterion.A final decision of conditional maintenance based monitoring can be made by integrating the results of on-line monitoring and off-line analysis.