Research On Status Evaluation And Fault Preditcion Of Key Component Of Power Plant Based On Oil Monitoring Multi-parameters

Bearing is a kind of key component widely used in power plant,and its state evaluation and fault prediction technology has received great attention.At present,the analysis model based on oil parameters is mainly based on offline monitoring methods,such as spectral and ferrography analysis models.In recent years,the online monitoring method has also received more and more attention.This paper mainly studies the bearing condition evaluation and fault prediction technology based on online oil parameters,and carries out the whole life experiment of bearing components,and obtains online oil data for experimental verification.The main contents are as follows:(1)A kind of status evaluation method combining grey relational model and k-means clustering is proposed to analyze the bearing whole life experiment data,which is used to divide the bearing's wear status.Thus the relationship between online multi-parameters of oil and bearing wear state is established.(2)In order to achieve the trend prediction of oil parameters,an unequal spacing grey trend prediction model based on online multi-parameters was established.Through calculation and prediction of the monitoring parameters of oil in armored vehicles with unequal interval,it has been shown that the grey prediction model can be used to predict the trend of oil data.(3)Based on the health status evaluation results using the established health status evaluation model,the measurement data is divided into training set and test set using cross validation method.The bearing health state partition model based on support vector machine(SVM)is studied,and the residual life prediction of bearing is studied by applying support vector regression theory,and the prediction model based on support vector regression is established.(4)The bearing life test bed was constructed,and the bearing life test was carried out for a certain type of bearing.Multi-parameter monitoring data and working status data through the bearing whole life were obtained,which provided experimental data support for bearing status evaluation and fault prediction models and algorithms.