Abnormal Detection Method of Condition Monitoring Data of Pumped Storage Unit Based on IPSO-DBSCAN
The status monitoring data of pumped storage unit is affected by factors such as collection equipment fail-ures and communication equipment abnormalities,and there are some abnormal data in the dataset,which has a negative impact on the subsequent assessment and prediction of the health status of the units.Therefore,a unit anomaly data de-tection model based on improved particle swarm optimization algorithm and DBSCAN density clustering algorithm was proposed.The model improves the particle swarm optimization algorithm to address the problem of easily falling into lo-cal optima,and then introduced contour coefficient as fitness function to optimize the parameters of DBSCAN.Finally,the correlation coefficient was used to evaluate the effect of eliminating outliers.The measured guide vane opening,active power and lower frame vibration data of a domestic pumped storage unit from early February to late March 2020 were used for example analysis.The results show that the proposed method can effectively detect the abnormal data of vibra-tion monitoring,and the correlation coefficient between the data after removing the outlier is improved,which lays a data foundation for the subsequent unit health status assessment and prediction.
万方数据
维普
