A Sensor Fault Diagnosis Method Based on EEMD and XGBoost
Sensors play an important role in key fields such as industrial interconnection,and it is crucial to ensure the reliability of their data through fault diagnosis methods.This paper proposes a sen-sor fault diagnosis method based on EEMD and Extreme Gradient Boosting(XGBoost)to improve the accurate identification and timely processing ability of sensor faults.Voltage and current sensor data were collected,and a dataset containing normal and abnormal operating states was established using fault injection.EEMD was utilized for denoising and feature extraction of the data,followed by training with XGBoost to establish the fault diagnosis model.The constructed model was validated using a parti-tioned validation set.Experimental results indicate that the proposed method is significantly better than other methods,at least 11.95%higher than the F1-score of the suboptimal method.It can better capture the fault characteristics of sensor data,has high diagnostic accuracy and strong robustness,low consump-tion,strong practical value,and can effectively ensure the accuracy and security of data in industrial in-terconnection,glass intelligent manufacturing and other fields.
万方数据
维普
