Transformer Fault Prediction Method Based on Multiple Linear Regression

2019 4th Asia Conference on Power and Electrical Engineering

能源学;电工学

Jiafeng, Qin^1 ; Chao, Zhou^1 ; Longlong, Li^1 ; Demeng, Bai^1 ; Wenjie, Zheng^1

State Grid Shandong Electric Power Research Institute, China^1

关键词: Correlation analysis;    Correlation relations;    Equipment fault diagnosis;    Fault characteristics;    Multiple linear regressions;    Occurrence probability;    Parameter prediction;    Prediction accuracy;   

Others  :  https://iopscience.iop.org/article/10.1088/1757-899X/486/1/012037/pdf
DOI  :  10.1088/1757-899X/486/1/012037

来源: IOP

【 摘 要 】

This paper mainly analyzes the transformer fault evolution rule. Meanwhile it analyzes the correlation relation between transmission and transformation equipment fault characteristic parameter and the fault. This paper establishes the fault characteristic parameter multi-factor forecast model and the equipment fault diagnosis model, which realizes the transformer fault occurrence probability, the fault type and the fault position real-time accurate forecast. It includes data acquisition and preprocessing, multi-factor prediction of fault characteristic parameters, correlation analysis of characteristic parameters and fault types, failure probability of various faults, equipment failure probability and fault diagnosis. In this paper, a transformer characteristic parameter prediction model based on environment and other factors is proposed. The example analysis of transformer characteristic parameter multi-factor prediction model based on multiple linear regression (MRL) shows that the multi-factor prediction model can effectively consider the influence of external factors on the variation of characteristic parameter, and the prediction accuracy is higher.