Failure Rate Prediction Model of Substation Equipment Based on Weibull Distribution and Time Series Analysis

Jingjing Wang,Hui Yin

doi:10.1109/access.2019.2926159

Jingjing Wang, Hui Yin

Open Access

https://doi.org/10.1109/access.2019.2926159

Copy DOI

Abstract

The failure rate is an important indicator to assess the reliability of substation equipment, and the failure rate prediction is an effective way to master the operation status of substation equipment. According to the characteristics of failure rate data, such as failure period segmentation and stochastic variation, this paper proposes a combined prediction method by the data decomposition of failure rate time series and the segmentation theory of failure periods and then establishes a new failure rate prediction model of the substation equipment based on the Weibull distribution and time series analysis. Compared with the traditional Weibull distribution function model which cannot describe the stochastic variation of failure rate data and cannot identify the failure period automatically, the proposed model in this paper uses the minimum sum algorithm of residual squares obtained by the Weibull distribution to accurately identify the demarcation point between different failure periods and then establishes the autoregressive moving average (ARMA) model to achieve combined prediction. Finally, the effectiveness of the proposed model is verified by the engineering example and comparison with the traditional failure rate prediction methods.

Highlights

The electric power industry is a basic industry of national economy
The trend components establish segmentation prediction model based on Weibull distribution, and the stochastic components establish autoregressive moving average (ARMA) model by time series analysis theory
In view of the traditional Weibull distribution function model which can’t describe the stochastic variation of failure rate data and the traditional mathematical models which can’t identify failure periods, this paper proposes a failure rate prediction method based on data decomposition of failure rate time series and segmentation theory of failure periods, and establishes a new failure rate prediction model of substation equipment based on Weibull distribution and time series analysis

Summary

INTRODUCTION

The electric power industry is a basic industry of national economy. With large-scale construction and development of power grids, safe and reliable electric power supply has raised a higher requirement for electric power system. Ensuring the safety and reliability of substation equipments is an important prerequisite for safe operation of electric power system [2], [3]. The failure rate prediction, which is an effective way to master the operation status of substation equipment and the theoretical basis for determining state maintenance, can ensure safe operation of electric power system and reduce equipment. Combined with previous research results and the problem of the discontinuity in the predictive failure rate at demarcation point of different failure period, this paper proposes a new failure rate prediction method based on data. The trend components establish segmentation prediction model based on Weibull distribution, and the stochastic components establish ARMA model by time series analysis theory. The validity of proposed model in this paper is verified by the engineering example and model comparisons

PREDICTION METHOD AND PROCESS

PARAMETER ESTIMATION OF WEIBULL DISTRIBUTION FUNCTION

COMPARE WITH OTHER METHOD

Findings

CONCLUSION