Prediction of Power Outage Quantity of Distribution Network Users under Typhoon Disaster Based on Random Forest and Important Variables

Min Li,Ling Zhu,Xianqiang Li,Yong Huang,Hui Hou,Hao Geng,Jufang Yu,Xiao-Shun Zhang

doi:10.1155/2021/6682242

Abstract

Typhoons can have disastrous effects on power systems. They may lead to a large number of power outages for distribution network users. Therefore, this paper establishes a model to predict the power outage quantity of distribution network users under a typhoon disaster. Firstly, twenty-six explanatory variables (called global variables) covering meteorological factors, geographical factors, and power grid factors are considered as the input variables. On this basis, the correlation between each explanatory variable and response variable is analyzed. Secondly, we established a global variable model to predict the power outage quantity of distribution network users based on Random Forest (RF) algorithm. Then the importance of each explanatory variable is mined to extract the most important variables. To reduce the complexity of the model and ease the burden of data collection, eight variables are eventually selected as important variables. Afterward, we predict the power outage quantity of distribution network users again using the eight important variables. Thirdly, we compare the prediction accuracy of a model called the No-model that has been used before, Linear Regression (LR), Support Vector Regression (SVR), Decision Tree Regression (DTR), RF-global variable model, and RF-important variable model. Simulation results show that the RF-important variable model proposed in this paper has a better effect. Since fewer variables can save prediction time and make the model simplified, it is recommended to use the RF-important variable model.

Highlights

Typhoon disasters may lead to a large area of power outage for distribution network users. e prediction of the power outage quantity of distribution network users under a typhoon disaster can effectively improve the accuracy of disaster prevention and reduction
In the light of the aforesaid scenario, this paper proposed a prediction method of power outage quantity of distribution network users based on Random Forest (RF) algorithm
Ten-level wind circle radius Landing area research area), and wind level contributed little to the accuracy of the prediction model. erefore, this paper focuses on the analysis of the variables that contribute a lot to the prediction model, and analyzes their impact on the power outage quantity of distribution network users

Summary

Introduction

Typhoon disasters may lead to a large area of power outage for distribution network users. e prediction of the power outage quantity of distribution network users under a typhoon disaster can effectively improve the accuracy of disaster prevention and reduction. Some scholars have successfully used the data-driven method to assess the risk to power systems under typhoon disasters Statistical learning models, such as linear model, are firstly applied to evaluate the power outage in hurricane weather in [4]. En, based on geographical grid division, the negative binomial regression model was used to predict the power outages quantity of distribution network users under Hurricane [17]. (3) To accelerate the evaluation efficiency under disasters, the importance of each explanatory variable is mined in this paper On this basis, we extract eight most important variables to establish a novel RFimportant variable model to predict the power outages quantity of distribution network users. E explanatory variables include meteorological factors (such as maximum wind speed, wind direction, rainfall, etc.), geographical factors (such as altitude, slope, underlay type, etc.), and power grid factors

Evaluation index

RF-Global Variable Modeling and Analysis

Analysis of Modeling Important Variables

Conclusion