| China,with its vast territory and long coastline,is one of the countries most seriously affected by typhoon disasters in the world.As an extreme meteorological disaster with high intensity and relatively long duration,typhoons often cause a lot of damage to distribution network equipment with low design level in coastal areas,which in turn causes power outages to customers and seriously affects the safe and stable operation of distribution network and normal power consumption of customers.Therefore,it is important to carry out research on the prediction of distribution network damage and optimization of emergency repair and restoration under typhoon disaster to realize quantitative assessment of pre-disaster risk,formulate emergency repair and restoration decision strategy,and improve the level of disaster prevention and mitigation.In this paper,the prediction of distribution network damage mainly includes the prediction of the number of damaged distribution network towers and the prediction of distribution network outage grid.Firstly,this paper takes distribution network towers as the research object and carries out the research on the prediction of the number of damaged towers under typhoon disaster.This paper first divides the affected area into 1km×1km grids,and extracts multi-source heterogeneous data such as meteorological data,distribution network data and geographic data as the original sample data.Then,in order to improve the quality of sample data and ensure the performance of model prediction,the original sample data are pre-processed,including categorical variable processing,feature variable screening,standardization and other operations.Then,the number of damaged towers in the distribution network in the grid is used as the response variable,and the prediction model of the number of damaged towers in the distribution network under typhoon disaster is established based on the gradient boosting decision tree algorithm.Finally,the model was validated by selecting the damage data of Xuwen County,Guangdong Province under typhoon "Rammasun" and "Kalmaegi ",and comparing the prediction results with the actual results to verify the validity of the model.Secondly,this paper conducts a study on the prediction of distribution network outage grid under typhoon disaster.First,the meteorological data,distribution network data,and geographic data related to customer outages under typhoon disasters are extracted in a 1km×1km grid as the samples for distribution network outage grid prediction.Then,the data pre-processing work such as standardization,categorical variable processing,and construction of feature variables is performed on the sample data.Again,the Stacking integrated learning method is used to integrate eight algorithms,including random forest,gradient boosting decision tree,adaptive boosting,K-nearest neighbor,support vector machines,extra trees,decision tree,and extreme gradient boosting,to build a two-layer Stacking integrated learning model containing base learner layer and meta-learner layer for distribution network outage grid prediction under typhoon disaster.Finally,we use the data of typhoons "Rammasun","Kalmaegi" and "Mujigae" in Xuwen County,Guangdong Province as the research object,and carry out the grid prediction of distribution network outage,and visualize the prediction results.The feasibility and effectiveness of the constructed model in predicting outage grids are verified by comparing the actual distribution of outage grids.Finally,a distribution network emergency repair and restoration optimization study is carried out under typhoon disasters,and a distribution network emergency repair and restoration optimization model is proposed that includes two stages before and after the disaster.First,in the pre-disaster stage,the number of damaged towers and outage grids in the affected area are predicted to determine the resources required for post-disaster emergency response,and the probability of distribution network line failure is calculated based on the probability of distribution network line failure model under typhoon disaster.Secondly,in the post-disaster that is the second stage,the temporary dispatch center is used as the starting point to get the post-disaster emergency repair scenario using scenario generation and reduction technology,and the post-disaster emergency repair and recovery optimization model is established using mobile distributed power supply islanding and repair team collaborative repair and recovery optimization to get the post-disaster distribution network emergency repair and recovery optimization strategy.Finally,taking the IEEE-33 bus system as an example,it is verified through scheme comparison that the proposed method in this paper can effectively reduce the loss of load volume loss in the post-disaster repair phase,and also proves the necessity of setting up a temporary dispatch center.In summary,this paper presents a systematic theoretical study on the optimization of distribution network damage prediction and emergency repair recovery under typhoon disasters.The research results can provide theoretical guidance for the power sector to grasp the information of distribution network damage under typhoon disasters and formulate emergency strategies after disasters. |
