Identification And Correction Of Abnormal Load Data In Distribution Network

In the power system,the distribution network plays a crucial role in the entire power system as a key link directly connected to consumers.At the same time,the validity and reliability of the load data in the distribution network are essential for the distribution network and even the entire power grid.The accurate and reliable load data in the distribution network can meet the demand of load forecasting,power decision-making and scheduling production.Therefore,it is very important to identify and modify the abnormal load data of distribution network to improve the validity and reliability of load data in order to establish a safe,efficient and sustainable power system.The clustering method provides a good model to eliminate the abnormal load data,there are many studies on the application of various types of clustering algorithms in the identification and correction of abnormal load data.Among these researches,K-means clustering algorithm is a kind of more classic clustering method,which has the advantages of simple,easy to implement,fast convergence speed and so on,but also has the obvious shortcoming that it is sensitive to the randomly selected initial cluster centers and easy to obtain the local optimal solution.Based on in-depth exploration and research,this thesis presents an improved clustering method based on the improved firefly algorithm and K-means algorithm to identify and correct the abnormal load on the distribution network,it can take full advantage of the characteristics of firefly algorithm and clustering algorithm,to achieve complementary advantages.Firefly algorithm is another new type of intelligent optimization algorithm of bionic swarm generated after particle swarm optimization algorithm.The improved firefly algorithm,through studying and improving the location updating mode of standard firefly algorithm,can effectively avoid the problem of the solution set falling into the trap of local optimal solution possibility.The main idea of the optimization of clustering method is to use the improved firefly algorithm to detemine the best initial clustering center for K-means clustering,and to optimize the clustering results while having the advantages of simple and efficient K-means algorithm and fast calculation speed,thereby getting fast and superb clustering results.According to the best clustering result,the characteristic curve is extracted and the band-pass matrix is used to build the identification model of abnormal load data,and combine applications preprocessing and the geometric mean modification method can realize the good identification and correction of abnormal load data.In the analysis of actual examples,the optimized clustering algorithm in this thesis is compared with the experimental results based on the standard K-means algorithm and the algorithm combining particle swarm optimization algorithm with k-means algorithm.The experimental results show that the optimized clustering algorithm has better identification and correction effects for abnormal load identification and correction.