Improvement Of FCM Clustering Algorithm And Application In GIS Cable Management System

Electric power industry is the basic resource of national economic development,which is related to the stable economic growth and people's daily life.As the transmission medium of the power system,the safe and efficient operation of the cable is an important guarantee for the national economy and the quality of life of the residents.The traditional cable management has a series of disadvantages,such as the low degree of digitalization,the incomplete original information,the inaccurate location of artificial census information,the blindness of cable planning,construction and repair and maintenance.The geographical information system(GIS)combined with maps,remote sensing and computer science,which has important applications in different scenarios,especially in power line management.Therefore,the scientific and effective establishment of cable management system based on GIS is very important to the cable management of power supply companies.This paper mainly discusses the new algorithm and new technology in the development and application of cable management system based on GIS.The main contents are as follows:(1)An improved clustering algorithm based on gray wolf optimizer and fuzzy C-means for clustering spatial data of cable is proposed.For the problems of fuzzy C-means(FCM)algorithm including easily affected by the initial clustering center and prone to premature convergence.Gray wolf optimizer(GWO)algorithm is combined with fuzzy C-means algorithm and a new fuzzy clustering algorithm(GWO-FCM)is proposed.In the proposed algorithm,the GWO algorithm is used an excellent global optimization to optimize the center of FCM,the effect of clustering is improved through simulation of wolf excellent search for prey behavior to find a set of optimal cluster centers.(2)A Hilbert R-tree index algorithm based on hybrid clustering to establish the index mechanism of cable spatial data is proposed.According to the Hilbert R-tree indexing algorithm of uneven distribution in the space object,the node is easy to cause the dead space and a large amount of overlap,especially in the overlap area of leaf nodes and other nodes increases,resulting in multiple queries,reducing the efficiency of retrieval.This paper presents a Hilbert R-tree index mechanism based on hybrid clustering,the hybrid clustering algorithm of gray wolf optimization combined with fuzzy clustering is integrated into Hilbert R-tree indexing.The specificmethod is to use Hilbert R-tree encoding of spatial object,and then use the hybrid clustering method to classify the data,to reduce the overlap of each node and access memory I/O times,so as to improve the retrieval efficiency.(3)A cable management system based on GIS is developed.In order to meet the scientific management of power cable and combined with the theoretical innovation proposed in this paper,a cable management system based on geographic information system technology is developed,and successfully applied in the State Grid Yiwu power supply company.The Hilbert R-tree indexing algorithm combined with hybrid clustering was applied to the system development.This paper introduces the construction target of the cable management system,the system structure design,function design,development environment and tools.The key technologies and the corresponding functions of the system are analyzed.