A Knowledge Graph-based Reliability Assessment Method for Overhead Distribution Lines

Overhead distribution lines are the key to safe and reliable power supply due to their large scale, wide distribution, various types and complex operating environment. Aiming at the problems of complex data analysis and difficult reliability assessment of overhead distribution line operation and maintenance text, this paper makes full use of multi-source heterogeneous data such as line operation and maintenance logs and equipment ledger to construct overhead distribution line knowledge graph. Based on the fault mechanism and correlation analysis of individual line equipment, the equipment time-varying failure rate model and the overall reliability evaluation model of overhead distribution line based on dynamic Bayesian network are constructed. Firstly, according to the different structural characteristics of multi-source heterogeneous data, data association rules are established to construct the pattern layer, and then entities are extracted from the text data to construct the data layer to form the overhead distribution line knowledge graph. Secondly, based on the knowledge graph, repetitive power failure factors are analyzed, and the correction coefficients and weights of major power failure factors are calculated, and the real-time failure rate model of overhead distribution equipment is constructed. On this basis, the dynamic Bayesian network model is used to establish the overall reliability evaluation model of overhead distribution lines, and simulate the fault evolution process of multi-state equipment. Finally, an example is given based on the actual data of a county distribution network in Zhejiang Province. The results show that the proposed method can effectively evaluate the operational reliability of overhead distribution lines, improve the level of digital operation and maintenance management of overhead distribution lines, and ensure the safe and stable operation of power grids.