User Power Supply Reliability Assessment And Prediction In Low-Voltage Distribution System

The power supply reliability management is always an importantpart in measuring power quality in the electric power industry. Indeveloped countries, the power supply reliability management commonlycovers the entire power grid, however, in our country，it is usuallyimplemented in high-voltage and medium-voltage systems, leaving thelow-voltage system no such management. In recent years, because of therapid development in industrial and people’s daily lives, the societyrequires higher on the power supply reliability, especially in thelow-voltage distribution system which directly transfer the electric to theterminal users, thus our electric power industry began to perfect its powersupply reliability management, putting the low-voltage system powersupply reliability into the whole reliability management system.The power supply reliability management includes the reliabilitystatistics and prediction. To extend reliability management to thelow-voltage distribution system in our country will first face twoproblems: one is how to establish a reasonable reliability statistics indexsystem for the low-voltage system, the other is how to predict reliabilitylevel under the condition that the reliability parameters of someequipment are not available. Thus, this paper puts forward feasiblestatistical and forecast plans of low-voltage distribution system userpower supply reliability based on our current system automation level.Firstly, after compare and analysis the current reliability statisticsindex systems both at home and abroad, and regard with the presentsituation of low-voltage distribution system in our country, this paper brings about a new low-voltage system reliability statistics index system.This system gives the definitions of each indicators and calculationformulas, and clears boundary conditions. Then the paper comes up witha reliability prediction plan using the ANN algorithm. This plan relays noton equipment reliability parameters, but some basic data like customers’number, transmission line length as its parameters, and use thelow-voltage distribution system reliability statistical results in the pastfew years as samples to train the ANN in order to predicate the reliabilitylevel in the future. At last, through three examples this paperdemonstrates the feasibility and effectiveness of the reliability statisticsand prediction methods this paper puts forward.