Reliability calculation of transformer based on GA optimized BP

PANG Er-jun1，2， YU Hong3， TANG Gui-ji2， GAO Kuo1，2
（1. Graduate Workstation of North China Electric Power University & Yunnan Power Grid Corporation，
Kunming 650217, China；  2. Department of Mechanical Engineering，North China Electric Power University， Baoding 071003， China；  3. Electric Power Research Institute， Yunnan Electric Power Test & Research Group
 Co.， Ltd.， Kunming 650217， China）

Abstract： In order to solve the problems that the time for calculating the reliability is too long when the transformer reliability assessment is carried out，the research of theory and simulation was investigated. The theory of the reliability of transformer based on FTA was briefly described，which experts' experience play a great role，the reason that the reliability is low for some transformers was been in analysis. After the calculation of the reliability for 138 transformers，the method was presented to fit the reliability of a single transformer based on BP neural network，in which the architecture and the training samples were determined. The weights and threshold of the BP network were optimized through GA，10 sets of data were used to verify the accuracy of the reliability. The results indicate that the BP neural network can fit the reliability of the transformer effectively，improving the rate and the error is in the acceptable range；the optimization of the BP neural network based on GA improved the accuracy of the fitted reliability.
Key words： transformer； reliability； fault tree analysis（FTA）； BP； the optimized BP based on GA