Fault diagnosis for reactor based on Bayesian network

Oil-immersed shunt reactor as a high voltage power grid in the means of reactive power compensation is now essential for power system; it directly affects the health of the entire power grid safety and stable operation. Bayesian network based on its solid mathematical theory, knowledge representation and reasoning effective in dealing with the issue of complex systems have great advantages, coupled with the maturity of DGA technology development, the health of the reactor to make a valid diagnosis possible. In this paper, DGA of reactor input variables for the Bayesian networks attribute data, health status is output on the diagnosis of the reactor were reasonable and effective evaluation.