Research on Remaining Useful Lifetime Prediction Methods of Main Transformer in Nuclear Power Station

Abstract

Abstract The main transformer of the NPP is the key equipment connects with public power grid and it is also the important electrical equipment to provide power for the power plant electrical equipment during the shutdown of the generator set. In consequence, it’s failure will cause huge consequences and economic losses. Therefore, the research of reaming life prediction technology for nuclear power plants(NPP) is of great significance to the economic operation and safety of NPP. In this paper, according to DL/T984-2005 and other related literature, it is determined that the aging condition of solid insulating fiber material is the main factor to determine the life prediction of transformer. The main aging mechanism of that is summarized, and the effect of temperature is emphatic introduced. In general, the ambient temperature and load curve are the important factors affecting the aging of transformer insulation, furthermore affecting the life of transformer. Therefore, based on GB/T15164-94 recommended ambient temperature and transformer load, this paper calculates the transformer life loss model. On this basis, this paper introduces the basic particle filter (PF) and Kalman filter (KF) algorithm flow. Based on the physical formula, a data model is developed to estimate the transformer life loss. Besides, the results are compared and analyzed. The study finally found that there were no significant differences in the accuracy of the predictions. However, considering the fact that the aging of transformer is a monotonic decreasing process, the particle filter based on degradation rate has a slight advantage in life prediction.