Research on Classification Method of Transients in Nuclear Power Plant Based on Curve Similarity Comparison

The transient in the primary circuit of nuclear power plant is one of the important references to characterize the fatigue of the primary circuit materials. The key information of the transient data is the transient species, start time, and duration. The detailed data is stored in a special data recording computer in the form of time series data. Dynamic time warping (Dynamic Time Warping, DTW) algorithm is an intelligent algorithm which can compare the similarity of curves at different time scales. Some researchers have used the designed transient data and virtual transient to verify the availability of the algorithm for transient classification. According to the characteristics of transient data classification, the DTW algorithm is optimized and modified to realize the similarity calculation between real transient data, and the similarity calculation results are given. The real transient data of M310 unit is used to test the algorithm, and the results show that the optimized DTW algorithm has good applicability, and the classification results can meet the daily use needs of transient engineers in power plants.