基于DTW技术的电站故障诊断

Power Plant Fault Diagnosis Based on DTW

传统的基于Euclidean距离的相似搜索无法处理电站DAS信号之间存在相位差的问题。基于DTW距离的相似搜索可有效地解决上述问题,提高DAS序列数据的分类精度。通过两个实际应用案例,验证了DTW技术在电站故障诊断中的有效性。

The classical similarity search method which is based on Euclidean distance cannot deal with phase difference, which is ubiquitous between DAS signals in. power plants. Based on DTW distance, similarity search can overcome the obstacle and increase the classification accuracy of data sequences in DAS. Two cases are used to verify the effectiveness of DTW in power plant fault diagnosis.