부산대학교 도서관

Extracting useful information about condition of system and equipment based on power disturbance data is significant to ensure the efficiency and safety of them. However, sensitive and reliable disturbance detection is fundamental for the proactive application of disturbance data. Existing detection methods generally applies to specific disturbances, such as power quality disturbances and fault disturbances, which are pertinent but lack of universality and adaptability. A generic high-sensitivity power disturbance detection method based on the singular value distribution law of differential waveform is proposed. The singular value range (SVR) is used to distinguish disturbance from normal waveform. According to the historical SVR sequence reference value and variance variation of differential data, the threshold is adaptively calculated to enhance the universality and detect disturbances. The proposed method is tested through a large number of simulation data and field-measured data, which verifies its high detection sensitivity for various disturbances and effective performances on weak and complex disturbances under ambient noise.