A location and detection method for transient power quality disturbance using SVD-ILMD

Swiftly and accurately analyze non-stationary disturbance signals within the power grid， a location and detection method for transient power quality disturbance that combines singular value decomposition （SVD） and improved local mean decomposition （ILMD） is proposed. First， noise information is processed by using ILMD and a fuzzy membership function threshold to mitigate noise interference. Then， a difference signal is formulated， and a sliding window SVD is employed to amplify the disturbance features while further suppressing noise interference. In conclusion， an adaptive threshold truncation-based approach for localizing and detecting transient power quality disturbances is proposed， utilizing the feature-enhanced signal. Simulation analysis and algorithm comparisons confirm that the proposed method exhibits precise location， robust resistance to noise， and low computational complexity. Moreover， it demonstrates excellent performance in detecting zero-crossing and minor disturbances.