A novel adaptive detection method for generic power quality waveform abnormality

The power quality monitoring data are widely applied in equipment fault detection and diagnosis, such as cable fault detection. Successfully detecting the abnormal signal from the power quality data is particularly important. This paper pro posed a generic abnormality detection method based on the residual-singular-value (RSV). The proposed method takes advantage of the essential characteristics of power quality waveform and the singular value decomposition (SVD) is performed on a matrix constructed by residual signals. Abnormality can be detected by measuring the change of singular values. In addition, the features of noise and singular value is excavated, then the adaptive threshold is proposed for abnormality detection. Simulation data and the field-measured data are both used to verify the feasibility and accuracy of the proposed algorithm.