Sampling Method For Power Quality Signal Based On Wavelet Multi-scale Compressived Sensing

Abstract:Signal sampling is the prerequisites for analy-zing and optimizing the power quality. The speeding change of power quality signal not only brings great challenge to the sampling, but also brings massive data. Compressived sensing theory can compress signal while sampling which reduce not only the requirement of sampling device but also the storage space.First, according to international standards, we modeled a series of common power quality signal, including the singal interference model and multiple interference model. We established the feasibility of applying compressived sensing theory to power quality signal sampling. And then, we simulated this method on various power quality signals. Simulation results show that the compressived sensing theory has the ability to reconstruct the power quality signal with a low ratio, low error under the interferences like harmonic,voltage change or inter-harmonic.According to the characteristics of harmonic signal, we proposed an improved method for rotation measurement matrix. The simulation results show that the improved rotation measurement matrix, reconstruction result could be more accurate under lower compression ratio and lower error. The stable performance is much better than ordinary rotation matrix and Gaussian random measurement matrix.Finally, we proposed the wavelet multiscale compressived sensing method to increase the reconstruction accuracy of power quality signal with pluse or short-term shock interference. The simulation result show that the method performs well not only to signal with pulse or shock interference but also harmonic or voltage change etc. The universal appliction illustrate that this method doesn’t depend on the type of signal. The good peformance meets the requirements of power quality analysis.