Analysis Methods Of Power Quality Complex Disturbance Based On Phase Space Reconstruction And Atomic Decomposition

In recent years,with the large-scale use of various nonlinear power electronic devices and sensitive load in the modern industry,and the emergence of electricity access,such as micro grid and distributed power,the power quality problems have appeared to be increasingly prominent.The ability of analysising the vast data from real-time monitoring of the gird is the key to improve the power quality.Due to the change of disturbance events and the complex disturbances,there may be some disturbance events in the data from the power quality monitoring device.The waveform data may contain various forms of the signal components,which occur at different times.Only by getting the signal components which represent different disturbance events,can we analysis the basic disturbance event better.This paper makes a deep research on the disturbance waveform recording and disturbance basic events recognition.Atomic decomposition has great advantage of analysising the discontinuous,segmented and complex signals.So atomic decomposition is proposed to extract the feature parameters in this paper.Because of the long disturbance waveform recording,when this method is used to decompose the whole disturbance waveform recording,it may lead to the non-ideal result,such as the large amount of inner product calculation and inaccurate matching parameters of fundamental frequency component and disturbance.This is not conductive to the extraction and identification of the disturbances.Phase space reconstruction theory is the new development of nonlinear time series,which can excavate the potential characters of signals.So phase space reconstruction is proposed to analysis the whole disturbance waveform recording before atomic decomposition.Phase trajectory is reconstruction by determining reconstruction parameters.According to the difference of phase trajectory,fundamental frequency and disturbances component is separated.Then atomic decomposition is used to extract feature parameters of fundamental frequency and disturbances component in determined waveform section contrapuntally,thus the interaction between fundamental and disturbance can be avoided and the computational efficiency of inner product can be improved.At the same time,based on fast fourier transform,particle swarm optimization and matching pursuit,hierarchical searching method is proposed to extract the feature parameters.The logic identification method is established to identify the power quality disturbance signal based on the relationship between fundamental frequency component and disturbance.The simulative results show that this method can get the type,time sequence and component of basic events from disturbance waveform recording,and the validity and correctness of this method are verified.