The Sdudy On Detection And Classification Of Hybrid Power Quality Disturbances

With the rapid development of national economy, the electric power plays a more and more important role in modern production and life. A large number of non-linear and unbalanced loads are put into use and cause a series of power quality problems. At the same time, the demand for power quality (PQ) is high, and the contradiction between supply and demand is more and more outstanding. Power quality detection and identification is the basis of fault diagnosis and evaluation. Disturbance recognition is a very complicated scientific problem, because the actual power quality is very complex, and the difference is not very clear signal disturbance mixing phenomenon. For recognition problems of hybrid PQ disturbances, the three aspects, the detection algorithm, and the recognition algorithm of transient and mixed PQ disturbances are discussed in the paper.In order to find an effective method for detection of PQ parameters, the paper researches the disturbances detection algorithm, and presents the PQ disturbances detection based on complete ensemble empirical mode decomposition (CEEMD), improved ensemble local mean decomposition (IELMD) and Hilbert vibration decomposition (HVD). ①The CEEMD algorithm not only can effectively avoid the mode mixing problem, but also greatly reduce the noise interference. In this approach, the residue of added white noises can be extracted from the mixtures of data and white noises via pairs of complementary ensemble IMFs with positive and negative added white noises. The disturbed signals of intrinsic mode functions (IMF) component decomposed by the CEEMD method offset the added white noise and ensure the completeness of the original signal. The start-stop moment, amplitude and frequency of the power quality disturbances can be detected according to the transient characteristics. ② The local mean decomposition (LMD) is a self-adaptive time-frequency analysis method. Mode mixing phenomenon which makes the decomposition results distortion may be produced when LMD is performed. The paper presented an improved ensemble local mean decomposition (IELMD) power quality disturbance detection and time-frequency analysis method. The method is composed of two parts:ELMD and Hilbert transform. Firstly, to extract the product function (PF) of the signal based on ELMD, and to obtain the instantaneous amplitude of the signal from the PF components of the amplitude modulation function. Then, apply the Hilbert transform to obtain the instantaneous frequency of PF. The IELMD method could effectively locate the beginning and ending time of the disturbances occurred. ③ The Hilbert vibration decomposition (HVD) method is introduced to detect the parameters of the power quality disturbances. The proposed HVD method is based on the Hilbert transform (HT), just as Hilbert-Huang transform (HHT), but do not involve complicated empirical mode decomposition (EMD). Low-pass filter is applied to remove the rapidly varying asymmetrical oscillating part and to leave only the useful frequency. The validity of the HVD method is verified by simulation.In transient power quality disturbance classification, the paper proposes a novel method using spectral kurtosis based on the Choi-Williams distribution combined with RMS algorithm to detect and identify the transient disturbances. Firstly, the voltage swell, sag and interrupt are seen as a class, and we use the spectral kurtosis based on Choi-Williams distribution divide the disturbances into transient pulse, transient oscillation and amplitude disturbance. Then, we use RMS algorithm curve to identify the different amplitude of the voltage perturbation. The algorithm do not need to use any classifier and greatly simplifies the process of algorithm and computation time.The simulation results show that both the spectral kurtosis based on CWD and the RMS algorithm can effectively extract the disturbance characteristic quantities, and have a good anti-noise capability. The method can effectively identify five kinds of disturbance and has a higher recognition rate for single and mixed disturbance.In the hybrid power quality disturbance classification, the method for power quality mixed disturbance classification based on time-frequency domain is presented in the paper. Various time-frequency analysis algorithms are combined to extract features. Firstly, the disturbance signals are processed with complete ensemble empirical mode decomposition and modified incomplete S-transformation, and nine time-frequency domain characteristics are extracted. Then, the characteristics are input into the decision tree for the disturbance identification. In this method, interferences between the single disturbances are fully considered and effectively suppressed through the complementary amount of time-frequency domain features. The simulation and experiment results show that the method can effectively recognize the power quality multiple disturbances including voltage sag, voltage swell, voltage interruption, transient impulsive, transient oscillation, harmonics, fluctuation and their mixed ones.