Dissertation > Excellent graduate degree dissertation topics show

Power Quality Disturbances Detection and Identification Based on Signal Processing Theory

Author: WangLiXia
Tutor: HeZhengYou
School: Southwest Jiaotong University
Course: Proceedings of the
Keywords: Power Quality Noise Canceling Disturbance detection and localization Disturbance Identification Time-frequency distribution Stratified analysis system
CLC: TM933.4
Type: Master's thesis
Year: 2010
Downloads: 225
Quote: 0
Read: Download Dissertation

Abstract

Modern power system load change makes power quality problems have become increasingly prominent, seriously affecting the safe operation of power system stability, and to the production and life has brought huge economic losses. In summary and analysis of previous research, based on the combination of signal processing related theories of power quality disturbance signal denoising, detection and identification of issues. Introducing based on wavelet transform and mathematical morphology noise cancellation algorithm, based on the research methods used in these two types of power quality disturbance signal adaptive noise cancellation and algorithm selection problem. Reconstruction factor established to evaluate the signal reconstruction algorithm ability to use two methods typical representative signal and two basic algorithm common power quality disturbance signal is analyzed by calculating the reconstruction of factors, discussed two algorithms for different adaptive signal. Through simulation analysis, the following conclusions: For non-transient pulse or high frequency oscillation signal disturbance, two algorithms are effective, according to the calculation of the different requirements of speed and accuracy to be selected; For transient pulse and oscillatory perturbations with signal de-noising based on wavelet transform is better than denoising based on mathematical morphology method. Using mathematical morphology mutations in information retention signal of good results, research based on mathematical morphology short duration power quality disturbance detection and positioning. Describes and analyzes three kinds of disturbance detection based on mathematical morphology and localization method, which is based on the first order derivative and morphological gradient method, based on morphological gradient and soft thresholding method based on decomposition and hat dq transformation method, through Comparison of three kinds of simulation methods in the analysis of voltage dips, voltage sags l, electromagnetic transient oscillation signal aspects of adaptability, able to correctly detect and locate the disturbance occurred at any time, the results show that decomposition and hat on dq transformation method the detection of the zero disturbance has a good effect, thus selecting hat on dq transformation method decomposition and the measured data, the disturbance detection and location analysis. Cohen category bilinear time-frequency distribution is a real sense of joint time-frequency distribution, the paper discusses the Cohen time-frequency distribution of power quality disturbance signal applied to identify the possibilities. Is given based on the rearrangement of quadratic time-frequency distribution of new power quality detection method, first using instantaneous reactive power theory and generalized morphological filter will power quality disturbance signal component and the fundamental component separation, reuse rearrangement two sub-type time-frequency distribution analysis of components of the disturbance, resulting in better performance when the disturbance frequency component of aggregate time-frequency joint distribution. Simulation examples show that the method intuitive when expressed disturbance signal frequency characteristics, but in the feature extraction and reduce the computational quantify the amount of cross-term problem solving, there are still some difficulties. Presents a linear time-frequency distribution and the binary threshold characteristic matrix of power quality disturbance classification. First, a combination of two linear time-frequency distribution (window Fourier transform and S-transform) the advantage is given to characterize the signal extracted features five characteristics and its binarization processing; On this basis, established based on binary threshold characteristic matrix perturbation classification criterion, characterized by a binary matrix with a threshold value is determined by comparing the type of disturbance. Rise against voltage sags, dips, interruptions, cut marks, oscillations, pulses, harmonics, harmonic and dips, harmonics and temporary l nine common disturbance simulation results show that this method has higher accuracy rate (gt ; 98%), indicating that the proposed method is correct and effective. Established a hierarchical power quality disturbances recognition system, given some of the characteristics and structure of the system, each function module specific algorithm. The hierarchical power quality disturbance detection systems include the fundamental and separation module disturbance, the disturbance time feature extraction and classification module seven functional modules, by dq transform, generalized morphological filtering, Fourier transform simple calculation method of signal analysis combination, layer by layer extracted amplitude perturbation time singular entropy perturbation frequency domain characteristics and classification, the final layers of classification based on the results of the integrated signal disturbance type identification. For seven common single power quality disturbances and some mixed power quality disturbance simulation results show that method has better classification results.

Related Dissertations

CLC: > Industrial Technology > Electrotechnical > Electrical measurement techniques and instruments > Power the number of measurements and instrumentation > Power measurement,power meter
© 2012 www.DissertationTopic.Net Mobile