Study On Detection Method Of Typical Power Quality Problems Based On Sparse Representation

In an ideal power supply system, only the sine signal with constant frequency is provided to the users uniquely. However, with the use of non-linear loads and the occurrence of grid failures, lots of power quality problems(i.e., Harmonics, Interharmonics, Transient impulse, Transient oscillation) are brought into the grid signal, which lead to the inconvenience to the normal operation of equipment and to the safe operation of power system. In order to adopt appropriate measures to improve power quality and reduce the negative influences brought by the disturbances, it is of great significance to analyze and detect the power quality problems.Based on the sparse characteristics of the typical power quality problems, the sparse representation theory is introduced to the analysis and detection of the typical signals of the power quality.In order to analyze and detect the steady-state power quality problems(i.e., Harmonics, Interharmonics), a kind of signal sparse representation method based on harmonic dictionary is proposed in this thesis. Based on basis pursuit theory, a mathematical model of signal representation is constructed in this method. Meanwhile, the harmonic dictionary corresponding to the characteristics of the signal is also built. The augmented Lagrangian method is used to solve the mathematical model. Based on the proposed method, the steady-state components in the signal are transferred to the sparse coefficients in the frequency domain. The accurate analysis of the steady-state components provides favorable and useful data for controlling harmonics.In order to analyze and detect the transient power quality problems(i.e., Transient impulse, Transient oscillation), the morphological component analysis is incorporated into the basis pursuit theory to construct a dual basis pursuit mathematical model of the signal; an impulse dictionary is built based on the characteristics of the transients in the signal. Furthermore, a union dictionary combining the impulse dictionary and the harmonic dictionary is proposed in this thesis. And the alternative direction multiplier method is used to solve the constrained optimization model. Based on the proposed method, the different components in the signal can be separated sparsely. Since the different components can be accurately detected by the proposed method, it is very important to provide a scientific and accurate data support for controlling the transient power quality problems.