| With the rapid development of power semiconductor devices,the proportion of sensitive loads in modern power system is also increasing,and power users have increasingly strict requirements for power quality.Power quality problems have become a common concern for both power suppliers and consumers,and voltage sag is currently considered the most common power quality problems.Monitoring voltage sag events in the grid,determining their sources and optimize compensation and governance configuration schemes are crucial for the stable operation of the power system.Distribution networks account for a large proportion in the power system and are more prone to voltage sags than high-voltage transmission networks.Therefore,it has a great significance to study the detection,identification,and compensation of voltage sags in distribution networks.The main research work of this article is as follows:(1)Simulation research was conducted on different types of voltage sag disturbance sources by using Matlab/Simulink software,and the characteristics of various voltage sag waveforms were studied and analyzed.Further research was focused on several multi-level composite sags,and the effective value waveform of voltage sag was obtained through simulation analysis.(2)A voltage sag detection algorithm based on maximum overlapping discrete wavelet transform(MODWT)is proposed in this thesis.The algorithm uses MODWT to achieve accurate detection of voltage sag start and end times and phase jumps;in response to the issue of MODWT not being suitable for detecting voltage sag amplitude and unable to detect frequency,the approximate signal after three-layer decomposition of MODWT is subjected to Teager energy operator envelope to detect voltage sag amplitude and frequency changes.(3)For the identification of voltage sag disturbance sources,this thesis proposes a voltage sag disturbance source identification method based on numerical derivative dynamic time warping(VDDTW),which combines the natural advantages of dynamic time warping(DTW)algorithm in processing time series.The algorithm combines the effective value of voltage sag with the selected resolution.Firstly,the waveform library required by the algorithm is constructed by the simulation of the sag disturbance source described above and subjected to normalization processing.Then,the waveform library is divided by three features.Finally,the VDDTW algorithm is used to compare the normalized effective value waveform of voltage sag to the waveform in the wavelet library to identify the disturbance source.(4)For the optimal configuration of voltage sag compensation equipment,this thesis uses voltage sag data to identify the proportion of sensitive loads in the system.Based on the proportion of sensitive loads in the system,combined with the probability of voltage sag occurrence,the probability of sensitive equipment shutdown,and the cost of DVR compensation,a mathematical model and model solving method for DVR optimal configuration are proposed.Finally,the IEEE33 node distribution network is used as an example to verify the calculation.The voltage sag detection method studied in this thesis can achieve fast and accurate detection,and the disturbance source identification algorithm has an accuracy of 95.05% in simulation examples.The compensation scheme calculated by the proposed optimization configuration model can improve economic benefits and have good engineering application value.There are 63 figures,12 tables and 83 references in this thesis. |