Research On Voltage Dip Online Detecting System

Voltage dip is a kind of dynamic power quality problem with high probability. Its impact on the sensitive electrical equipment is serious, and every year the economic loss caused by voltage dip is large. Therefore, monitoring the voltage dip characteristic features quickly and accurately is an important basis for voltage dip assessment and governance.Firstly, the paper pays great attention to the commonly used voltage dip detection algorithm(RMS method, defect voltage method, wavelet transform, dq transformation), analyzes their principle in detail. On this basis, several MATLAB simulation comparisons that containing noise and harmonic are given, and the characteristics of various algorithms are summarized. The paper also gives an assessment of advantages and disadvantages of traditional algorithms when detecting voltage dip(begin-end time, phase angle jump, dip depth).As the traditional algorithms relay too much on the detection of dip depth, they have poor real-time performance when detecting the beginning and ending moment of dip by setting RMS threshold. Therefore, a method of locating the begin-end moment of voltage dip based on the residual error curve is proposed. Through the research on the establishment principle of the best auto-regressive model(AR model), the paper gives the method to predict the next sampling point data on the basis of AR model. Then, the beginend moment of dip is detected by the abrupt change point of the residual curve. The method is completely independent of the detection of dip depth, and there is no need to use real-time filter. So the method’s real-time performance is good. In addition, the new method’s detection performance under different conditions is verified by MATLAB simulation.As the traditional algorithms take too much time on the detection of dip depth and phase angle jump, the paper gives a method that includes zero position matching and sliding RMS based on advanced prediction. On the basis of AR model, the method takes the way of predicting a cycle of sampling point data in advance that reduces the detecting delay obviously. In addition, to match zero position accurately, the paper also designs an IIR filter which is used to filter locally. At last, the new method’s detection performance is verified by MATLAB simulation.On the basis of the proposed new method of voltage dip detection, the paper designs a kind of voltage dip online detecting system which includes the detecting device based on ARM and the master station monitoring software based on VB6.0. The μC/OS-II embedded operating system core is transplanted into ARM processor, and corresponding tasks are planned according to the hardware function module. The master station monitoring system can monitor voltage dip information online with the new algotithm based on hardware, and a SQLite database is established for real-time storage and query. Finally, the performance of voltage dip online detecting system is tested, and the results show that it has high detection precision and reliable Zig Bee communication ability.