Abstract: For the problem that it is difficult to accurately obtain the user's sensitive equipment parameters and the production process structure in voltage sag mitigation, a method is proposed for identifying sensitive loads based on available voltage sag monitoring data. Discrete wavelet transform is performed on the voltage and current recorded when the voltage sag occurs to calculate the active power trajectory. The singular value decomposition method is used to identify the starting point of the transition section of the trajectory, and the characteristics of the trajectory and its relationship with the load voltage sag response are obtained. Combined with the voltage tolerance curve of sensitive equipment, the sag and load-response events are grouped. The active power and its recovery ratio after the end of the voltage sag is then calculated based on the recorded data, thereby the active power percentage of typical sensitive loads is identified. Through Matlab/Simulink simulation, the accuracy and validity of the proposed method and indexes are proved, which has obvious engineering significance.

Key words: voltage sag, monitoring data, sensitive load, load-response events, singular value decomposition, load identification, waveform analysis