Fast and precise voltage sag detection method for dynamic voltage restorer (DVR) application

• The articles in the field of voltage sag detection method is reviewed.
• Important criteria to evaluate performance of any voltage sag detection method are discussed.
• A voltage sag detection method is proposed.
• Analytically, performance of the proposed method is investigated under frequency variation.

Voltage sag detection method has a significant impact on the performance of dynamic voltage restorer (DVR). This paper proposes a fundamental voltage amplitude detection method which is applicable in both single/three-phase systems for DVR applications. Advantages of the proposed method include application in distorted power grid without any low-pass filter, precise and reliable detection, simple computation and implementation without using a phased locked loop and lookup table. The proposed method has been simulated and implemented experimentally and tested under various conditions considering all possible cases such as different amounts of voltage sag depth (VSD), different amounts of point-on-wave (POW) at which voltage sag occurs, harmonic distortion, line frequency variation, and phase jump (PJ). A detailed procedure to design the threshold band for comparator is presented. According to the achieved results and detailed analysis, maximum ripple of the calculated fundamental voltage amplitude is ±0.3% in normal condition of line frequency and ±1% in the case of ±1 Hz variation of 60 Hz line frequency. In addition, the error between the actual fundamental voltage amplitude and the calculated one is almost zero in normal condition of line frequency and maximum of ±0.25% in the case of line frequency variation with excessive-harmonic distortion. Furthermore, detection time (DT) of the proposed method is analyzed in detail for different amounts of VSD, POW and PJ. The best and worst detection time of the proposed method were measured 1 and 8.8 ms, respectively. Finally, the proposed method has been compared with other methods available in the literature.