An improved state‐space average model of the ultra‐high voltage inverter based on wavelet decomposition and reconstruction

Abstract

SummaryUltra‐high voltage inverters are widely used as grid‐connected devices in new energy grids, and the state‐space average model is the most practical modeling method for the inverter. However, the accuracy of this modeling method is significantly reduced because it ignores the high‐frequency switching process, which contains overshoot voltage and other factors that affect system stability. Moreover, the amplitude of the overshoot voltage in the ultra‐high voltage inverter will be huge, which may lead to the damage to switching devices, circuit overheating, and even the collapse of the system. Therefore, the wavelet decomposition and reconstruction methods are used in this paper to extract the transient characteristics of the inverter and integrate them into the state‐space average model. Compared with the traditional state‐space average model, the improved model method can obtain a more accurate overshoot voltage peak value with the same simulation step size. Then, it can provide a more precise reference for the withstand voltage in the device selection of a high voltage level system. The single‐phase full‐bridge inverter is taken as an example to illustrate the construction process based on the wavelet decomposition and reconstruction method. Finally, the simulation waveform is verified to be in good agreement with the experiment.