Abstract: DC-link capacitors are key components for energy buffering in converter systems and are widely used in fields such as locomotive traction, aerospace, wind power generation and photovoltaic power generation. In harsh working environments, the performance of DC-link capacitors deteriorates, the failure rate increases, and they are easily damaged. Currently, the existing methods for state identifying of capacitors are easily affected by sensor measurement noise, resulting in a significant difference between the capacitance identification results and the actual values. To solve these problems, this paper proposes a capacitor state identification method based on the pre-charging process. By constructing the pre-charging model, the mapping relationship between capacitance and voltage signal is obtained, and the influence of measurement noise is minimized. Then, the recursive auxiliary variable least square method is used to estimate the parameters of the pre-charging model, so as to obtain the accurate capacitor capacity identification results. The verification results show that this method can effectively resist the interference of measurement noise, and has good identification effect when the signal sampling frequency is greater than 500 Hz. The capacitance identification error is 0.569% when the sampling frequency reaches 1 kHz.

Key words: capacitor, capacitance, pre-charging, state identification