With the improvement of people's living standards, more and more nonlinear loads appear in the power grid system, bringing harmonic pollution, three-phase imbalance and other power quality problems to the low-voltage distribution network.In order to eliminate harmonics and ensure the safe and reliable operation of the power grid, the active power filter APF is widely used in low-voltage distribution networks.However, the main component of APF is the power switch IGBT, which is very prone to failure at high voltage. Therefore, this paper investigates the IGBT open-circuit fault diagnosis method based on wavelet packet decomposition and particle swarm optimization BP neural network.By analyzing the characteristics of the grid-side current waveform and neutral current waveform under single-tube fault conditions, the low-frequency energy value and waveform coefficient extracted by wavelet packet decomposition are used as fault characteristic values. the fault diagnosis is carried out by particle swarm optimization BP neural network, which shows the effectiveness of the proposed diagnostic strategy.