基于自适应谱线增强的光伏电弧故障检测方法OA

The complex operating conditions of photovoltaic(PV)systems make arc fault detection particularly challenging,especially when arc signals exhibit weak amplitudes and spectral characteristics that are easily obscured by background noise,making it difficult to accurately capture their transient behaviors and nonlinear features.Building upon direct-current(DC)arc fault detection and incorporating time-frequency domain analyses of fault characteristics in four DC and PV system scenarios,this paper proposes a photovoltaic arc fault processing method based on adaptive line enhancement(ALE).The proposed method not only effectively addresses detection challenges in DC systems but also adapts well to photovoltaic systems,particularly under inverter startup noise interference.First,fast fourier transform(FFT)is employed to obtain frequency-domain information.Signal preprocessing is then performed using the density of extreme points combined with a fluctuation factor.Subsequently,a cuckoo search optimized adaptive filter(CSOAF)is utilized for noise suppression and signal smoothing.An improved Teager energy operator(TEO)based on a three-point symmetric difference scheme is introduced to enhance the nonlinear transient characteristics of the signal while suppressing noise components in other frequency bands.Features extracted from the enhanced signals are finally identified using residual connection-based stacked autoencoders(Res-SAEs).Experimental results demonstrate that the proposed method achieves a fault detection accuracy exceeding 98%,indicating its effectiveness and robustness for photovoltaic arc fault detection under complex operating conditions.