基于自适应形态学的直流故障电弧检测特征增强研究

Research on feature enhancement of DC arc fault detection based on adaptive morphology

光伏直流系统中的复杂系统噪声干扰使得故障电弧的特征难以有效提取,从而导致误判或漏判。因此,为了确保系统的可靠性,精准检测故障电弧并增强其检测特征至关重要。为此,研究了形态学对直流故障电弧检测特征的增强作用,提出了一种自适应数学形态学(mathematical morphology, MM)算法,利用粒子群算法结合信号特征对结构元素进行自适应动态寻优,提升信号信噪比同时增强检测特征,并与随机共振和小波分析算法进行对比,验证了提出方法具有很好的特征增强效果,可有效区分正常和故障电弧。最后结合支持向量机对检测特征进行决策,得到了高达96.23%的检测精度。自适应形态学算法能有效地增强故障电弧检测特征,适用于不同实验场景的自适应寻优,有助于精准、有效地检测直流故障电弧。

The complex system noise interference in photovoltaic DC systems makes it difficult to effectively extract the characteristics of fault arcs,leading to misjudgment or omission.Therefore,to ensure the reliability of the system,it is crucial to accurately detect fault arcs and enhance their detection features.For this reason,this article studies the enhancement effect of morphology on DC fault arc detection features and proposes an adaptive mathematical morphology(MM)algorithm,which uses particle swarm optimization algorithm combined with signal features to carry out adaptive dynamic optimization on structural elements,improve signal-to-noise ratio and enhance detection features at the same time.Compared with Stochastic resonance and wavelet analysis algorithms,the proposed method has a good feature enhancement effect,and can effectively distinguish between normal and fault arcs.Finally,a decision was made on the detection features using support vector machines,resulting in a detection accuracy of up to 96.23%.The adaptive morphological algorithm proposed in this article can effectively enhance the detection features of fault arcs and is suitable for adaptive optimization in different experimental scenarios,which helps to accurately and effectively detect DC fault arcs.