To ensure the safe and stable power supply of medium and low voltage distribution networks,a single-phase grounding fault diagnosis method based on linear decision functions was studied to improve the fault diagnosis effect.By introducing screening compression method and approximate matrix method into linear discriminant analysis,an improved two-step linear discriminant analysis method was obtained;the optimal single-phase grounding fault characteristics of medium and low voltage distribution networks was extracted in combination with Fisher's linear decision function;within a probabilistic neural network,input single-phase grounding fault characteristics and output single-phase ground fault diagnosis results.The experimental results demonstrate that the proposed method can effectively extract single-phase grounding fault features,and there is no confusion between the fault features of each category.It has excellent fault feature extraction performance and high fault diagnosis accuracy.
关键词
线性决策函数/中低压配电网/单相接地/故障诊断/近似矩阵法/概率神经网络
Key words
linear decision function/medium and low voltage distribution network/single-phase grounding/fault diagnosis/approximate matrix method/probabilistic neural network
维普
万方数据