基于HCMAC神经网络的故障电弧检测方法

低压故障电弧电流的有效值往往小于热脱扣器的动作电流,导致传统的电路保护装置失效.模拟低压电气线路中的故障电弧,利用电流采样装置对故障电弧电流进行采样.在分析故障电弧电流波形特征的基础上,利用HCMAC神经网络对电流波形的各周期均值、斜率和小波变换高频系数三个判据进行综合评判,有效克服了单一判据的局限性.仿真结果表明,基于HCMAC的故障电弧检测方法可有效识别故障电弧特征,提高辨识故障电弧的准确率.

Detection Method of Low Voltage Fault Arc Based on HCMAC Neural Network

The RMS current of low voltage fault arc is often less than the operating current of the heat relaese, which results in failure of the traditional circuit protection device. By simulating the fault arc phenomenon in low voltage electrical line, the fault arc current was collected with the help of current sampling device. Based on analysing the current wave characteristics and using HCMAC neural network to judge the three comprehensive evaluations cfiterian, mean, slope, and the wavelet transform frequency coefficients of fault arc current,it effectively overcomes the limitation of a single criterion. The simulation results show that with this method the fault arc characteristics can be effectively identified, and the accuracy of fault arc identification is improved.