基于剩余电流固有模态能量特征的生物触电故障诊断模型

针对未来低压电网剩余电流保护技术中,生物触电故障诊断与剩余电流之间具有不确定的潜在规律及关系映射,提出了一种基于剩余电流固有模态能量特征的生物触电故障诊断模型。首先应用Hilbert-Huang变换明确了生物触电故障时,剩余电流各固有模态能量在时间和各种频率尺度上的分布,其中低频IMF分量的能量占有率高达86.35%,建立了剩余电流固有模态能量特征的提取方法;然后以选取剩余电流各IMF分量5维度能量特征向量,为生物触电故障诊断模型提供有效特征的信息源,利用量子遗传计算的快速寻优性和神经计算的自适应性有机结合,建立了一种量子遗传模糊神经网络作为触电故障模式分类归属的决策系统,仿真试验准确率达到100%。为研发基于人体触电电流而动作的新型剩余电流保护装置,提供可靠的理论依据和方法支撑。

Fault diagnosis model for biological electric shock based on residual current intrinsic mode function energy features

Abstract: Residual current operated protective devices (RCDs) have a wide range of application in low-voltage power grids. RCDs play an important role in preventing electric shock hazard and avoiding fire disaster caused by ground fault. In general, the root mean square (RMS) value of whole leakage current signal detected is considered as the unique criterion to determine whether the protector acts or not. The traditional RCDs cannot classify and identify electric shock fault type automatically based on whole leakage current signal. Theoretical analysis and operation experience indicate that such a criterion is unavailable in identification if an organism electric shock fault has occurred. The uncertain potential regularity and mapping relations exist between the biological shock fault diagnosis and residual current. To decrease the malfunction and tripping phenomenon and increase the reliability and the rate of proper commissioning for RCDs, a fault diagnosis model for biological electric shock based on residual current intrinsic mode function (IMF) multidimensional energy features is proposed innovatively for residual current protection technologies in the future low-voltage power grid. First, the electric shock current of organism (animal) is decomposed into five IMF components and one residual component by Hilbert-Huang transform method. The energy share of low frequency component IMF4 and IMF5 is as high as 86.35%, which can meet the needs of more than 86% for the measured signal, and the correlation coefficient of the amplitude of the low frequency IMF components is up to 0.99 or more. And the distribution of IMF energy on time and various frequency scales is made clear when the biological electric shock fault occurs. Residual current signal performance information is converted into energy feature vectors. The extraction method of IMF energy features in residual current is established. Then, the five-dimensional energy eigenvector in each residual current IMF component is selected to provide effective characteristics information source for biological shock fault diagnosis model. By combining the rapid optimization of quantum genetic computation and the self-adaptability of neural computation, the quantum genetic fuzzy neural network is established as the decision system of electric shock failure mode classification. The method has the advantages of self-adaptive resolution, good fault tolerance, high robustness and high accuracy. And the accuracy of simulation experiment reaches 100%, and the method avoids the local minimum of traditional gradient descent learning algorithm and improves the learning efficiency. The network error is 0.00099758 when the optimal learning algorithm is iterated to 1 156 times, which meets the error accuracy requirement. The problem that the electric shock fault type is not identified effectively in the engineering is resolved using this method. The reliable theoretical basis and supporting method for developing new generation residual current protection device are provided to ensure the personal safety and the safe operation of the low-voltage power grid, which is based on the action being caused by electric current component of the human body electric shock.