基于间隙度的变换器早期故障诊断方法

Gradation and Classification of Early Fault for the Converter Based on Gap Metric

变换器早期故障特征微小,极易被电网复杂工况信号淹没,因此早期故障诊断可以避免电力系统故障发酵,保证供电安全。但传统基于电流、电压残差信号的方法难以实现早期故障诊断。引入间隙度作为一种新的度量工具,可以由系统间内部距离描述早期故障。通过卡尔曼滤波器观测的输出状态可以建立早期故障模型,将模型间的间隙度作为早期故障诊断特征,由系统内部衡量故障偏差,进行故障类别和故障强度诊断,提高系统稳定性。通过仿真实验搭建变换器模型,验证以间隙度为度量工具实现变换器早期故障诊断的有效性,通过与传统残差阈值方法比较,本文方法可以有效提高变换器早期故障诊断灵敏度同时降低误判率。

The converter early fault characteristics are tiny and are easily swamped by the signals of complex working conditions of the power grid, so early fault diagnosis can avoid power system fault fermentation and ensure power supply safety. As the traditional methods based on current and voltage residual signals are difficult to achieve early fault diagnosis, However, the traditional method based on residual signal of current and voltage is difficult to achieve early fault diagnosis. The gap degree was introduced as a new measurement tool, which can describe the early fault by the internal distance between systems. The output state observed by Kalman filter can establish early fault model, taking the inter-model gap as the early fault diagnosis feature, fault deviation was measured within the system, fault category and fault intensity were diagnosed, and system stability was improved. Through simulation experiments, the converter model was built to verify the effectiveness of early fault diagnosis using gap degree as the measurement tool. Compared with the traditional residual threshold method, the proposed method can effectively improve the sensitivity of early fault diagnosis and reduce the false diagnosis rate.