【摘要】 电弧故障是引起电气火灾的重要原因之一。针对串联电弧故障随机性、多样性和隐蔽性等带来的诊断难题,为提高故障诊断率,设计了一种新的串联电弧故障诊断方法。借助高频电流传感器和高速数据采集系统采集串联电弧故障电流,通过分形维数定量衡量高频电流信号的混沌特性,以便提取串联电弧故障的特征信息,以盒维数和关联维数构造串联电弧故障的特征向量,采用最小二乘支持向量机对电流信号的特征向量进行分类,实现了线路正常与串联电弧故障状态的正确区分。运用所建立的实验平台验证了整个诊断方法的有效性,实验结果表明,串联电弧故障诊断率达到98%以上,所设计的诊断方法具有良好的泛化能力。更多还原

【Abstract】 Arc fault is one of the important reasons for electrical fires. Due to the diagnostic difficulties of randomness, diversity and concealment in series arc faults, a new diagnostic method for series arc fault is designed in this paper to raise the fault diagnostic rate. Series arc fault currents were acquired by high frequency current transformers and a high-speed data acquisition system. To extract characteristic information of series arc faults, the chaotic characteristics of high frequency currents were quantitatively described by fractal dimensions. Afterwards, the featured vectors of series arc fault were constructed by the box dimension and the correlation dimension. And current characteristic vectors were classified by least squares support vector machine(LSSVM). Then, the normal state and the series arc fault state were correctly discriminated. Finally, the designed method was verified via the set-up experimental platform. The diagnostic rate of series arc fault is over 98.0%, which shows the designed method has good generalization ability.更多还原