基于混沌的故障电弧早期弧声检测方法研究

Study on the Detection Method of Early Sound Signature of Fault Arc Based on Chaos

故障电弧发生之前伴生的微弱电弧声频率主要集中在5~10kHz附近的特征频段里,准确快速检测该频段内特征弧声是否存在,即可判断故障电弧是否将要发生,实现故障电弧的早期预测预警;基于此,提出基于混沌的故障电弧早期弧声信号检测方法,建立以Duffing振子为基础的混沌检测系统;混沌检测系统对可听波7kHz附近3%频段内微弱的早期弧声信号非常敏感,对其它频率的声信号具有良好的抑制作用;当没有电弧声产生时,系统处于混沌状态;而电弧声一旦产生,系统立刻切换到间歇混沌状态;根据间歇混沌的存在与否可以判断特征弧声是否已经产生;试验结果表明,该方法能够可靠检测出微弱故障电弧早期弧声信号。

The signature band of sound wave is within the range from 5kHz to 10kHz before fault arc taking place.By monitoring the existence of weak arc in this band,early detection and warning of switchgear arc fault can be realized.The proposed technique detects the arc sound signature nearby the frequency of 7kHz based on Duffing chaos oscillator.The arc sound is recorded using fiber microphone and imported into the chaos detection system.The time course of the detecting system state variables switches form chaos into intermittent chaos as soon as arc sounds signature taking place.Observing the motion state of chaotic detecting system can identify arc sounds.The result of experiments has shown that arc sounds can be reliably detected under strong background of electromagnetic noise by using the method which afford a possibility of forecast and early warning to fault arc.