非侵入式电力负荷的辨识和监测

Nonintrusive electrical load identification and monitoring

本文提出了一种基于Mallat小波变换的电力负荷瞬态和动态轨迹综合辨识方法。其核心原理在于不同种电力负荷的动态和静态频域特征不尽相同;通过对负荷的电流和电压进行实时采样,并对其有功和无功功率进行实时计算,然后通过Mallat小波变换,将电压与电流、有功与无功功率的分层频谱轨迹特征图实时计算出。通过预先对每种负载的预学习和辨识,在实际使用中便可准确地对各类负载的使用及运行状态做出识别。当系统辨识的负载与预存的数据出现严重差异,可以推测出所接入的设备出现异常,或者有新的负载被接入;系统便可以启动相应的报警和应急处理机制,从而达到对电力负载的实时监控和管理的目的。

In this paper a Mallat wavelet transform based electrical load transient and steady state identification model is proposed. It is based on the fact that different electrical appliance has different transient and steady state spectrum signa-tures. Through load voltage and current sampling, active and reactive power calculation, and using Mallat wavelet trans-form FPGA, the trajectory patterns between voltage and current, between active power and reactive power for each wavelet output frequency segment could be obtained. This group of spectrum trajectory signatures for each type of load could be learnt in advance, so that the load can be identified when being used. When there is a large difference between the identified results and pre-learnt signature data, it means a load could be working abnormally or an alien device has been branched in. The system could then trigger warning and response mechanism, so as to achieve the objective of monitoring and controlling the connected electrical appliances.