基于PLA-GDTW支持向量机的非侵入式负荷监测方法

Non-intrusive Load Identification Method Based on PLA-GDTW Support Vector Machine

非侵入式负荷监测方法(non-intrusive load monitoring method,NILM)嵌入智能电表终端,可以进行全面的家庭能耗监测和用电优化管理,实现用户和电网的双向互动,推进智能用电的发展。提出了一种基于分段线性近似及高斯动态弯曲核(piecewise linear approximation Gaussian dynamic time warping kernel,PLA-GDTW)支持向量机的非侵入式负荷监测方法。首先分析家用电器暂态过程的有功功率变化,用暂态时间和暂态功率跳变值进行负荷预筛选,然后对暂态有功功率的波形进行分段线性近似(piecewiselinear approximation,PLA)从而实现波形特征提取和降维表示,最终将波形特征输入到以高斯动态弯曲核(Gaussiandynamic timewarpingkernel,GDTW)为核函数的支持向量机进行分类识别。实验结果显示所提出方法对于持续变化负荷及多负荷同时运行的情况均具有较高的识别率和较快的识别速度。

The(non-intrusive load monitoring method)(NILM)embedded in the smart meter terminal can monitor comprehensive household energy consumption and optimize power management to achieve interaction between the user and the grid and develop smart grid.A non-intrusive load identification method based on PLA-GDTW(piecewise linear approximation Gaussian dynamic time warping kernel)support vector machine is proposed.Firstly,the transient power of devices is analyzed,and devices are pre-filtered with transient time and power step value.Then,extract feature and reduce dimension of active power waveform with piecewise linear approximation algorithm.Finally,based on Gaussian dynamic time warping kernel support vector machine classifiers,the devices are identified.The result shows that the recognition rate is high and recognition speed is fast even though continuously variable devices need to be identified or multiple devices are running at the same time.