电器辨识算法与系统研究

Research on the Electric Appliance Identification Algorithm and System

家庭用电负荷分解数据的加工利用,对于提高节能减排水平以及促进智能电网的发展具有重要意义。负荷监测有侵入式和非侵入式两种主流监测方式。但是它们都因各自的缺点而难以得到广泛应用。通过对这两种监测方式的研究,设计了一种新的负荷监测方式。通过将非侵入式结构和侵入式结构相融合,基于电流差异性和基尔霍夫电流定律,建立了电力干线数学模型。采用C-均值法对实时监测到的电流进行分类,达到辨识效果。对9种常用的家用电器分3个辨识节点进行试验。试验2 h内,辨识的准确率可达99.98%。该系统对于推动家庭服务类人工智能的发展具有积极作用。

Processing and utilization of the decomposed data of household electricity loadis of great significance for the improvement of energy-saving emission reduction and the development of smart grid. Currently, there are two mainstream load monitoring methods, invasive and non-invasive. However, these are both difficult to be widely used because of their certain disadvantages. Through researching these monitoring methods, a novel system for load monitoring is applied. The non-intrusive structure is fused into intrusive structure, based on current difference and Kirchhoff5 s current law, the mathematical model of power trunk is established. The classification of the current data detected in real-time is implemented by the C-means method,the effect of identification is reached. Experiments are conducted for 9 kinds of household appliances that are connected to 3 identification nodes. During two-hour experiment, the accuracy of identification can reach 99. 98%. The system promotes the development of artificial intelligence for household services.