非侵入式电力负荷多目标分解框架

现有基于最优化的非侵入式负荷分解方法存在两个问题:使用一到两个特征对家庭负荷的分解效果差;而使用三个及以上特征作为用电设备辨识的目标函数难度高。提出非侵入式电力负荷多目标分解框架,解决传统方法利用特征数少、加权系数难确定等问题。以有功功率、无功功率、视在功率、谐波和电流波形作为电器运行状态的目标函数,建立多目标优化负荷分解模型。利用多目标进化算法(Multi-Objective Evolutionary Algorithm,MOEA)对实测用电数据进行负荷分解求得Pareto最优解集。最后通过多准则决策方法选出识别结果。实验结果表明,增加特征可提高MOEA算法对多个用电设备同时运行时识别准确率,且与当前主流算法相比,所提框架对家庭负荷分解的准确率更高。

A framework for non-intrusive load monitoring using multi-objective evolutionary algorithms

There are two problems in the state-of-the-art Non-Intrusive Load Monitoring (NILM) method based on optimization. Firstly, using only one or two features for load disaggregation is ineffective. Secondly, it is difficult to utilize three or more features for NILM methods based on optimization as the objective function of equipment identification. A framework for non-intrusive load monitoring using multi-objective evolutionary algorithms is proposed, which solves the problems of traditional methods such as using fewer features and difficult to determine weighting coefficients. Active power, reactive power, apparent power, harmonic and current waveforms are taken as objective functions of electric appliances operation state, and a multi-objective load disaggregation model is established. The measured power consumption data for different electric appliances is disaggregated by MOEAs to obtain Pareto optimal solution set. Finally, multi-criteria decision-making method is used to select the recognition result. The experimental results show that the recognition accuracy rate of MOEAs for the case that multiple appliances operate simultaneously is increased with the increasing number of features. Compared with the state-of-the-art NILM methods, the proposed framework has higher recognition rate for household load disaggregation. This work is supported by National Natural Science Foundation of China (No. 61772255, No. 61866025 and No. 61866026) and Natural Science Foundation of Jiangxi Province (No. 20181BAB202025).