基于随机森林算法和稳态波形的非介入式工业负荷辨识

非介入式工业负荷的准确辨识可以获取工厂内各负荷的运行情况，有利于需求侧智能用电管理。工业负荷由于采集暂态数据建模困难、需要高精度测量设备等特点，造成辨识方法复杂难以实现。针对这种情况，提出一种利用随机森林算法和稳态波形的非介入式工业负荷辨识方法。首先，通过事件监测工业负荷功率状态变化并提取稳态波形，根据工业负荷性能不同而引起的电流波形的差异性，构建单个负荷电流稳态波形的特征数据。然后，利用稳态波形高维度数据作为样本数据，采用随机森林算法中bootstrap（自助）抽样方法和CART算法生成多组决策树。最后，通过投票法对多组决策树进投票辨识得到工业负荷类型。仿真采用某工厂的实际运行负荷数据作为样本数据，通过组合负荷方法仿真比较验证所提辨识算法的有效性和快速性。仿真结果表明：所提的辨识算法准确率达到99%以上、辨识时间3.36 s，远超过贝叶斯辨识算法的准确率63.8%、时间6.15 s，可以有效实现非介入式工业负荷辨识。

Non-intrusive Industrial Load Identification Based on Random Forest Algorithm and Steady-State Waveform

Non-intrusive industrial load identification can accurately acquire the operation situations of each load in the plant,which is beneficial to the demand-side intelligent power management.The identification method for industrial load is complicated and difficult to implement due to the difficulty in collecting transient data for modeling and the demand for high-precision measuring equipment.Aiming at this situation,a non-intrusive industrial load identification method is proposed using random forest algorithm and steady state waveform.Firstly,the steady state waveform is extracted by monitoring the power state change of the industrial load through the event,and the characteristic data of the steady state waveform of the single load current is constructed according to the difference of the current waveform caused by different performance of the industrial load.Then,by using the high-dimensional data of the steady-state waveform as the sample data,the bootstrap sampling method and the CART algorithm in the random forest algorithm are used to generate multiple decision trees.Finally,the industrial load types are identified by voting multiple decision trees.The actual operating load data of a factory is used as the sample data for simulation,and the effectiveness and rapidity of the proposed identification algorithm are verified with simulation.The simulation results show that the accuracy rate of the proposed identification algorithm is more than 99%with the identification time of 3.36 s,which is much higher than the accuracy rate of 63.8%and the identification time of 6.15 s of the Bayesian identification algorithm.Therefore,the proposed identification algorithm can effectively realize the non-intrusive industrial load identification.