微型电脑应用

针对窃电现象普遍存在及反窃电形势日益严峻的问题，建立基于电力用户用电信息大数据的用户窃电行为辨识模型。通过对用电信息大数据的预处理和分析，建立能有效反映用户窃电行为的电压、电流、电量和功率因数的电气量异常检测模型。利用改进随机森林模型，实现电气量异常特征量与用户窃电行为的非线性映射预测诊断。建立用户窃电行为辨识计算实例，通过用户窃电行为辨识结果的对比分析表明辨识方法的有效性和优越性，可为供电公司对用户窃电行为的辨识及反窃电相关工作提供有效的技术参考和指导。

In view of the widespread phenomenon of power theft and the increasingly severe situation of anti-power theft, this paper establishes a user power theft behavior identification model based on big data of power consumption information of power users. Through the pre-processing and analysis of big data of power consumption information, the abnormal detection model of electricity consumption that can effectively reflect the user's power theft behavior is established. The improved random forest model realizes the nonlinear mapping prediction and diagnosis of the abnormal characteristic quantity of electricity consumption and the behavior of electricity theft. A calculation example of user behavior identification is established, and the effectiveness and superiority of the identification method in this paper are demonstrated by comparing the identification results of user behavior identification. The research results can provide effective technical reference and guidance for the power supply company to identify the power theft behavior of users and anti-power theft related work.