高效的用户异常用电行为检测是保证电网安全稳定的基础，针对现有用户异常用电行为检测方法存在的准确率差和效率低等问题，基于电能信息采集系统，提出一种将改进K-means算法、改进动态时间规整算法(dynamic time warping, DTW)和Stacking集成学习算法相结合的用户异常用电行为检测策略。通过改进K-means算法完成用户用电行为的特征曲线聚类，通过改进的时间动态弯曲算法完成用户异常用电初筛，通过Stacking集成学习算法判断用户异常用电行为。通过实验验证其优越性。结果表明，所提方法与常规检测方法相比具有最佳检测性能，检测准确率为99.93%，检测时间为8.32 s。可为电网安全提供技术支撑。
英文摘要:
Efficient detection of abnormal user electricity consumption behavior is the foundation for ensuring the safety and stability of the power grid, in response to the problems of poor accuracy and low efficiency in existing detection methods for abnormal user electricity consumption behavior, a user abnormal electricity consumption behavior detection strategy is proposed based on an electric energy information collection system, which combines an improved K-means algorithm, an improved time dynamic bending algorithm, and a Stacking ensemble learning algorithm. By improving the K-means algorithm to cluster the characteristic curves of user electricity consumption behavior, the time dynamic bending algorithm to complete the initial screening of user abnormal electricity consumption, by using the Stacking ensemble learning algorithm to determine user abnormal electricity consumption behavior. Its superiority is verified through experiments. The results indicate that, the proposed method has the best detection performance compared to conventional detection methods, with a detection accuracy of 99.93% and a detection time of 8.32 seconds. It can provide certain assistance for the safety of the power grid.
