低压窃电负荷小，难以被及时发现，给电力企业造成了巨大的经济损失。文中基于差分整合移动平均自回归模型(Auto-regressive Integrated Moving Average model, ARIMA)和递归贝叶斯算法，构建了一种针对配电网低压窃电行为的识别方法，该方法结合用户历史数据对低压用户与台区表夜间各时段电力负荷数据进行分析，并算出用户窃电概率，从而发现用户是否存在窃电行为。仿真与实际结果表明：该方法对及时准确发现窃电行为，提高配电线路线损治理效率具有重要意义。
英文摘要:
The energy theft activity of low voltage is difficult to be discovered because of its small load, but it causes huge economic losses to utility companies. This paper introduces an approach based on auto-regressive integrated moving average model (ARIMA) and recursive Bayesian method to identify low voltage energy theft activities in distribution network system. Combing with the historical period data of users, this approach analyzes the power load data of low-voltage users and station meters at various periods of time at night, and calculates the probability of energy theft, so as to find whether users have energy theft behaviors. The simulation and case results show that the proposed approach not only can discover energy theft activities accurately with high calculation speed, but improves the efficiency of energy loss management of distribution lines.