AMI环境下基于深度学习的异常用电监测方法

Abnormal power consumption monitoring method based on deep learning in AMI environment

为解决现有异常用电监测方法,在AMI环境下存在的监测性能低的问题,提出基于深度学习的异常用电监测方法。根据用户异常用电机理,构建相应的异常用电模型。在该模型下,设置不同异常用电类型的用电负荷序列标准特征。在AMI环境下,利用智能电表设备采集用电计量数据,利用深度学习算法求解用电异常监测指标。通过计算异常监测指标与设置标准特征之间的关系,得出异常用电监测结果。通过性能测试实验得出结论:设计监测方法的最高准确率为95.5%,且监测面积不小于研究区域面积的90%,即设计方法在监测精度和范围两个方面均满足应用要求。

In order to solve the problem of low monitoring performance of existing abnormal power consumption monitoring methods in AMI environment,an abnormal power consumption monitoring method based on deep learning is proposed.According to the abnormal power consumption mechanism of users,the corresponding abnormal power consumption model is constructed.Under this model,the standard characteristics of power load series of different abnormal power consumption types are set.In the AMI environment,the intelligent meter equipment is used to collect the power consumption measurement data,and the deep learning algorithm is used to solve the power consumption anomaly monitoring index.By calculating the relationship between abnormal monitoring indicators and setting standard characteristics,the abnormal power consumption monitoring results are obtained.Through the performance test experiment,it is concluded that the maximum accuracy of the design monitoring method is 95.5%,and the monitoring area is not less than 90%of the research area,that is,the design method meets the application requirements in both monitoring accuracy and scope.