基于MEA-BP的用户窃电识别模型构建及仿真

The construction and Simulation of user stealing identification model based on MEA-BP

针对传统窃电识别精度低的问题,结合BP神经网络和思维进化算法的原理,提出一种基于(Mind Evolutionary algorithm and back propagation,MEA-BP)的窃电识别模型。在该模型中,通过MEA的子种群趋同操作和异化操作,实现BP神经网络最优权值和阈值的优化;然后在上述优化基础上,构建基于用电参量的BP识别模型,并将测试数据导入模型中,实现异常用电客户的精准判断。仿真结果表明,经MEA优化后的算法只需三次趋同即找到最优BP参数,同时仿真识别的结果与实际结果一致,具有较高的精度。

in view of the low precision of traditional electric stealing identification,a new model of electric stealing identification based on MEA BP is proposed based on the principle of BP neural network and thought evolution algorithm.In this model,the optimal weight and threshold of BP neural network are optimized by the convergence operation and alienation operation of the sub population of MEA.Then,based on the above optimization,the BP identification model based on the power consumption parameters is constructed,and the test data is introduced into the model to realize the accurate judgment of abnormal consumers.The simulation results show that the algorithm after mea optimization only needs three convergence to find the optimal BP parameters.At the same time,the simulation results are consistent with the actual results,and have high accuracy.