基于深度学习的配电网NTL检测方法探究
Deep Learning-based NTL Detection for Distribution Networks
赵广旭 1黄小燕2
作者信息
- 1. 许昌开普检测研究院股份有限公司,河南许昌 461000
- 2. 许继电气股份有限公司,河南许昌 461000
- 折叠
摘要
为了更准确地检测配电网中的非技术损失(NTL),提出基于粒子群优化(PSO)算法和反向传播(BP)神经网络的检测方法.与传统的BP神经网络相比,PSO-BP神经网络在多个评价指标上表现出良好的性能,准确率、精确率、召回率、F1-Score及AUC值分别提高到86.98％、87.20％、86.68％、86.94％和86.65％.PSO算法的加入显著提高了网络权重和偏置值的优化程度,从而增强了模型对NTL的检测能力.
Abstract
Aiming at more accurate detection of non-technical losses(NTL),a detection method based on particle swarm optimization(PSO)algorithm and back propagation(BP)neural network is proposed.By comparing with conventional BP neural networks,PSO-BP neural networks have shown good performance indicated by multiple evaluation indicators,as accuracy,precision,recall,F1 Score,and AUC values have increased to 86.98％,87.20％,86.68％,86.94％,and 86.65％,respectively.The introduction of of PSO algorithm significantly improves the optimization of network weights and bias values,thereby enhancing the model's ability to detect NTL.
关键词
深度学习/配电网/非技术损失/神经网络/粒子群算法Key words
deep learning/distribution network/non-technical loss/neural networks/particle swarm optimization引用本文复制引用
出版年
2024
NETL
NSTL
万方数据