摘要
针对电网中的用电异常行为,为了解决大部分传统检测方法效率低下以及当前的机器学习模型存在局限性等问题,提出了一种基于时间卷积网络的端到端的用户用电异常检测模型。结果表明,文章提出的方法在中国国家电网公司(SGCC)收集的电表数据集上表现出的分类效果优于现有的支持向量机(SVM)、logistic回归(LR)、卷积神经网络(CNN)和长短期记忆网络(LSTM)等方法。
Aiming at the abnormal electricity consumption behavior in the power grid,in order to solve the problems of inefficiency of most traditional detection methods and the limitations of current machine learning models,this paper uses deep learning methods to propose a terminal based on Temporal Convolutional Network(TCN)end-to-end user electricity anomaly detection model.The results show that the method proposed in this paper shows better classification effects on the electricity meter data set collected by the State Grid Corporation of China(SGCC)than the existing support vector machines(SVM),logistic regression(LR),convolutional neural networks(CNN)and Long Short-Term Memory(LSTM)and other methods.
出处
《科技创新与应用》
2021年第27期145-147,共3页
Technology Innovation and Application
关键词
用电异常检测
非技术性损失
时间卷积网络
electricity anomaly detection
non-technical losses
temporal convolutional network(TCN)
