摘要
随着光伏发电装机容量逐年增长,从光伏发电设备采集的电流、电压等数据规模呈现指数爆炸。针对此类情况,提出基于人工智能的方案。通过均值漂移聚类的方式将光伏大数据进行降维操作,而后利用相似矩阵将光伏数据转化为矩阵块,最后使用三维卷积神经网络进行训练、预测故障。实验表明,本方案的技术路线可以有效进行光伏发电设备的故障诊断。
As the installed capacity of photovoltaic power generation increases year by year,the scale of current,voltage and other data collected from photovoltaic power generation equipment presents an exponential explosion.In view of this situation,this paper proposes a scheme based on artificial intelligence.In this scheme,the dimension of photovoltaic big data is reduced by drift clustering,and then the photovoltaic data is transformed into matrix blocks by using similarity matrix.Finally,3D Convolutional Neural Network is used for training and fault prediction.Experiments show that the technical route of this scheme can effectively diagnose the fault of photovoltaic power generation equipment.
出处
《电力系统装备》
2021年第24期109-110,共2页
Electric Power System Equipment
关键词
均值漂移聚类
三维卷积神经网络
相似矩阵
mean shift clustering algorithms
3D convolutional neural networks
similar matrix