基于多模型融合集成学习的智能变电站二次设备状态评估

为准确评估智能变电站二次设备运行状态,建立了二次设备状态评估指标体系,并结合多种机器学习算法的差异性,提出了基于多模型融合集成学习的二次设备状态评估法。该方法采用双层结构,上层中利用划分好的数据对数个基学习器进行k折验证,下层中利用全连接级联神经网络融合多个基学习器,并采用改进的列文伯格-马夸尔特算法训练该神经网络加速模型收敛。实例分析表明,所提出的方法可以准确地评估二次设备的运行状态,并为智能变电站系统和二次设备的检修工作提供指导意见。

A state assessment method for intelligent substation secondary equipment based on multi-model ensemble learning

In order to accurately evaluate the operational status of secondary equipment in an intelligent substation, this paper establishes an evaluation index system of the secondary equipment status. Combined with the differences of various machine learning algorithms, it also proposes a secondary equipment condition evaluation method based on multi-model ensemble learning. The method adopts a double-layer structure. In the upper layer, k-fold verification is carried out by dividing the data into several base learners. In the lower layer, a fully connected cascaded neural network is used to fuse multiple base models, and the improved Levenberg Marquardt algorithm is used to train the neural network to accelerate model convergence. The case analysis shows that the proposed method can accurately evaluate the operational status of the secondary equipment, and provide guidance for the maintenance of the intelligent substation system and secondary equipment. This work is supported by the Science and Technology Project of State Grid Corporation of China (No. 521200190081).