基于深度学习的SESN供电网络电能质量评估

Power Quality Assessment of SESN Power Supply Network Based on Deep Learning

为保证超大规模储能供电网络安全可靠地运行,提出一种基于深度学习的电能质量评估方法。首先,依据电能质量国家标准,提出电能质量评估指标5等级划分法,建立基于深度置信网络(DBN)的电能质量评估模型。然后,考虑储能供电网络的构成、运行特征和储能可用容量,将储能系统荷电状态(SOC)预估结果作为其电能质量评估的另一指标,构建基于DBN的超大规模电池储能供电网络系统电能质量评估仿真模型。最后,按照电能质量指标等级分类对模拟数据做标准化处理,构造基于DBN的电能质量评估训练数据集。通过将结果与其他评估方法进行比较,验证了该方法的优越性。

In order to ensure the safe and reliable operation of the super-large-scale energy storage power supply network,a power quality assessment method based on deep learning is proposed.Firstly,according to the national standard of power quality,a method of 5-level division for power quality assessment indices is proposed,and a power quality assessment model based on deep belief network(DBN)is established.Then,considering the composition,operation characteristics and energy storage available capacity of energy storage power supply network,the SOC estimated result of energy storage system is taken as another index for power quality assessment,a simulation model for power quality assessment of super-large-scale energy storage power supply network system based on DBN is constructed.Finally,the simulation data are normalized according to the classification of power quality indices,and the training data set of power quality assessment based on DBN is formed.The advantages of this method are shown by comparing the results with other assessment methods.