基于CNN和LSTM混合网络的电动汽车充电桩运行状态预测方法

随着电动汽车的大规模发展,公共充电桩运行数量和充电量逐年增长。然而,充电桩运行始终存在故障频发、运维难度大和维修成本高等问题,并且传统故障检测方法效率低下。因此提出了一种基于卷积神经网络(CNN)和长短期记忆(LSTM)网络的混合网络电动汽车充电桩运行状态预测方法,可以实现对电动汽车充电桩运行状况的综合评估。在特征数据输入阶段,对充电桩运行状态的关键指标进行分析,通过CNN提取运行状态影响因素的特征量,再利用LSTM判断和预测充电桩运行状态,从而实现对充电桩潜在故障的预警。试验结果表明,该方法预测准确率高、实用性强,能较准确地反映和预测充电桩的运作状态,可实际用于充电桩故障预测与运维检修。

Prediction Method of Electric Vehicle Charging Pile Operating State Based on CNN and LSTM Hybrid Network

With the large-scale development of electric vehicles, the number of public charging piles in operation and the charging capacity are increasing year by year. However, there are many problems in the operation of charging pile, such as frequent failures, difficult operation and high maintenance costs, and traditional fault detection methods are inefficient. Therefore, a hybrid network prediction method of electric vehicle charging pile operation state based on convolution neural network (CNN) and long short term memory (LSTM) network is proposed, which can realize the comprehensive evaluation of the operation state of electric vehicle charging piles. In the feature data input stage, the key indicators of the operation state of the charging pile are analyzed, and the characteristic quantities of the influencing factors of the operation state are extracted through CNN. Then, the operation state of the charging pile is judged and predicted by LSTM, so as to realize the early warning of potential failures of the charging pile. The experimental results show that the method has high prediction accuracy and strong practicability. It can accurately reflect and predict the operation state of charging piles, and can be used in actual charging pile fault prediction and operation and maintenance.