摘要
针对电动直臂车的特殊工况,提出了一种基于双向长短期记忆神经网络(Bi-LSTM)的电动直臂车荷电状态(SOC)估计模型和方法。该方法将电池的工作电压、电流及表面温度作为输入,采用双向传递的两层LSTM神经网络进行训练,再将两次得到的结果进行拼接作为最终输出。实验结果表明,该方法比传统前馈(BP)神经网络和单向LSTM神经网络具有更好的估计性能,并且可以精确估计不同环境温度下的电池及整车SOC。
According to the special operating condition of electric telescopic boom aerial work platform,a novel state of charge(SOC)prediction model is proposed based on bi-directional long short-term memory(Bi-LSTM)neural network.Voltage,current and surface temperature of battery are considered as inputs,and the forward LSTM layer and the backward LSTM layer can be used to provide complete timing information for the output.Experimental results show that the proposed model improves estimated performance compared with traditional back propagation(BP)neural network model and one-directional LSTM model.Furthermore,the proposed model can accurately estimate the SOC of battery and vehicle under different ambient temperatures.
作者
张艺迪
孙晖
ZHANG Yi-di;SUN Hui(Zhejiang University,College of Electrical Engineering,Hangzhou 310027,China)
出处
《能源工程》
2023年第1期37-42,共6页
Energy Engineering
