铅酸动力电池高阶PNGV改进模型研究

Research on improved high order PNGV model for lead-acid power batteries

为了提高动态工况下铅酸动力电池荷电状态(SOC)的估算精度,针对现有Thevenin模型和PNGV模型仅有一个极化环节、动态工况下暂态响应精度不高的缺点,进行了铅酸动力电池等效电路模型的改进研究。在PNGV模型基础上增加极化环节,建立了改进高阶PNGV等效电路模型和对应的Simulink模型。采用高速可编程电子负载和电压电流数据采集卡搭建实验平台、以EVF-38铅酸动力电池为对象进行了不同倍率放电、脉冲放电和动态工况放电实验,根据采集的实验数据,运用遗传算法(GA)对改进模型极化环节参数在电流加载和卸载条件下分别进行了辨识。将动态工况下模型输出的仿真电压与实验实测电压进行比较,验证了模型的精度。改进后的高阶PNGV模型脉冲放电和动态工况仿真端电压最大相对误差分别为0.986%和2.155%,相对于原PNGV模型,动态工况实验仿真精度最大提高了1.313%,实验结果验证了铅酸动力电池高阶PNGV改进模型具有更好的准确性和动态性能。

For the existing Thevenin model and PNGV model that have only one polarization segment, and the accuracy is not high for the transient response characteristics of battery model under dynamic working conditions, the improvement research of the equivalent circuit model of lead-acid power battery was carried out to improve the SOC estimation accuracy. A Simulink model and a high order PNGV model were established, which improved by adding polarization segments based on the original PNGV model. An experimental platform was constructed by using high-speed programmable electronic load and voltage and current data acquisition card. Taking EVF-38 leadacid battery as the object, the experiments of different rate discharge, pulse discharge and dynamic condition discharge were carried out. Based on the experimental data, the GA algorithm is applied for model polarization segments parameters identification. The deferent parameters of polarization segments under loading and unloading conditions were calibrated respectively. The model accuracy was verified by comparing the simulated voltage of the model output with the experimental measured voltage under dynamic condition. The maximum relative errors of simulation terminal voltage using the improved PNGV model were 0. 986% and 2. 155% in the pulse discharge and dynamic discharge conditions. Compared with the original PNGV model, the simulation accuracy increased by up to 1. 313%. It was verified by the experimental results that the high order improved PNGV model has better accuracy and dynamic performance.