摘要
针对现有车载锂电池荷电状态(SOC)估算方法存在误差大、动态模型建立困难及模型适应性差等问题,文章提出一种采用动态神经网络(NARX)来实时估算车载锂电池SOC的方法。首先由改进人工鱼群算法寻优得到最佳NARX的结构参数,以建立初步的锂电池SOC估算模型;其次根据不同工况更新模型中的结构参数,增强估算模型自适应性。估算模型的输入分别为锂电池的电压、电流、欧姆内阻,其中锂电池的欧姆内阻是依据电池模型通过递推最小二乘法在线辨识获取,随温度和电池健康状态(SOH)实时变化。在两种典型工况下做车载锂电池放电试验,并采用两种不同锂电池来验证估计模型的适应性,结果表明该方法的预测精度高、适应性好,两种工况下的估计误差均低于0.045,两种锂电池下的估计误差均不超过0.043。
Aiming at the problems of large errors,difficulties in establishing dynamic models and poor adaptability of the existing methods for estimating SOC of on-board lithium batteries,a method for real-time estimation of SOC of on-board lithium batteries is proposed by using dynamic neural network(NARX).Firstly,the optimal NARX structural parameters were obtained by the optimization of the improved artificial fish swarm algorithm to establish a preliminary lithium battery SOC estimation model.Secondly,the structural parameters in the model were updated according to different working conditions to enhance the adaptability of the estimation model.The input of the estimation model is voltage,current and ohmic resistance of lithium battery respectively,among which the ohmic resistance of lithium battery is obtained online through recursive least-squares method according to the battery model,and changes in real time with temperature and battery health state(SOH).The discharge test of on-board lithium battery was carried out under two typical working conditions,and different lithium batteries were used to verify the adaptability of the estimation model.The results show that the prediction accuracy of this method was high and the adaptability was good.The estimation error under two working conditions was lower than 0.045,and the estimation error under two lithium batteries was less than 0.043.
作者
陈德海
邹争明
丁博文
华铭
CHEN Dehai;ZOU Zhengming;DING Bowen;HUA Ming(School of Electrical Engineering and Automation,Jiangxi University of Science and Technology,Ganzhou 341000,China)
出处
《江西理工大学学报》
CAS
2020年第3期96-102,共7页
Journal of Jiangxi University of Science and Technology
基金
江西省自然科学基金资助项目(20151BAB206034)。
关键词
车载锂电池
荷电状态
NARX
人工鱼群算法
递推最小二乘法
on-board lithium battery
state of charge
NARX
artificial fish swarm algorithm
recursive least-squares method
