Supervisory long-term prediction of state of available power for lithium-ion batteries in electric vehicles
Supervisory long-term prediction of state of available power for lithium-ion batteries in electric vehicles
复制标题
电动汽车锂离子电池可用电量状态的长期监管预测
DOI：
10.1016/j.apenergy.2019.114006
复制
发表时间：
2020
期刊：
影响因子：
11.2
通讯作者：
Deng Zhongwei
中科院分区：
文献类型：
--
作者：
Yang Lin;Cai Yishan;Yang Yixin;Deng Zhongwei
电池可用功率状态（SOAP）对于改善电动汽车（EV）的能量管理和保护电池免受损坏至关重要。本文提出了一种新的监督长期预测方案的SOAP电动汽车锂离子电池。监督长期预测表示SOAP是在EV的未来长期驾驶条件的监督下在线预测的，而不是在恒定的工作限制下的传统方法。首先，为准确捕捉电池动态特性，建立了包含多参数动态开路电压的电池模型，并采用带自适应遗忘因子的最小二乘法对电池参数进行在线辨识。提出了一种基于自适应两步滤波器的电池状态估计算法，以提高状态估计的精度。还建立了电动汽车电池长期功率需求（LTPD）预测模型。基于改进的电池模型和预测的电池状态，特别是在预测的LTPD的监督下，最后提出了新的监督长期电池SOAP预测方法，使预测的实用性和准确性。该算法对电池的长期荷电状态和SOAP进行在线联合预测。通过实验系统地评估了该方法对错误初始值、不同电池老化水平和环境温度的鲁棒性。实验结果表明，与传统方法相比，长期电池SOAP预测误差降低了85.9%。
The battery state of available power (SOAP) is crucial to improve the energy management of electric vehicles (EVs) and protect batteries from damage. This paper proposes a novel supervisory long-term prediction scheme of SOAP for lithium-ion batteries in electric vehicles. The supervisory long-term prediction denotes that the SOAP is online predicted under the supervision of the EV’s future long-term driving conditions, instead of the traditional approaches under the constant working limitations. Firstly, to accurately capture the battery dynamics, a battery model incorporated with multi-parameters dynamic open circuit voltage is established, and the least square approach with an adaptive forgetting factor is applied to online identify the battery parameters. A new battery state estimation algorithm based on an adaptive two step filter is then proposed to improve the accuracy of the state estimation. A battery’s long-term power demand (LTPD) prediction model is also established for EVs. Based on the improved battery model and predicted battery states, especially under the supervision of the predicted LTPD, the novel supervisory long-term battery SOAP prediction approach is finally put forward to make the prediction practical and accurate. The long-term state of charge (SOC) and SOAP of battery are online co-predicted by the derived algorithms. The robustness of the proposed approach against erroneous initial values, different battery aging levels and ambient temperatures is systematically evaluated by experiments. The experimental results verify the long-term battery SOAP prediction error reduced by 85.9% when compared with that by traditional approaches.
登录
查看更多内容
影响因子：
11.2
作者：
Linfeng Zheng;Jianguo Zhu;Guoxiu Wang;Tingting He;Yiying Wei
通讯作者：
Linfeng Zheng;Jianguo Zhu;Guoxiu Wang;Tingting He;Yiying Wei
影响因子：
11.2
作者：
Claudio Burgos-Mellado;M. Orchard;Mehrdad Kazerani;R. Cárdenas;D. Śaez
通讯作者：
Claudio Burgos-Mellado;M. Orchard;Mehrdad Kazerani;R. Cárdenas;D. Śaez
影响因子：
6.4
作者：
S. Panchal;I. Dincer;M. Agelin-Chaab;R. Fraser;M. Fowler
通讯作者：
S. Panchal;I. Dincer;M. Agelin-Chaab;R. Fraser;M. Fowler
影响因子：
11.2
作者：
M. Abbasi;M. Taki;Amin Rajabi;Li Li-Li;Jiangfeng Zhang
通讯作者：
M. Abbasi;M. Taki;Amin Rajabi;Li Li-Li;Jiangfeng Zhang
影响因子：
9.2
作者：
Roscher, Michael A.;Sauer, Dirk Uwe
通讯作者：
Sauer, Dirk Uwe