由于传统无迹卡尔曼滤波估算方法具有局限性,为了能准确估算动力电池荷电状态(state of charge,SOC),提出了一种基于无迹卡尔曼粒子滤波的动力电池SOC估算方法.以三元锂电池为研究对象,建立了电池二阶RC等效电路模型,通过对电池进行充...由于传统无迹卡尔曼滤波估算方法具有局限性,为了能准确估算动力电池荷电状态(state of charge,SOC),提出了一种基于无迹卡尔曼粒子滤波的动力电池SOC估算方法.以三元锂电池为研究对象,建立了电池二阶RC等效电路模型,通过对电池进行充放电试验辨识出模型参数,并验证模型准确性.采集了实际工况下的电池数据,分别用无迹卡尔曼滤波算法、粒子滤波算法和无迹卡尔曼粒子滤波算法估算电池SOC,在MATLAB中进行了仿真试验,并对估算的电池SOC进行比较.结果表明:无迹卡尔曼粒子滤波算法可以快速准确地估算出电池SOC,误差小于2.5%,优于另外2种算法.展开更多
为提高对动力电池的荷电状态(state of charge, SOC)估算精度、动力电池的健康状态(state of health, SOH)对锂电池性能的影响,提出一种扩展卡尔曼滤波(extended kalman filtering, EKF)联合估算算法。根据现有的实验数据,分析锂电池特...为提高对动力电池的荷电状态(state of charge, SOC)估算精度、动力电池的健康状态(state of health, SOH)对锂电池性能的影响,提出一种扩展卡尔曼滤波(extended kalman filtering, EKF)联合估算算法。根据现有的实验数据,分析锂电池特性,构建二阶RC等效电路模型,并进行参数辨识,搭建MATLAB仿真平台联合EKF算法进行SOC估算,将仿真结果与真实数据进行对比,结果表明,EKF联合估算SOC比EKF估算SOC误差精度约高1.2%,且抗干扰能力更强。展开更多
Lithium element has attracted remarkable attraction for energy storage devices, over the past 30 years. Lithium is a light element and exhibits the low atomic number 3, just after hydrogen and helium in the periodic t...Lithium element has attracted remarkable attraction for energy storage devices, over the past 30 years. Lithium is a light element and exhibits the low atomic number 3, just after hydrogen and helium in the periodic table. The lithium atom has a strong tendency to release one electron and constitute a positive charge, as Li<sup> </sup>. Initially, lithium metal was employed as a negative electrode, which released electrons. However, it was observed that its structure changed after the repetition of charge-discharge cycles. To remedy this, the cathode mainly consisted of layer metal oxide and olive, e.g., cobalt oxide, LiFePO<sub>4</sub>, etc., along with some contents of lithium, while the anode was assembled by graphite and silicon, etc. Moreover, the electrolyte was prepared using the lithium salt in a suitable solvent to attain a greater concentration of lithium ions. Owing to the lithium ions’ role, the battery’s name was mentioned as a lithium-ion battery. Herein, the presented work describes the working and operational mechanism of the lithium-ion battery. Further, the lithium-ion batteries’ general view and future prospects have also been elaborated.展开更多
锂电池的状态估计和主动均衡是提高电池性能和延长使用寿命的关键技术,针对参数模型的荷电状态(State of charge,SOC)估计方法忽略电动汽车实际工况而导致的估计偏差较大的问题,提出一种基于遗传算法的极限学习机(GA-ELM)神经网络算法...锂电池的状态估计和主动均衡是提高电池性能和延长使用寿命的关键技术,针对参数模型的荷电状态(State of charge,SOC)估计方法忽略电动汽车实际工况而导致的估计偏差较大的问题,提出一种基于遗传算法的极限学习机(GA-ELM)神经网络算法来估计电池的荷电状态SOC,通过遗传算法优化了ELM的参数,提高估计精度和泛化能力,并在UDDS工况数据下进行训练与测试。同时采用双向Buck-Boost均衡拓扑结构,该拓扑结构能够快速实现电池间的能量传递,同时又降低了传递路径的复杂性。通过遗传算法的极限学习机估计出的SOC作为均衡变量,利用Matlab/Simulink仿真平台进行试验。结果表明,提出的GA-ELM神经网络平均误差为0.15%,而传统的ELM神经网络平均误差为0.56%,因此提出的神经网络能够更精确地估计SOC;同时电池组之间能够快速完成能量均衡,证明了所提方案的可行性。展开更多
针对电池储能(battery energy storage system,BESS)平抑风电波动过程中电池单元荷电状态(state of charge,SOC)均衡性较差且未考虑风储净收益的问题,提出了风电波动平抑下考虑SOC均衡及收益的BESS功率分配策略。首先,建立综合考虑售电...针对电池储能(battery energy storage system,BESS)平抑风电波动过程中电池单元荷电状态(state of charge,SOC)均衡性较差且未考虑风储净收益的问题,提出了风电波动平抑下考虑SOC均衡及收益的BESS功率分配策略。首先,建立综合考虑售电收益、弃风惩罚、缺电惩罚及BESS运行成本等多个因素的风电并网指令优化模型,以并网指令波动率、电池组SOC标准差等多个因素为约束条件,提出改进算术优化算法(improved arithmetic optimization algorithm,IAOA)求解该优化模型。然后,将BESS划分为两个电池组,设计了BESS双层功率分配方法(double-layer power allocation method,DPAM),上层将BESS充放电指令分配给两个电池组,下层根据最大充放电功率原则或新型SOC均衡原则将电池组充放电指令分配给各自的电池单元。最后,通过仿真对所提策略进行了验证。仿真结果表明:IAOA加快了寻优速度,提高了寻优精度;DPAM提升了电池组内电池单元SOC的均衡速度,改善了均衡程度;提出的功率分配策略进一步降低了风电并网波动率,同时提高了风储系统净收益。展开更多
文摘由于传统无迹卡尔曼滤波估算方法具有局限性,为了能准确估算动力电池荷电状态(state of charge,SOC),提出了一种基于无迹卡尔曼粒子滤波的动力电池SOC估算方法.以三元锂电池为研究对象,建立了电池二阶RC等效电路模型,通过对电池进行充放电试验辨识出模型参数,并验证模型准确性.采集了实际工况下的电池数据,分别用无迹卡尔曼滤波算法、粒子滤波算法和无迹卡尔曼粒子滤波算法估算电池SOC,在MATLAB中进行了仿真试验,并对估算的电池SOC进行比较.结果表明:无迹卡尔曼粒子滤波算法可以快速准确地估算出电池SOC,误差小于2.5%,优于另外2种算法.
文摘为提高对动力电池的荷电状态(state of charge, SOC)估算精度、动力电池的健康状态(state of health, SOH)对锂电池性能的影响,提出一种扩展卡尔曼滤波(extended kalman filtering, EKF)联合估算算法。根据现有的实验数据,分析锂电池特性,构建二阶RC等效电路模型,并进行参数辨识,搭建MATLAB仿真平台联合EKF算法进行SOC估算,将仿真结果与真实数据进行对比,结果表明,EKF联合估算SOC比EKF估算SOC误差精度约高1.2%,且抗干扰能力更强。
文摘针对传统BP神经网络估算电池SOC过程中,存在初始权值和阈值对预测精度影响较大的问题,引入Tent混沌映射和自适应收敛因子对灰狼算法(GWO)进行改进,改善灰狼算法易陷入局部最优、后期迭代效率不高的缺点。将改进灰狼算法(improved grey Wolf algorithm,IGWO)与BP神经网络模型结合,得到BP神经网络最优初始权值和阈值,提高预测精度和收敛速度。对锂电池充放电实验数据预处理,得到样本数据。利用MATLAB进行仿真验证,结果表明,IGWO-BP神经网络算法的预测精度相较于传统BP神经网络算法、GWO-BP神经网络算法更优,基于改进灰狼优化BP神经网络估算电池SOC的方法的绝对误差能控制在1.53%以内,有效提高了预测精度和收敛速度。
文摘Lithium element has attracted remarkable attraction for energy storage devices, over the past 30 years. Lithium is a light element and exhibits the low atomic number 3, just after hydrogen and helium in the periodic table. The lithium atom has a strong tendency to release one electron and constitute a positive charge, as Li<sup> </sup>. Initially, lithium metal was employed as a negative electrode, which released electrons. However, it was observed that its structure changed after the repetition of charge-discharge cycles. To remedy this, the cathode mainly consisted of layer metal oxide and olive, e.g., cobalt oxide, LiFePO<sub>4</sub>, etc., along with some contents of lithium, while the anode was assembled by graphite and silicon, etc. Moreover, the electrolyte was prepared using the lithium salt in a suitable solvent to attain a greater concentration of lithium ions. Owing to the lithium ions’ role, the battery’s name was mentioned as a lithium-ion battery. Herein, the presented work describes the working and operational mechanism of the lithium-ion battery. Further, the lithium-ion batteries’ general view and future prospects have also been elaborated.
文摘锂电池的状态估计和主动均衡是提高电池性能和延长使用寿命的关键技术,针对参数模型的荷电状态(State of charge,SOC)估计方法忽略电动汽车实际工况而导致的估计偏差较大的问题,提出一种基于遗传算法的极限学习机(GA-ELM)神经网络算法来估计电池的荷电状态SOC,通过遗传算法优化了ELM的参数,提高估计精度和泛化能力,并在UDDS工况数据下进行训练与测试。同时采用双向Buck-Boost均衡拓扑结构,该拓扑结构能够快速实现电池间的能量传递,同时又降低了传递路径的复杂性。通过遗传算法的极限学习机估计出的SOC作为均衡变量,利用Matlab/Simulink仿真平台进行试验。结果表明,提出的GA-ELM神经网络平均误差为0.15%,而传统的ELM神经网络平均误差为0.56%,因此提出的神经网络能够更精确地估计SOC;同时电池组之间能够快速完成能量均衡,证明了所提方案的可行性。
文摘针对电池储能(battery energy storage system,BESS)平抑风电波动过程中电池单元荷电状态(state of charge,SOC)均衡性较差且未考虑风储净收益的问题,提出了风电波动平抑下考虑SOC均衡及收益的BESS功率分配策略。首先,建立综合考虑售电收益、弃风惩罚、缺电惩罚及BESS运行成本等多个因素的风电并网指令优化模型,以并网指令波动率、电池组SOC标准差等多个因素为约束条件,提出改进算术优化算法(improved arithmetic optimization algorithm,IAOA)求解该优化模型。然后,将BESS划分为两个电池组,设计了BESS双层功率分配方法(double-layer power allocation method,DPAM),上层将BESS充放电指令分配给两个电池组,下层根据最大充放电功率原则或新型SOC均衡原则将电池组充放电指令分配给各自的电池单元。最后,通过仿真对所提策略进行了验证。仿真结果表明:IAOA加快了寻优速度,提高了寻优精度;DPAM提升了电池组内电池单元SOC的均衡速度,改善了均衡程度;提出的功率分配策略进一步降低了风电并网波动率,同时提高了风储系统净收益。
