无迹卡尔曼滤波(unscented Kalman filter,UKF)是锂离子电池荷电状态(state of charge,SOC)估计的常用算法之一。然而在实际应用中,由于受到外界环境温度变化、电池容量退化等不确定性干扰,以及非高斯过程噪声的影响,需要进一步提高算...无迹卡尔曼滤波(unscented Kalman filter,UKF)是锂离子电池荷电状态(state of charge,SOC)估计的常用算法之一。然而在实际应用中,由于受到外界环境温度变化、电池容量退化等不确定性干扰,以及非高斯过程噪声的影响,需要进一步提高算法的性能才能更有效地保证估计精度。基于此,提出一种改进的无迹卡尔曼滤波算法(PO-RUKF)。首先,在UKF中引入H∞滤波提高算法的鲁棒性,用来克服各种干扰带来的不良影响。其次,利用鹦鹉优化算法对UKF的过程噪声协方差矩阵进行自适应调整,以解决滤波噪声参数先验确定的问题,从而提高滤波精度。最后,采用马里兰大学的FUDS和HPPC工况下的两种公开数据集进行了实验验证,结果表明,在不同的温度、电池容量退化状态以及不同的工况下,相比于传统的UKF算法以及鲁棒UKF算法,改进后的算法具有更高的SOC估计精度,平均绝对误差小于0.50%,均方根误差小于0.56%,此外还展现出更强的鲁棒性和普适性。证实所提方法可以为锂离子电池SOC估计提供更可靠、有效的技术支撑。展开更多
To address the limitations of the traditional parrot optimization algorithm(POA),such as susceptibility to local optima and slow convergence, a novel POA based on hybrid reverse learning and crisscross strategy(POABHR...To address the limitations of the traditional parrot optimization algorithm(POA),such as susceptibility to local optima and slow convergence, a novel POA based on hybrid reverse learning and crisscross strategy(POABHRLCS) was proposed in this paper. The key innovations of POABHRLCS include the following. First,Kent chaotic mapping is used for population initialization,enhancing the diversity of the initial population. Second,a hybrid reverse learning strategy combining lens imaging reverse learning and stochastic reverse learning is introduced to improve the algorithm's ability to escape local optima. Third,adaptive factors,including dynamic inertia weights and switching factors,are introduced to balance global exploration and local exploitation. Finally,a crisscross strategy employing horizontal and vertical crossover operations is used to maintain population diversity and prevent premature convergence. Extensive experiments on 23 benchmark functions demonstrate that POABHRLCS achieves faster convergence and higher solution accuracy compared to state-of-the-art metaheuristic algorithms.Furthermore,the algorithm outperforms other comparative algorithms in solving engineering constraint problems,such as the multi-disc clutch brake design and the three-bar truss volume optimization. These results confirm the practicality and effectiveness of POABHRLCS in both theoretical and real-world applications.展开更多
文摘无迹卡尔曼滤波(unscented Kalman filter,UKF)是锂离子电池荷电状态(state of charge,SOC)估计的常用算法之一。然而在实际应用中,由于受到外界环境温度变化、电池容量退化等不确定性干扰,以及非高斯过程噪声的影响,需要进一步提高算法的性能才能更有效地保证估计精度。基于此,提出一种改进的无迹卡尔曼滤波算法(PO-RUKF)。首先,在UKF中引入H∞滤波提高算法的鲁棒性,用来克服各种干扰带来的不良影响。其次,利用鹦鹉优化算法对UKF的过程噪声协方差矩阵进行自适应调整,以解决滤波噪声参数先验确定的问题,从而提高滤波精度。最后,采用马里兰大学的FUDS和HPPC工况下的两种公开数据集进行了实验验证,结果表明,在不同的温度、电池容量退化状态以及不同的工况下,相比于传统的UKF算法以及鲁棒UKF算法,改进后的算法具有更高的SOC估计精度,平均绝对误差小于0.50%,均方根误差小于0.56%,此外还展现出更强的鲁棒性和普适性。证实所提方法可以为锂离子电池SOC估计提供更可靠、有效的技术支撑。
基金supported by the National Key Research and Development Program of China(2022ZD0119000)
文摘To address the limitations of the traditional parrot optimization algorithm(POA),such as susceptibility to local optima and slow convergence, a novel POA based on hybrid reverse learning and crisscross strategy(POABHRLCS) was proposed in this paper. The key innovations of POABHRLCS include the following. First,Kent chaotic mapping is used for population initialization,enhancing the diversity of the initial population. Second,a hybrid reverse learning strategy combining lens imaging reverse learning and stochastic reverse learning is introduced to improve the algorithm's ability to escape local optima. Third,adaptive factors,including dynamic inertia weights and switching factors,are introduced to balance global exploration and local exploitation. Finally,a crisscross strategy employing horizontal and vertical crossover operations is used to maintain population diversity and prevent premature convergence. Extensive experiments on 23 benchmark functions demonstrate that POABHRLCS achieves faster convergence and higher solution accuracy compared to state-of-the-art metaheuristic algorithms.Furthermore,the algorithm outperforms other comparative algorithms in solving engineering constraint problems,such as the multi-disc clutch brake design and the three-bar truss volume optimization. These results confirm the practicality and effectiveness of POABHRLCS in both theoretical and real-world applications.
