基于粒子滤波算法的质子交换膜燃料电池健康状态估计

State of Health Estimation for Proton Exchange Membrane Fuel Cells Based on Particle Filtering Algorithm

质子交换膜燃料电池的老化过程影响其输出性能,为了更精确地控制输出功率,需要考虑燃料电池的老化和功率衰退趋势。以功率-电流曲线作为健康状态指标,在以往研究的基础上进行改进,考虑开路电压在老化过程中的变化,增加老化模型中老化因子的个数,建立了质子交换膜燃料电池功率和内部部件老化之间的映射关系,基于极化曲线推导了半机理功率衰减模型,并基于正则化粒子滤波算法设计了老化率模型,结合功率衰减实现对燃料电池健康状态的估计。在测试数据集上进行了仿真试验,并和试验测试数据进行对比,结果表明,该方法能对长期性能衰减模型进行预测,相比于已有的研究方法,该方法能通过老化率参考值和功率衰减模型更准确地估计质子交换膜燃料电池的健康状态和性能衰减趋势,随着训练时间的缩短,估计精度较之提升,尤其是在训练时间长度100h,估计时间长度250h,误差相对下降率达65.69%。

The aging process of a proton exchange membrane fuel cell(PEMFC)affects its output performance,and in order to accurately control output power,it is necessary to consider the aging and power degradation trends of the PEMFC.In this paper,the power-current curve is used as an indicator of the state of health(SOH).Based on previous studies,improvements have been made by considering changes in open-circuit voltage during the aging process.The number of aging factors in the aging model has been increased and the mapping relationship between the PEMFC power and the aging of the internal components is established.A semi-mechanical power degradation model is derived based on polarization curves,and an aging rate model has been designed using the particle filter algorithm.Combining the power decay analysis,the paper estimated the fuel cell's SOH.Simulations were carried out on the test dataset and compared with experimental test data.The results show that the method can predict the long-term performance decay model.Furthermore,compared with existing research methods,the proposed method estimates the SOH and performance decay trend of PEMFCs more accurately through the use of aging rate reference values and the power decay model.With reduced training time,there is an improvement in estimation accuracy.Especially when the training time is 100 hours and the estimation time is 250 hours,the error's relative decrease rate reaches 65.69%.