【摘要】 电池剩余寿命(RUL)预测对电池健康管理至关重要。该文针对传统电池RUL预测过于依赖电池容量而容量数据难以直接获取的问题,提出对可直接在线测量的电池放电电压进行小波包能量熵(WPEE)提取替代容量表征电池退化。此外,利用提取的WPEE建立分数阶灰色退化模型(FGM)融合改进无迹粒子滤波(AUPF),构建FGM-AUPF算法框架,最终实现电池RUL预测。实验结果表明,该文构建的FGM-AUPF算法框架分别利用电池放电电压WPEE和容量作为退化表征量,均能准确地预测电池RUL,且用电池放电电压WPEE预测得到的结果相对误差不大于5.96%。更多还原

【Abstract】 Accurate prediction of battery remaining useful life(RUL) is one of the key technologies of battery management systems. The traditional methods depend heavily on the battery capacity and the capacity data is difficult to obtain directly. Therefore, the wavelet packet energy entropy(WPEE) extracted from the battery discharge voltage was proposed to replace the capacity charactering the battery degradation. Then the extracted WPEE was used to construct a fractional grey model(FGM), and the model was applied to be fused with the adaptive unscented particle filter(AUPF) for realizing battery RUL prediction. The experimental results show both the battery discharge voltage WPEE and battery capacity can be used as the degradation characterization indicator under the proposed FGM-AUPF algorithm framework to achieve battery RUL prediction. And the relative error of the battery discharge voltage WPEE prediction results is no more than 5.96%.更多还原