为了能更准确的估算锂离子电池的健康状态(SOH)以进行更换以及降级使用等措施来防止危险事故的发生,文章采用了基于改进萤火虫算法优化的卷积神经网络模型,实现高精度的锂离子电池SOH估算。用NASA随机步长电池数据集进行分析并提取电池健康因子,采用改进萤火虫算法对卷积神经网络的权值进行优化,搭建电池SOH估算模型,并加入高斯白噪声对输入数据进行数据增强,然后与传统的算法进行对比仿真分析。结果验证了本方法的有效性及高精度性,有利于维护大型矿卡的电池健康状态并及时更换提高矿卡运输安全性。

In order to more accurately estimate the state of health ( SOH) of lithium - ion batteries for replacement and downgraded use, measures are taken to prevent dangerous accidents, a convolutional neural network model based on improved firefly algorithm optimization was adopted to achieve high-precision estimation of lithium-ion battery SOH. The NASA random step battery dataset was used to analyze and extract the battery health factor, the weight of the convolutional neural network was optimized using the improved firefly algorithm, the battery SOH estimation model was built, and Gaussian white noise was added to enhance the input data. Then, compared with the traditional algorithm, the results verify the effectiveness and high accuracy of the proposed method, which is conducive to maintaining the battery health status of large mining trucks and replacing them in time to improve the transportation safety of trucks.