The Combined Prediction Method Of Lithium-ion Battery Life Based on CEEMDAN-LSTM
-
摘要: 锂离子电池进行准确的寿命预测在电池健康管理中具有重要意义。针对目前锂离子电池寿命预测方法预测结果不准确的问题,文中在单一LSTM预测模型的基础上,采用了自适应噪声完全集成的经验模态分解(CEEMDAN)算法将容量分为主退化趋势和若干局部退化趋势,然后使用长短期记忆神经网络(LSTMNN)算法分别对所分解的若干退化数据进行容量预测,最后将若干容量预测结果进行叠加组合后得到锂离子电池的寿命预测结果。验证结果表明,所提出的CEEMDAN-LSTM锂离子电池组合预测模型较之EMD-LSTM预测模型平均预测精度提高了4.86%。
-
关键词:
- 电池健康管理 /
- 锂离子电池 /
- 剩余使用寿命 /
- 长短期记忆神经网络 /
- 自适应噪声完全集成经验模态分解
Abstract: Accurate life prediction of lithium-ion batteries is of great significance in battery health management.Based on the current lithium ion battery life prediction method to predict the results the problem of inaccurate, LSTM single forecast model is presented in this paper, on the basis of using the adaptive noise fully integrated algorithm of empirical mode decomposition (CEEMDAN) points is given priority to the capacity degradation trends and some partial degradation, and then use the Long and short term memory neural network (LSTMNN) algorithm respectively volume projections for the decomposition of degradation data, finally several capacity prediction results after superposition combination of lithium ion battery life prediction results.Verification results show that the proposed CEEMDAN-LSTM combination prediction model improves the average prediction accuracy by 4.86% compared with the EMD-LSTM prediction model. -

计量
- 文章访问数: 688
- HTML全文浏览量: 69
- 被引次数: 0