The invention discloses a power battery residual endurance prediction method based on a time sequence large model algorithm, and belongs to the technical field of battery life prediction, and the method comprises the following steps: S1, collecting vehicle static data features; s2, collecting vehicle operation dynamic data, wherein the time from the beginning of the previous charging to the beginning of the next charging is regarded as a complete cycle; s3, calculating the difference between two cycles; s4, using a sequence formed by the characteristics X (i) of a complete cycle from the beginning of the previous charging to the beginning of the next charging as the input of the training model; s5, using the driving kilometers of a complete cycle from the beginning of the previous charging to the beginning of the next charging as a target value of the training model; s6, carrying out important feature cleaning based on an XGBoost regression model; s7, a time sequence large language model Time-LLaMA is trained; and S8, predicting the remaining endurance time of the power battery by using the trained Time-LLaMA.
本发明公开了一种基于时间序列大模型算法的动力电池剩余续航预测方法,属于电池寿命预测技术领域,包括以下步骤:S1:采集车辆静态数据特征;S2:采集车辆运行动态数据:将上一次充电开始到下一次充电开始视为完整的一个循环;S3:计算两次循环间差分;S4:用上一次充电开始到下一次充电开始这样一个完整循环的特征X(i)组成的序列作为训练模型的输入;S5:用上一次充电开始到下一次充电开始这样一个完整循环的行驶公里数作为训练模型的目标值;S6:基于XGBoost回归模型进行重要特征清洗;S7:训练时间序列大语言模型Time‑LLaMA;S8:利用训练完成的Time‑LLaMA对动力电池剩余续航时间进行预测。
Power battery residual endurance prediction method based on time sequence large model algorithm
一种基于时间序列大模型算法的动力电池剩余续航预测方法
2024-04-09
Patent
Elektronische Ressource
Chinesisch
Power battery endurance mileage estimation system and method
Europäisches Patentamt | 2020
|Model based on time sequence analysis and residual matching
Europäisches Patentamt | 2021
|Unmanned aerial vehicle battery endurance flight capability prediction system
Europäisches Patentamt | 2023
|Rail vehicle traction battery remaining capacity endurance mileage prediction method and system
Europäisches Patentamt | 2020
|Vehicle endurance mileage prediction method and device
Europäisches Patentamt | 2023
|