The invention relates to the technical field of artificial intelligence, and particularly discloses a battery life estimation method based on AI deep learning, and the method comprises the steps: carrying out the accurate estimation of the SOH of a battery through a physical parameter model based on the battery, obtaining an SOH estimation value, carrying out the accurate estimation of an SOC through a hybrid model based on the battery, and obtaining an SOC estimation value, finally, according to the SOH estimated value and the SOC estimated value, an AI/ML deep learning architecture is used for conducting residual useful life prediction, the complete battery predicted life cycle is obtained, the RUL is estimated through the reference energy value calculated through the method based on a physical model, the RUL estimation accuracy and precision are improved, the second-layer AI/ML mixed algorithm tracks parameters changing along with time, and the estimation accuracy and precision of the RUL are improved. According to the system, the driving mode, the battery running time and the endurance mileage are provided, wireless flashing, calibration and updating are allowed, and the customer satisfaction degree is improved to the maximum extent, so that the driving mode, the battery running time and the endurance mileage are understood intelligently.
本发明涉及人工智能技术领域,具体公开了一种基于AI深度学习的电池生命预估方法,先通过基于电池的物理参数模型对电池SOH准确估算,获得SOH估算值,再通过基于电池的混合模型进行SOC准确预估,获得SOC估算值,最后根据SOH估算值和SOC估算值使用AI/ML深度学习架构进行剩余有用寿命预测,获得完整电池预测生命周期,通过采用基于物理模型的方法计算的参考能量值来估计RUL,提高了RUL的估算准确性和精度,其中的第二层AI/ML混合算法,它跟踪随时间变化的参数,并应用智能来了解它将以多快的速度降级或达到使用寿命,提供了对驾驶模式、电池运行时间和续航里程的洞察,并允许无线刷写校准更新,最大限度的提高客户满意度。
Battery life estimation method based on AI deep learning
一种基于AI深度学习的电池生命预估方法
2023-05-30
Patent
Electronic Resource
Chinese
IPC: | B60L PROPULSION OF ELECTRICALLY-PROPELLED VEHICLES , Antrieb von elektrisch angetriebenen Fahrzeugen |
BATTERY LIFE ESTIMATION METHOD AND BATTERY LIFE ESTIMATION APPARATUS
European Patent Office | 2016
|Battery aging estimation with deep learning
IEEE | 2017
|European Patent Office | 2015
|