A lithium battery state of charge prediction method based on deep learning
By using deep learning-based methods, combined with electrochemical principles and data-driven approaches, a lithium battery state-of-charge (SOC) prediction model was constructed. This model solves the problem of inaccurate prediction in existing technologies, achieves more accurate SOC prediction, and improves the performance of the battery management system and battery life.
CN122410318APending Publication Date: 2026-07-17BEIJING HYPERSTRONG TECH CO LTD
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- BEIJING HYPERSTRONG TECH CO LTD
- Filing Date
- 2026-04-28
- Publication Date
- 2026-07-17
AI Technical Summary
Technical Problem
Existing technologies lack lithium battery state-of-charge prediction models that combine electrochemical principles with data-driven methods, leading to inaccurate predictions.
Method used
A deep learning-based approach is used to predict SOC by measuring the initial SOC value, obtaining dynamic SOC data through ampere-hour integration, and combining parameters such as total current, individual cell voltage, and temperature to construct a deep learning neural network model.
Benefits of technology
It enables more accurate prediction of lithium battery state of charge, improves the accuracy and safety of the battery management system, extends battery life, optimizes charging and discharging strategies, and reduces operating costs.
✦ Generated by Eureka AI based on patent content.
Smart Images
Figure CN122410318A_ABST
Abstract
本发明公开了一种基于深度学习的锂电池荷电状态预测方法，涉及荷电状态预测技术领域，本发明对单体锂电池在充分放电后静置时，通过测量电池电压并依据预先建立的电压与SOC对应表确定初始SOC值，作为模型训练样本；同时采用安时积分法对充放电过程中的电流数据进行积分，计算出电池动态SOC值，作为标签数据；在不同SOC状态下记录电池运行参数，并对原始数据进行噪声滤除、异常值剔除及插补处理；基于上述处理数据，本发明利用分类预测或回归预测算法构建SOC预测模型，其构建方案可选用深度学习神经网络、XGBoost、随机森林或支持向量机等算法；通过训练得到的模型对电池单体的SOC进行在线预测，并将预测结果记录于数据管理系统中。
Need to check novelty before this filing date? Find Prior Art