Battery abnormality recognition method and device

By using a hierarchical physical constraint neural network model to quickly identify internal battery parameters, the problem of low efficiency in battery state detection is solved, enabling early identification and efficient detection of battery anomalies.

CN122109833APending Publication Date: 2026-05-29HANGZHOU ZHONGHEN ELECTRIC CO LTD +1
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
HANGZHOU ZHONGHEN ELECTRIC CO LTD
Filing Date
2026-01-13
Publication Date
2026-05-29

AI Technical Summary

Technical Problem

Existing technologies have low battery status detection efficiency, making it difficult to meet the real-time requirements of online monitoring in battery management systems.

Method used

A pre-trained hierarchical physical constraint neural network model, including a hard constraint model and a soft constraint model, is used to quickly identify the set of internal parameters of the battery by receiving current, voltage and sampled target time data. The parameters are adjusted using a particle swarm optimization algorithm until the error between the predicted voltage and the measured voltage is minimized. Combining the robustness of the hard constraint model and the high accuracy of the soft constraint model, battery anomaly identification is achieved.

Benefits of technology

It enables rapid and accurate battery status detection, improves the detection efficiency of the battery management system, and can identify battery anomalies at an early stage, meeting the rapid diagnostic needs of vehicle or energy storage systems.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122109833A_ABST
    Figure CN122109833A_ABST
Patent Text Reader

Abstract

The application discloses a battery abnormality identification method and device. The method comprises the following steps: receiving the power of a target battery; detecting the power data by using a pre-trained hierarchical physical constraint neural network model to obtain the predicted value of the battery internal parameter set of the target battery, wherein the hierarchical physical constraint neural network model is used for predicting the positive and negative electrode solid-liquid phase lithium ion concentration corresponding to the battery internal parameter set of the target battery, the positive and negative electrode solid-liquid phase lithium ion concentration corresponding to the battery internal parameter set of the target battery is used for determining the predicted voltage, and the value of the battery internal parameter set of the target battery corresponding to the predicted voltage is the predicted value of the battery internal parameter set of the target battery in the case that the error between the predicted voltage and the voltage value in the power data is minimum; and comparing the predicted value of the battery internal parameter set of the target battery with the threshold value of each parameter in the battery internal parameter set at the target time, and determining whether the target battery is abnormal according to the comparison result.
Need to check novelty before this filing date? Find Prior Art