Battery SOH Checking Using AI-Predicted Voltage Curves
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for evaluating the state of health (SOH) of batteries are inefficient and time-consuming, making it difficult to accurately assess battery health during continuous use, which is crucial for electric vehicles, portable electronics, and energy storage systems.
Innovation Solution
An apparatus and method using pre-trained AI learning models to rapidly and accurately evaluate SOH by analyzing charge and discharge characteristics of a test battery, incorporating a discharge unit, charge unit, data collection, sensor, and prediction units to generate and predict voltage curve data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a full charge/discharge test is performed to accurately measure SOH, then measurement precision is improved, but loss of time increases significantly
Solution Approach 1:
The patent applies preliminary action by pre-training AI learning models offline using comprehensive full charge/discharge test data. Once trained, the model can rapidly predict SOH during actual use without requiring time-consuming full tests. The pre-processing unit also performs preliminary data cleaning and normalization before feeding data to the AI model, further accelerating the evaluation process while maintaining accuracy.
Solution Approach 2:
The patent uses copying by creating a virtual model of the battery's charge/discharge behavior through AI learning. Instead of physically performing full charge/discharge cycles, the system copies the essential characteristics through the trained AI model that predicts voltage curves and SOH values. This virtual copy enables rapid assessment without the time cost of actual full cycles.
2Productivity
If SOH evaluation is performed during continuous use, then productivity is improved, but measurement precision deteriorates due to incomplete charge/discharge cycles
Solution Approach 1:
The system performs preliminary action by pre-training the AI model with comprehensive full charge/discharge data before deployment. This pre-training enables the model to accurately interpret partial charge/discharge data encountered during continuous use, maintaining measurement precision while enabling online evaluation during normal operation.
Solution Approach 2:
The patent applies dynamics by adapting the evaluation method to work with varying charge/discharge states during continuous use. The AI model dynamically adjusts predictions based on the current state of charge and discharge characteristics, enabling accurate SOH assessment regardless of whether the battery is in a partial or complete charge/discharge cycle.
3Loss of time
If rapid SOH evaluation is implemented, then loss of time is reduced, but measurement precision deteriorates
Solution Approach 1:
The patent resolves this contradiction through preliminary action by performing extensive model training and data pre-processing before actual SOH evaluation. The pre-trained AI model and pre-processed data enable rapid predictions without sacrificing accuracy, as the computationally intensive work has already been completed in advance.
Solution Approach 2:
The patent replaces the mechanical system of physical charge/discharge cycling with an information-based AI prediction system. Instead of physically charging and discharging the battery to assess SOH, the system uses the trained AI model to predict SOH from voltage data, dramatically reducing time while maintaining or improving precision through advanced algorithms.
Data Source
AI summary
An apparatus for checking a state of health (SOH) includes a discharge unit that discharges a test battery, a charge unit that charges the test battery by providing charge power, a data collection unit that selects points in the discharge or charge process and collects input data in the discharge or charge process or at the selected points, a sensor unit that senses a voltage of the test battery, a data generation unit that receives the input data and the voltage and generates charge voltage curve data or discharge voltage curve data of the test battery, a first prediction unit that predicts the discharge voltage curve data based on the charge voltage curve data by using a pre-trained first AI learning model, and a second prediction unit that checks the SOH of the test battery based on the discharge voltage curve data by using a pre-trained second AI learning model.