Battery SOC-OCV Estimation from Rest-Period Operation Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Accurately determining the full charge capacity (FCC) of secondary batteries in electric vehicles is costly and difficult for business operators due to the lack of accessible SOC-OCV curves, necessitating costly data collection methods.
Innovation Solution
A calculation system that acquires operation data from multiple time points, extracts sample data during battery rest periods, and estimates the SOC-open circuit voltage (OCV) characteristic using a management device to generate the SOC-OCV curve without requiring initial data acquisition.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If business operators collect operation data to create SOC-OCV curves for accurate FCC determination, then measurement precision improves, but loss of time and manufacturing cost increase significantly
Solution Approach 1:
The patent uses machine learning models to copy and replicate the SOC-OCV characteristics of batteries without physically collecting extensive operation data. The model learns from limited data and generates accurate SOC-OCV curves, eliminating the need for time-consuming data collection processes while maintaining measurement precision.
Solution Approach 2:
The patent performs preliminary actions by pre-training machine learning models with battery characteristics data before actual FCC determination is needed. This preliminary model training enables rapid SOC-OCV curve generation without requiring real-time data collection, thus reducing time loss while maintaining accuracy.
2Measurement precision
If business operators collect operation data to create SOC-OCV curves for accurate FCC determination, then measurement precision improves, but manufacturing cost increases
Solution Approach 1:
The patent replaces expensive physical data collection processes with machine learning-based copying of battery characteristics. The model replicates SOC-OCV curves computationally, eliminating costs associated with acceleration tests, data collection infrastructure, and manual processing while maintaining high measurement precision.
Solution Approach 2:
The patent substitutes mechanical data collection methods (acceleration tests, physical measurement systems) with computational machine learning approaches. This replacement eliminates the need for expensive test equipment, test vehicles, and manual data gathering processes, significantly reducing manufacturing costs while preserving accuracy.
3Measurement precision
If SOC-OCV curves are acquired through acceleration tests or data collection, then estimation accuracy improves, but device complexity and operational complexity increase
Solution Approach 1:
The patent uses machine learning models to copy battery SOC-OCV characteristics without requiring complex data collection systems. The model internally represents battery behavior, eliminating the need for complex measurement equipment, test protocols, and data management infrastructure while maintaining estimation accuracy.
Solution Approach 2:
The patent enables the battery management system to self-determine SOC-OCV characteristics using the machine learning model without external data collection infrastructure. The model processes available data autonomously and generates accurate curves, reducing system complexity by eliminating dedicated data collection hardware and manual processing systems.
Data Source
AI summary
In a calculation system, a data acquisition unit is configured to acquire operation data of a battery, the operation data including at least voltages and currents measured by a management device at a plurality of time points, and states of charge (SOCs) estimated based on at least one of the voltages and the currents, the management device being configured to manage the battery. An extraction unit is configured to, extract, as sample data, a set of an SOC and a voltage from sets of the SOCs and the voltages the plurality of time points included in the operation data in a period in which the battery is regarded as resting, the period being specified based on the currents. An estimation unit is configured to estimate an SOC-open circuit voltage (OCV) characteristic of the battery based on the extracted sample data.