Battery Cell State Estimation Using Correlated External Variables

Solution Overview

Problem

Existing battery management systems struggle to accurately estimate internal variables of a battery cell due to low correlation between internal and external variables, leading to inaccurate assessments of the battery's chemical state.

Innovation Solution

A battery management apparatus and method that utilizes a sub-multilayer perceptron to analyze correlations between observable external variables and unobservable internal variables, filtering and selecting only those with sufficient correlation for accurate estimation.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Quantity of substance

If all external variables are used to estimate internal variables, then more data is available for estimation, but the estimation accuracy decreases due to low correlation between unrelated variables

Engineering Contradiction:

Improvenumber of external variables usedVSAvoidestimation accuracy of internal variables

Core Design Contradiction:

Quantity of substanceVSMeasurement precision

Solution Approach 1:

The patent extracts and selects only those external variables that have a correlation coefficient greater than a predetermined threshold from the set of all external variables. This filtering process removes variables with low correlation, ensuring that only relevant external variables are used for estimating internal variables, thereby maintaining estimation accuracy while using an appropriate quantity of data.

Inventive Principle:

Principle #2Taking out (Extraction)

2Quantity of substance

If external variables with low correlation are included in the estimation model, then more observational data is utilized, but the internal variable estimation becomes greatly different from the actual chemical state

Engineering Contradiction:

Improveamount of observational data usedVSAvoidaccuracy of internal variable estimation

Core Design Contradiction:

Quantity of substanceVSReliability

Solution Approach 1:

The patent extracts and selects only those external variables that have a correlation coefficient greater than a predetermined threshold from the set of all external variables. This filtering process removes variables with low correlation, ensuring that only relevant external variables are used for estimating internal variables, thereby maintaining estimation accuracy while using an appropriate quantity of data.

Inventive Principle:

Principle #2Taking out (Extraction)

3Measurement precision

If correlation analysis is performed between all external and internal variables, then the most accurate model can be built, but the computational complexity and data processing requirements increase

Engineering Contradiction:

Improveaccuracy of internal variable estimationVSAvoidcomplexity of data processing system

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent extracts and selects only those external variables that have a correlation coefficient greater than a predetermined threshold from the set of all external variables. This filtering process removes variables with low correlation, ensuring that only relevant external variables are used for estimating internal variables, thereby maintaining estimation accuracy while using an appropriate quantity of data.

Inventive Principle:

Principle #2Taking out (Extraction)

Data Source

PatentUS12586827B2Battery management apparatus, battery management method and battery pack

Publication Date: 2026.03.24 LG ENERGY SOLUTION LTD

AI summary

There are provided a battery management apparatus, a battery management method and a battery pack. The battery management apparatus sets at least one of a plurality of external variables as a valid external variable for each internal variable using a plurality of observational data sets associated with the external variables that can be observed outside a battery cell and a plurality of desired data sets associated with the internal variables that are unobservable outside the battery cell. The observational data set associated with respective valid external variable is used for the machine learning of sub-multilayer perceptron necessary to estimate respective internal variable.