Battery Cell Specification Using ML Parameter Matching

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Developers face challenges in specifying target battery cells due to the complexity of characteristics and the vast number of reference cells, making it difficult to determine relevant guidance for designing new cells, especially with the exponential growth of battery R&D data.

Innovation Solution

A computer-implemented method using a machine-learning model that estimates a matching score for reference battery cells with respect to target requirements, identifies relevant production parameters, and provides a degree of freedom for human developers to manually adjust further parameters, reducing computational complexity and improving interpretation.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Loss of information

If a machine-learning model identifies all relevant production parameters from a large number of reference cells, then the completeness of technology guidance is improved, but the computational complexity and difficulty of interpretation increase

Engineering Contradiction:
Improvecompleteness of technology guidanceVSAvoidcomputational complexity
Core Design Contradiction:
Loss of informationVSDevice complexity

Solution Approach 1:

The patent extracts only the most relevant production parameters from the large set of reference cell data using the machine-learning model. Instead of processing all parameters, the system identifies and extracts the critical subset that has the strongest correlation with target cell properties, thereby reducing computational complexity while maintaining guidance completeness.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The patent segments the large number of production parameters into hierarchical groups based on their relevance to target cell properties. The machine-learning model divides the parameter space into core parameters (high relevance) and secondary parameters (lower relevance), allowing the system to focus computational resources on the most important factors while maintaining interpretability.

Inventive Principle:
Principle #1Segmentation

2Measurement precision

If a machine-learning model processes extensive battery R&D data to provide comprehensive guidance, then the accuracy of target cell specification is improved, but the ease of operation for human developers deteriorates

Engineering Contradiction:
Improveaccuracy of target cell specificationVSAvoidease of interpretation for developers
Core Design Contradiction:
Measurement precisionVSEase of operation

Solution Approach 1:

The patent applies local quality by providing different levels of detail for different parameters. The machine-learning model identifies which parameters require precise numerical values versus which parameters can be specified with broader ranges or qualitative descriptions, allowing human developers to focus their attention on the most critical parameters while maintaining overall specification accuracy.

Inventive Principle:
Principle #3Local quality

3Extent of automation

If the model identifies a complete set of production parameters automatically, then the extent of automation is improved, but the degree of freedom for human developers to adjust parameters deteriorates

Engineering Contradiction:
Improveautomatic parameter identificationVSAvoiddegree of freedom for human adjustment
Core Design Contradiction:
Extent of automationVSAdaptability or versatility

Solution Approach 1:

The patent implements dynamics by creating a flexible parameter specification system where the machine-learning model provides automated recommendations but human developers can dynamically adjust any parameter. The system allows developers to override automated suggestions, modify parameter values based on additional considerations, and iteratively refine the target cell specification, thereby maintaining both automation efficiency and human adaptability.

Inventive Principle:
Principle #15Dynamics

Data Source

PatentEP4321883A1Method and system for specifying a target battery cell
Publication Date: 2024.02.14 CUSTOMCELLS HOLDING GMBH
  • EP4321883A1 patent drawingFigure 1~3
  • EP4321883A1 patent drawingFigure 4
  • EP4321883A1 patent drawing

AI summary

The invention relates to a method of specifying a target battery cell, comprising: providing, for at least one reference battery cell, a set of reference production parameters and at least one reference cell property, providing predefined target requirements defining at least one target cell property of the target battery cell, inputting the set of reference production parameters and the target requirements into a model, wherein in response the model: estimates a matching score of the reference battery cell with respect to the target battery cell, and/or identifies at least one target production parameter from the set of reference production parameters which is estimated to be relevant to achieve the target requirements.