Battery Temperature Prediction via Machine Learning Formula

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing methods for predicting the highest battery temperature are time-consuming and costly, often requiring physical tests on multiple candidate batteries, leading to inefficiencies and potential inaccuracies in temperature prediction.

Innovation Solution

An apparatus and method utilizing a preprocessing device and temperature determination device with machine learning capabilities to predict battery temperature, incorporating cell design parameters, simulation analysis conditions, and RC values through a temperature calculation formula derived via machine learning.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If physical tests and analysis simulations are performed on multiple candidate batteries to predict highest temperature, then prediction accuracy is improved, but time consumption and cost increase significantly

Engineering Contradiction:
Improvetemperature prediction accuracyVSAvoidprediction time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The system performs preliminary actions by pre-processing battery design data and extracting RC parameters before the actual temperature prediction is needed. Training data is prepared in advance, and the machine learning model is trained offline, so that when temperature prediction is required, the system can quickly query results without performing time-consuming physical tests or simulations.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The invention creates a virtual copy of the battery system through machine learning modeling. Instead of physically testing actual batteries, the system uses trained ML models that replicate battery thermal behavior based on design parameters. This virtual copying allows rapid prediction without consuming physical battery resources or time.

Inventive Principle:
Principle #26Copying

2Measurement precision

If physical tests and analysis simulations are performed on multiple candidate batteries to predict highest temperature, then prediction accuracy is improved, but manufacturing cost increases

Engineering Contradiction:
Improvetemperature prediction accuracyVSAvoidprediction cost
Core Design Contradiction:
Measurement precisionVSEase of manufacture

Solution Approach 1:

The system replaces expensive physical testing and simulation processes with a virtual machine learning model. Once the model is trained, predicting temperatures for multiple candidate batteries becomes a low-cost computational query rather than requiring physical prototypes or simulation resources.

Inventive Principle:
Principle #26Copying

Solution Approach 2:

The system changes the approach from physical/ computational experimentation to data-driven parameter analysis. By extracting RC parameters and using them as inputs to the ML model, the system transforms the prediction process into a parameter-based calculation that avoids the costs of physical testing and detailed simulations.

Inventive Principle:
Principle #35Parameter changes

3Productivity

If machine learning is used to predict battery temperature through trained models, then prediction time is reduced, but model training and data preparation complexity increases

Engineering Contradiction:
Improveprediction speedVSAvoidsystem complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The system segments the temperature prediction task into distinct phases: data preparation phase (extracting RC parameters from battery design data), model training phase (offline ML training), and prediction phase (quick querying). This segmentation allows the complex work to be done once during preparation, enabling rapid predictions afterward.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The machine learning model acts as an intermediary between battery design parameters and temperature predictions. Instead of directly simulating thermal physics, the ML model learns the relationship between design parameters (including RC values) and temperature outcomes, serving as a computational mediator that simplifies the prediction process.

Inventive Principle:
Principle #24Intermediary (Mediator)

Data Source

PatentEP4629127A1Apparatus and method for predicting maximum temperature
Publication Date: 2025.10.08 LG ENERGY SOLUTION LTD
  • EP4629127A1 patent drawingFigure 1
  • EP4629127A1 patent drawingFigure 2
  • EP4629127A1 patent drawingFigure 3

AI summary

An apparatus for predicting a highest temperature may include a preprocessing device including a first processor and a first memory storing a plurality of programs; and a temperature determination device including a second memory and a temperature prediction learning device. The first processor may generate input data including cell design parameters for a candidate battery cell constituting a candidate battery module and simulation analysis conditions for a simulation predicting the highest temperature for the candidate battery cell. The second memory may store the input data, and during a learning operation, the temperature prediction learning device may perform machine learning by using training data among the input data stored in the second memory, and determine a temperature calculation formula predicting the highest temperature of the candidate battery cell through machine learning.