Battery SoC Estimation with Sectioned ECT Parameter Updates
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing battery state-of-charge estimation methods face challenges in maintaining accuracy due to battery cell deterioration and the need for frequent parameter updates, which can lead to increased computational operations and reduced update speed.
Innovation Solution
A battery optimization method that uses an electrochemical thermal (ECT) model to segment state of charge (SoC) into sections, selectively adjusts parameters based on operating data, and updates the ECT model parameters to minimize optimization loss, employing Bayesian optimization to adjust diffusion parameters and maintain accurate SoC estimation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If battery parameters are frequently updated to maintain SoC estimation accuracy, then estimation accuracy is improved, but computational operations increase and update speed decreases
Solution Approach 1:
The patent segments the battery parameter update process into two distinct parts: diffusion parameters that are frequently updated using Bayesian optimization, and other parameters (such as film resistance, active material volume fraction, and OCP offset) that are updated less frequently. This segmentation allows the system to maintain high SoC estimation accuracy by frequently updating the most critical parameters while avoiding the computational burden of updating all parameters equally often, thus resolving the contradiction between accuracy and update speed.
2Reliability
If all ECT model parameters are updated frequently, then model accuracy is improved, but computational complexity increases
Solution Approach 1:
The patent applies local quality by treating different ECT model parameters with different update frequencies based on their individual characteristics and importance. Diffusion parameters receive frequent updates with high computational resources, while other parameters like film resistance and active material volume fraction receive less frequent updates. This localized differentiation of update strategies optimizes model accuracy for each parameter type while reducing overall computational complexity.
Solution Approach 2:
The patent implements parameter changes by dynamically adjusting the update frequency and methodology for different ECT model parameters. Specifically, diffusion parameters are updated using Bayesian optimization which adapts the update process based on observed data quality and model performance, while other parameters are updated using less computationally intensive methods. This adaptive parameter change strategy maintains model accuracy while managing computational complexity.
3Measurement precision
If Bayesian optimization is used to adjust diffusion parameters, then training accuracy is improved, but operational burden increases
Solution Approach 1:
The patent segments the parameter optimization task by applying Bayesian optimization specifically to diffusion parameters, which are identified as the most critical for SoC estimation accuracy. Other parameters are updated using simpler methods or less frequently. This segmentation concentrates the computational effort of Bayesian optimization on the parameters that provide the greatest accuracy improvement, thereby reducing the overall operational burden while maintaining high training accuracy.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A battery optimization method and apparatus is provided. The battery optimization method includes selectively adjusting a parameter set related to a corresponding SoC section among a plurality of parameters of an electrochemical thermal (ECT) model based on operating data of each SoC section, and updating the plurality of parameters of the ECT model based on adjusted parameter sets.