Battery Surface SOC Estimation for Dynamic Voltage Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing battery state estimation methods, particularly using equivalent circuit models, struggle to accurately reflect dynamic changes in battery states during charging or discharging, leading to increased estimation errors.
Innovation Solution
A battery management system that estimates a surface state of charge (SOC) based on measured current, temperature, and historical SOC data, using models to determine coefficients and parameters that account for reaction rates and diffusion resistances at the electrode surface, allowing for precise estimation of terminal voltage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If an equivalent circuit model using SOC is used to estimate terminal voltage, then the estimation is suitable for static battery states (open circuit), but the current effect cannot be reflected in dynamic states (continuous charging/discharging), leading to increased estimation error
Solution Approach 1:
The patent combines the equivalent circuit model (which performs well in static states) with a neural network model (which captures dynamic current effects) into a hybrid estimation system. The neural network processes current measurements and SOC values to generate corrections that are applied to the equivalent circuit model outputs, enabling accurate terminal voltage estimation in both static and dynamic battery states.
Solution Approach 2:
The patent creates a composite estimation model that integrates two different modeling approaches: the physics-based equivalent circuit model and the data-driven neural network model. This composite model leverages the strengths of both approaches, where the equivalent circuit model provides the baseline estimation and the neural network provides dynamic corrections based on current effects, achieving high accuracy across all battery operating conditions.
2Ease of manufacture
If the equivalent circuit model is used, then the model structure is simple and easy to implement, but it cannot accurately reflect current effects during continuous charging or discharging
Solution Approach 1:
The patent divides the estimation task into two segments: the equivalent circuit model handles the baseline terminal voltage estimation and is easy to implement, while the neural network model specifically handles the current effects correction. This segmentation allows each component to be optimized independently, maintaining the simplicity of the equivalent circuit model while adding the accuracy of the neural network for dynamic conditions.
Solution Approach 2:
The neural network acts as an intermediary that processes current measurements and SOC values to generate correction terms. These corrections are then applied to the equivalent circuit model outputs, serving as a bridge that transfers the neural network's ability to capture dynamic effects to the simple equivalent circuit model structure.
Data Source
AI summary
A battery apparatus receives a measured current of a battery and estimates a surface SOC representing a potential at an electrode surface of the battery based on a plurality of parameters. The plurality of parameters includes a first parameter determined based on the measured current and a second parameter determined based on an SOC of the battery.