Vehicle Battery Property Estimation With Adaptive Cell Modeling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In hybrid or electric vehicles, existing battery management systems face challenges in accurately estimating and predicting slowly varying battery parameters like impedance and capacity due to variations in cell properties over time and changing operating conditions, which affects state-of-charge estimation and battery health monitoring.
Innovation Solution
A method and system using a function approximator, such as a tile coding model, to estimate battery properties by training on measured parameters like current, temperature, and voltage, updating the model based on predetermined conditions, and adapting to changing battery properties over time and different operating conditions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a predetermined battery cell model is used to determine battery property values, then measurement precision is improved, but device complexity increases due to the need for continuous model training and updating
Solution Approach 1:
The function approximator automatically updates itself using measured battery parameter values from normal operation. The system performs self-training without external intervention, where the battery management system continuously refines the model using real-world data from voltage, current, and temperature measurements during vehicle operation.
Solution Approach 2:
The system changes the parameters of the function approximator based on measured battery behavior. By adjusting the approximator parameters using optimization algorithms (such as least squares or gradient descent) applied to measured data, the system adapts the model parameters to match actual battery characteristics under varying operating conditions.
2Reliability
If battery parameters are monitored continuously, then reliability is improved, but use of energy increases due to continuous measurement and processing
Solution Approach 1:
The system uses a function approximator that processes battery parameters in a simplified manner rather than full continuous analysis. By using an approximated model with predetermined conditions for updates, the system achieves adequate monitoring reliability without the excessive energy consumption of continuous full-model processing.
3Adaptability or versatility
If the function approximator is updated frequently, then adaptability is improved, but productivity decreases due to increased computational load
Solution Approach 1:
The function approximator is updated periodically based on predetermined conditions rather than continuously. The system determines whether to update by checking if specific conditions are met (such as sufficient new data accumulation or significant changes in battery behavior), creating a periodic update pattern that balances adaptability with computational efficiency.
Solution Approach 2:
The system uses feedback from measured battery parameter values to determine when updates are necessary. By monitoring the quality and quantity of measured data, the system provides feedback to the update mechanism, triggering updates only when the feedback indicates improved accuracy is needed, thus avoiding unnecessary computational operations.
Data Source
AI summary
The invention relates to a system for and method of determining a model for estimating battery properties in a hybrid or electrical drive system (900) for a vehicle (902), the method comprising: providing (51) a function approximator estimating a battery property as a function of two or more measured battery parameters; providing (52) a stream of measured battery parameter values as a function of time for a predetermined period to a predetermined battery cell model describing the relationship between the measured battery parameter and the battery property; determining (S3) a battery property value using the cell model based on a selected portion of the stream of measured battery parameter values; determining (S4) if a predetermined condition for the selected portion of the stream of measured battery parameter values is fulfilled; and if (S5) the predetermined condition is fulfilled, updating (S6) the function approximator based on the battery property value determined by the cell model.