Battery Parameter Estimation via Augmented Kalman Filter
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing battery management systems in hybrid electric vehicles (HEVs) face challenges in accurately estimating battery parameters, such as internal resistance and charge transfer impedance, leading to inefficiencies and reduced battery lifespan due to sensitivity to noise and variability in driving modes.
Innovation Solution
Implementing an extended Kalman filter with an augmented state vector that includes a time constant and proportionality factor between internal resistance and charge transfer impedance to improve observability and reduce parameter variability, allowing for precise battery model parameter estimation and control.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional battery management systems are used to estimate battery parameters, then the system structure is simple, but the measurement precision of battery parameters deteriorates due to sensitivity to noise and variability
Solution Approach 1:
The patent transforms the battery parameter estimation problem by changing the parameter representation through an augmented state vector that includes not only the original parameters (internal resistance, charge transfer impedance) but also their derivatives with respect to time. This parameter transformation enables the extended Kalman filter to track parameter variations more accurately while reducing sensitivity to measurement noise, thereby improving measurement precision without excessive complexity increase
Solution Approach 2:
The extended Kalman filter implements a feedback mechanism where the estimated battery parameters and their uncertainties are continuously updated based on new measurements. The filter uses the covariance matrix to weigh the reliability of measurements versus model predictions, providing optimal feedback-adjusted parameter estimates that reduce noise sensitivity while maintaining computational tractability
2Reliability
If battery parameters are estimated without considering observability, then the estimation process is simpler, but the reliability of parameter estimation deteriorates due to high sensitivity to noise
Solution Approach 1:
The patent applies parameter changes by augmenting the state vector to include derivatives of battery parameters with respect to time. This transformation improves observability because the time-derivative information provides additional constraints that make the parameter estimation more robust to noise. The extended state vector allows the filter to distinguish between actual parameter changes and measurement noise, thereby improving reliability
Solution Approach 2:
The patent introduces dynamics into the estimation process by modeling the time evolution of battery parameters through their derivatives. Instead of treating parameters as static, the augmented state vector captures their dynamic behavior, allowing the extended Kalman filter to track parameter changes while filtering out high-frequency noise. This dynamic approach enhances reliability without requiring excessive model complexity
3Measurement precision
If extended Kalman filter with augmented state vector is implemented, then the measurement precision of battery parameters is improved, but the device complexity increases
Solution Approach 1:
The patent changes the parameter representation by using an augmented state vector that includes parameter derivatives. This transformation improves measurement precision because the additional state variables provide more information for accurate estimation. The computational complexity increase is managed by efficiently formulating the Jacobian matrices and covariance updates specific to this augmented structure, rather than using generic high-complexity estimation methods
Data Source
AI summary
A vehicle includes a battery pack and at least one controller programmed to implement a model of the battery pack. The model includes parameters representing an internal resistance and charge transfer impedance of the battery pack, and identified from an extended Kalman filter having an augmented state vector. The augmented state vector is at least partially defined by a time constant associated with the charge transfer impedance and a variable representing a proportionality between the internal resistance and a resistance term of the charge transfer impedance to reduce observed variability of the parameters.