Distributed Battery State Estimation for Real-Time Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current battery management systems face challenges in accurately estimating battery states, particularly in real-time, especially on compute-limited resources, and in validating models to ensure precision, while also optimizing the useful lifetime and economic value of battery systems.
Innovation Solution
A system and method that includes a computing system with a state estimator, simulation engine, and model generator, which distributes computation between remote and local systems to enhance battery state estimation, uses modular models to select and validate battery models based on application-specific criteria, and enables real-time monitoring and prediction of battery performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If battery state estimation is performed in real-time on compute-limited resources, then responsiveness and operational utility are improved, but measurement precision and accuracy deteriorate
Solution Approach 1:
The system divides battery state estimation into two segments: a computationally intensive model validation and selection process performed offline, and a simplified real-time estimation process executed on compute-limited resources during operation. This segmentation allows complex computations to be done when computational resources are available, while real-time operations use pre-computed models that run efficiently on constrained devices.
Solution Approach 2:
The system performs preliminary model validation, selection, and optimization before real-time battery operation. During these preliminary phases, multiple candidate models are evaluated and validated against historical data to identify the optimal model for real-time deployment. This preliminary action ensures that the model selected for real-time execution is already optimized for both accuracy and computational efficiency.
2Measurement precision
If complex validation and model selection processes are implemented, then battery state estimation accuracy is improved, but device complexity and computational requirements increase
Solution Approach 1:
The validation and model selection process is segmented into distinct phases: data collection, model training, validation against multiple criteria, and deployment. Each phase produces intermediate results that can be independently evaluated and stored. This segmentation reduces the apparent complexity by breaking down the monolithic validation process into manageable, modular stages that can be executed sequentially.
Solution Approach 2:
The system creates simplified copies of complex validation results in the form of pre-computed model parameters, lookup tables, and optimized algorithms. These copies encapsulate the outcomes of extensive validation processes in a form that is computationally efficient for real-time execution, thereby reducing the complexity burden during operational phases.
3Reliability
If extensive model validation is performed to ensure threshold accuracy, then reliability is improved, but loss of time and computational resources increases
Solution Approach 1:
Model validation is performed as a preliminary action before real-time battery operation begins. During this offline phase, candidate models are validated against historical battery data using multiple validation criteria including accuracy metrics, computational efficiency, and robustness tests. The validation results are cached and reused, eliminating the need for repeated validation during real-time operations.
Solution Approach 2:
The validation process evaluates multiple model parameters and thresholds to identify the optimal configuration for real-time execution. By adjusting validation parameters such as accuracy thresholds, confidence levels, and computational constraints, the system balances the depth of validation against the time investment required, ensuring sufficient reliability without excessive computational overhead.
Data Source
AI summary
A system or method for determining a battery state can include receiving a set of sensor measurements; determining the battery state using a state estimator collocated with the battery; and determining parameters used by the state estimator using a second state estimator operating on a processor remote from the battery.