Real-Time Battery RUL Estimation Using AI Data Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing battery management systems face challenges in implementing real-time on-board applications for estimating the remaining useful life (RUL) of batteries due to the large amount of data required, making it difficult to provide accurate and timely estimates for critical applications.
Innovation Solution
A method and system that classify battery data into pre-defined classes based on charge, discharge, and impedance cycles, using artificial intelligence (AI) models to estimate both gross and fine RUL, with the option to trigger alerts when the estimated RUL approaches the end of life threshold, enabling precise monitoring and timely replacements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If existing battery management systems use traditional RUL estimation methods, then they can provide RUL estimates, but they require a large amount of data which makes real-time on-board implementation difficult
Solution Approach 1:
The patent segments battery data into multiple classes based on charge, discharge, and impedance cycles. By dividing the data into distinct classes, the system can apply specific estimation methods to each class, reducing the overall data processing complexity while maintaining estimation accuracy.
Solution Approach 2:
The patent implements a two-level estimation approach where a first level provides a gross RUL estimate and a second level provides a fine RUL estimate. This partial action approach allows the system to obtain satisfactory RUL estimates without processing all available data, thereby reducing computational complexity while maintaining sufficient accuracy for practical applications.
2Measurement precision
If battery management systems process large amounts of data for RUL estimation, then they can achieve accurate estimates, but real-time estimation becomes difficult to implement
Solution Approach 1:
By segmenting battery data into multiple classes based on operational characteristics (charge, discharge, impedance cycles), the system can process each class separately with optimized algorithms. This segmentation enables real-time processing by reducing the computational burden on any single data processing operation.
Solution Approach 2:
The two-level estimation method allows the system to achieve adequate RUL accuracy without processing the complete dataset in real-time. The first level provides a quick gross estimate, and the second level refines it when needed, enabling real-time implementation while maintaining sufficient accuracy.
3Measurement precision
If the system performs detailed second level RUL estimation for all battery classes, then fine RUL accuracy is achieved, but computational resources are wasted on classes where it is not needed
Solution Approach 1:
The patent applies different estimation methods to different battery classes based on their specific characteristics. The detailed second level estimation is applied only to specific classes where it provides necessary accuracy, while other classes use the simpler first level estimation. This local quality approach optimizes computational resource usage by matching the estimation method to the specific needs of each battery class.
Solution Approach 2:
The system dynamically adjusts the estimation approach based on battery class parameters. By changing the estimation method (from first level to second level) based on the identified battery class, the system achieves fine RUL accuracy only when necessary, thereby reducing overall computational resource consumption while maintaining required accuracy for critical cases.
Data Source
AI summary
A method of estimating a remaining useful life (RUL) of a battery includes: identifying a class of data of the battery in real time; determining whether a second level RUL estimation is set for the class; estimating a gross RUL by performing a first level RUL estimation in response to the second level RUL estimation not being set for the class; and estimating a fine RUL of the battery in response to the second level RUL estimation being set for the class.