Aging State Model Preparation Using Active Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for determining the aging state of electrical energy storage means, such as batteries, are costly and complex, requiring significant energy expenditure and time, especially for generating training data for aging state models, and lack direct measurement methods, relying on indirect model-based approaches that are computationally complex and not suitable for everyday use.
Innovation Solution
A method for preparing a data-based aging state model involves operating energy storage means under different load profiles, recording operational variable trends, determining aging states as labels, and selecting subsets for training data based on information measures from predictive covariance, using a hybrid model combining physical and data-based approaches, including probabilistic regression models like Gaussian processes, to improve accuracy and efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If initial measurement is provided for a defined number of energy storage means to generate training data, then the aging state model can be trained, but the energy expenditure and time required become significant and scale with the initial number of energy storage means being measured
Solution Approach 1:
The patent applies partial action by measuring only a subset of energy storage means at each measurement time point rather than all units. The system selectively identifies and measures only those units that provide maximum information gain for model training, reducing the number of measurements required while maintaining model accuracy. This is achieved through the information theoretic approach that calculates expected information gain for each unmeasured unit and selects the most informative subset.
Solution Approach 2:
The system uses the predictive covariance of the aging state model itself to identify which measurements would be most valuable. The model's own uncertainty estimates guide the selection of units to measure, allowing the system to self-optimize its measurement strategy without external intervention. The information measure based on predictive covariance enables the system to autonomously determine which measurements will most effectively reduce model uncertainty.
2Measurement precision
If initial measurement is provided for a defined number of energy storage means to generate training data, then the aging state model can be trained, but the time required becomes considerable, in particular if sufficient training data remains to be recorded for aged energy storage means
Solution Approach 1:
The patent reduces measurement time by performing partial measurements at selective time points rather than continuously monitoring all units. The system determines optimal measurement time points based on when expected information gain is maximized, and only measures a subset of units at each time point. This approach significantly reduces the total measurement time while accumulating sufficient training data for model training.
Solution Approach 2:
The system implements feedback by using the aging state model's predictive covariance to continuously guide measurement decisions. After each measurement, the model is updated and the information measure is recalculated, creating a closed-loop system that adapts measurement strategy based on accumulated data. This feedback mechanism ensures that measurements are performed at optimal times and on optimal units, minimizing total measurement time while maximizing model accuracy.
3Reliability
If operational variable data is continuously recorded and transferred to a central processing unit, then aging states of electrical storage means in a plurality of devices can be monitored, but the device complexity and data processing requirements increase
Solution Approach 1:
The patent extracts only the essential information needed for model training and monitoring from the continuous operational data stream. Instead of processing all raw operational variable data centrally, the system extracts key features and aggregates data at distributed units before transmission to the central processing unit. This extraction approach reduces the volume of data requiring central processing while maintaining the reliability of aging state monitoring.
Solution Approach 2:
The system segments the monitoring function by performing preliminary processing and information measure calculations at distributed measurement units rather than centralizing all processing functions. Each measurement unit independently calculates information measures and selects which data to transmit, dividing the overall processing burden across multiple distributed units. This segmentation reduces the complexity burden on any single central processing unit while maintaining comprehensive monitoring capability.
Data Source
AI summary
A method for initially preparing an at least partially data-based aging state model for an electrical energy storage means is disclosed. The method includes providing a number of energy storage means on a test bench for measurement based on a respective load profile, wherein the load profiles are different and characterize a chronological trend of at least one load-imposing operational variable for the energy storage means. The method also includes operating the number of energy storage means having the respective associated load profile and recording chronological operational variable trends. Further, the method includes at a predetermined evaluation timepoint, determining an aging state of a subset of the energy storage means as a label based on an input vector, and generating a training data set, which includes the operational variable trends and the determined label, for each energy storage means of the subset of the energy storage means. The method additionally includes selecting the subset of the energy storage means having the respective associated load profile based on an information measure for the subset of the energy storage means, the measure being determined using a predictive covariance of the data-based aging state model at at least one future timepoint.