Battery Life Prediction Using Adaptive SOH Aging Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for predicting the remaining life of lithium-ion batteries in uninterruptible power supply systems are limited in accuracy due to variations in aging trends caused by manufacturing tolerances, usage patterns, and environmental factors, leading to potential errors in reliability and economic losses.
Innovation Solution
A battery life prediction device and method that generates state-of-health data in real-time, uses a state data generator to create a health state model with a non-linear function, and adjusts the model using least squares approximation to accurately calculate the remaining battery life, incorporating real-time sensing data and past state data for improved accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If aging test data from manufacturers is used for prediction, then the method is simple to implement, but the prediction accuracy is low due to manufacturing tolerances and varying usage patterns
Solution Approach 1:
The system continuously monitors actual battery state-of-health data and uses it to correct and update the aging trend model. This feedback mechanism allows the prediction system to adapt to individual battery characteristics and actual usage conditions, significantly improving prediction accuracy while maintaining implementation simplicity through automated data collection and model updating.
Solution Approach 2:
The system performs preliminary corrections to the aging trend model using initial state-of-health data before making predictions. By pre-adjusting the model parameters based on actual battery performance data collected during early operation, the system accounts for manufacturing tolerances and establishes a more accurate baseline for future predictions.
2Device complexity
If traditional prediction methods are used, then the system complexity is low, but the reliability of power supply prediction is insufficient
Solution Approach 1:
The system dynamically updates the aging trend model parameters based on real-time state-of-health data. Instead of using static manufacturer-provided aging curves, the model adapts to the actual battery's degradation pattern, making the prediction system more reliable while managing complexity through efficient data processing algorithms.
Solution Approach 2:
The system changes the parameters of the aging trend model based on actual measured data. By adjusting model parameters such as degradation rates and threshold values according to real battery performance, the system achieves higher reliability without requiring a complete redesign of the prediction architecture.
3Measurement precision
If real-time state monitoring and model updating is implemented, then prediction accuracy is improved, but the computational requirements and processing time increase
Solution Approach 1:
The system performs partial model updates at predetermined intervals or when specific triggers are met, rather than continuously recalculating all parameters. This approach maintains prediction accuracy by updating the model sufficiently often while avoiding unnecessary computational overhead that would increase processing time.
Data Source
AI summary
A battery life prediction device, including a state data generator configured to receive information about a battery in real time and to generate state-of-health data; a state data storage configured to store the state-of-health data and past state data; and a battery life calculator configured to: generate a health state model based on the past state data, generate state prediction data based on the state-of-health data using the health state model, determine whether to modify the health state model based on the state prediction data and the state-of-health data, and calculate a remaining life of the battery, wherein the generating of the health state model includes generating an initial model including a non-linear function, determining an initial model coefficient using a least squares approximation based on the past state data, and generating the health state model based on the initial model coefficient and the initial model