Battery SOH Regression for Rapid Deterioration Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for detecting battery deterioration in electric vehicles are prone to errors due to measurement noise and sensor inaccuracies, leading to unreliable detection of rapid capacity deterioration.
Innovation Solution
A battery management system that includes a measurement unit, SOH estimation, deterioration regression curve generation, and rapid deterioration determination units to accurately assess battery health by analyzing voltage, current, and SOH data through curve regression and coefficient comparison.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If linear regression is performed on change amount of FCC or SOC to detect rapid deterioration, then deterioration detection method is established, but measurement errors and noise cause erroneous determination
Solution Approach 1:
The patent divides the time-series SOH data into multiple sections and performs curve regression on each section separately. By comparing deterioration coefficients between sections, the system can detect rapid deterioration while reducing the impact of measurement errors in individual data points. This segmentation approach isolates the effect of noise to specific sections rather than affecting the entire detection process.
Solution Approach 2:
The patent performs curve regression analysis on historical SOH data before making deterioration determination. By establishing a baseline deterioration trend through preliminary regression on past data, the system can compare current deterioration rates against this baseline, thereby reducing erroneous determinations caused by temporary measurement errors or noise.
2Reliability
If curve regression is performed on time-series SOH data, then deterioration trend can be analyzed, but calculation complexity increases
Solution Approach 1:
The patent segments time-series SOH data into multiple sections and performs curve regression on each segment separately. This segmentation reduces calculation complexity by breaking down a large dataset into smaller, more manageable sections, while still maintaining reliable deterioration detection through comparison of deterioration coefficients across sections.
Solution Approach 2:
The patent changes the parameter being analyzed from raw SOH values to deterioration coefficients obtained through curve regression. By focusing on the deterioration coefficient as the key parameter and comparing its changes between different time sections, the system simplifies the detection process while improving reliability.
Data Source
AI summary
SOH estimation unit (4613) estimates a state of health (SOH) of battery (E1, 41) based on measurement data of battery (E1, 41). Deterioration regression curve generation unit (4614) generates a deterioration regression curve of battery (E1, 41) by performing curve regression on a plurality of the SOHs specified in time series for battery (E1, 41). Rapid deterioration determination unit (4615) determines whether or not rapid deterioration has occurred in battery (E1, 41) based on a difference or a ratio between a deterioration coefficient of a deterioration regression curve of battery (E1, 41) generated based on the plurality of SOHs in a first data section and a deterioration coefficient of a deterioration regression curve of battery (E1, 41) generated based on the plurality of SOHs in a second data section.