Battery Failure Detection Using Partial Correlation Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing battery monitoring systems face challenges in detecting the initial stage of failure in secondary batteries, require large storage for cumulative data, and involve time-consuming capacity measurements.
Innovation Solution
A battery monitoring system that acquires multiple types of monitoring data for each secondary battery, performs sparsity regularization, calculates partial correlation coefficient matrices, and determines abnormality levels to detect battery failure.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If cumulative monitoring data is stored for fault diagnosis, then measurement precision is improved, but device complexity increases due to large storage requirements
Solution Approach 1:
The patent extracts only the essential features needed for fault diagnosis by calculating correlation coefficients between monitoring data items. Instead of storing and analyzing all cumulative data, the system extracts correlation relationships that capture the essential patterns for detecting battery failures, thereby reducing storage requirements while maintaining diagnostic accuracy
Solution Approach 2:
The patent transforms the monitoring data from raw cumulative values into correlation coefficient parameters. By changing the parameter representation from absolute data values to relational correlation coefficients, the system reduces the dimensionality and storage requirements while preserving the diagnostic information needed for detecting battery failures
2Measurement precision
If capacity measurement is performed for fault diagnosis, then measurement precision is improved, but loss of time increases due to time-consuming measurements
Solution Approach 1:
The patent replaces the time-consuming mechanical capacity measurement process with a computational approach using correlation analysis of existing monitoring data. Instead of performing actual capacity discharge/charge tests, the system uses mathematical correlation coefficients to infer battery health status, eliminating the time loss associated with physical measurements
3Measurement precision
If multiple types of monitoring data are acquired, then measurement precision is improved, but device complexity increases due to data processing requirements
Solution Approach 1:
The patent merges multiple types of monitoring data by calculating their inter-correlations. Instead of processing each data type separately, the system combines voltage, current, temperature and other monitoring data into a unified correlation coefficient matrix, where the relationships between different data types reveal failure patterns that would be difficult to detect individually
Data Source
AI summary
A battery monitoring system monitors states of at least two secondary batteries. In the system, a data acquiring unit acquires a plurality of types of monitoring data to monitor the state of each of the secondary batteries. A failure determining unit determines whether the secondary battery has failed. The failure determining unit performs sparsity regularization using the acquired monitoring data of each of the secondary batteries as variables and calculates a partial correlation coefficient matrix of the monitoring data. The failure determining unit calculates, as an abnormality level, a difference in a partial correlation coefficient, which is a component of the partial correlation coefficient matrix, between two partial correlation coefficient matrices respectively calculated using the monitoring data of the two secondary batteries. The failure determining unit determines that either of the two secondary batteries has failed when the calculated abnormality level exceeds a predetermined threshold.