Battery Cell Voltage Recovery Matrix for Slow Fault Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current battery management systems require significant computation and time to detect abnormalities in battery cells, often failing to detect slow changes in cell information, leading to delayed or missed detection of defective cells.
Innovation Solution
A battery management system that uses cell voltage as a single parameter for abnormality detection, generating an observation matrix from voltage history data and recovering it using principal components to accurately identify defective cells by analyzing differences between pre and post-recovery datasets.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple parameters (voltage, current, temperature) are monitored for abnormality detection, then detection accuracy is improved, but computation time and power consumption increase
Solution Approach 1:
The patent extracts only the voltage parameter from the set of multiple parameters (voltage, current, temperature) for abnormality detection. By focusing solely on voltage changes and their temporal patterns, the system achieves acceptable detection accuracy while significantly reducing computation time and power consumption requirements.
Solution Approach 2:
The patent transforms the voltage parameter into a time-series sequence and applies temporal pattern analysis. By changing how the voltage data is processed (from static values to temporal sequences), the system maintains detection sensitivity while reducing the computational burden compared to analyzing multiple parameters simultaneously.
2Reliability
If multiple parameters are monitored for abnormality detection, then detection comprehensiveness is improved, but power consumption increases
Solution Approach 1:
The patent extracts only the voltage parameter from the set of multiple parameters (voltage, current, temperature) for abnormality detection. By focusing solely on voltage changes and their temporal patterns, the system achieves acceptable detection accuracy while significantly reducing computation time and power consumption requirements.
3Speed
If rapid changes in cell information are monitored, then early abnormality detection is improved, but slow changes are missed
Solution Approach 1:
The patent performs preliminary action by continuously recording voltage time-series data and establishing baseline patterns before abnormalities occur. This allows the system to detect both rapid and slow changes by comparing current patterns against established baselines, improving detection accuracy for gradual degradation while maintaining rapid response capability.
Solution Approach 2:
The patent applies dynamic analysis by examining temporal patterns and trends in voltage data rather than static snapshots. This dynamic approach enables the system to adapt to both rapid changes (sudden deviations from pattern) and slow changes (gradual trend shifts), improving detection comprehensiveness across different degradation rates.
4Reliability
If cell information is monitored continuously, then abnormality detection capability is improved, but electrical energy consumption increases
Solution Approach 1:
The patent extracts only the voltage parameter from the set of multiple parameters (voltage, current, temperature) for abnormality detection. By focusing solely on voltage changes and their temporal patterns, the system achieves acceptable detection accuracy while significantly reducing computation time and power consumption requirements.
Data Source
Figure 1
Figure 2
Figure 3~4
AI summary
A battery management system according to the present disclosure includes a voltage measurement circuit configured to generate a voltage signal indicating a cell voltage of each of a plurality of battery cells connected in series, and a control unit. The control unit determines an observation matrix including a plurality of observation voltage vectors indicating a voltage history of each of the plurality of battery cells in a moving window having a predetermined size based on the voltage signal. The control unit determines a recovery matrix including a plurality of recovery voltage vectors corresponding to the plurality of observation voltage vectors in a one-to-one relationship. The control unit detects an abnormality of each of the plurality of battery cells based on a plurality of absolute error vectors indicating a difference between the plurality of observation voltage vectors and the plurality of recovery voltage vectors.