Battery Defect Diagnosis by Charging-State Data Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing battery diagnostic models have low sorting power and high false positive rates, particularly for atypical charging and discharging patterns, leading to unreliable performance.
Innovation Solution
A battery diagnosis apparatus and method that utilizes a sensor to measure test data and a processor to classify the data into multiple charging/discharging states, employing a preliminary diagnostic model to determine defects based on ensemble algorithms, ensuring high performance by requiring multiple state data sets to confirm defects.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If general diagnostic models are used to diagnose battery defects, then the diagnosis process can be performed, but the sorting power is less than 80% and the false positive rate is more than 10%, leading to unreliable performance
Solution Approach 1:
The patent divides battery test data into multiple charging/discharging state data sets based on reference values related to different charging/discharging states. Each data set is then diagnosed separately using a preliminary diagnostic model, and final defect determination is made by aggregating results from multiple state data sets. This segmentation approach improves diagnosis reliability by capturing defects across different operational states rather than relying on a single general diagnostic model.
Solution Approach 2:
The patent performs preliminary defect diagnosis for each charging/discharging state data set before making the final defect determination. This preliminary action allows the system to evaluate multiple diagnostic results and aggregate them to reach a more reliable final conclusion, thereby improving both sorting power and reducing false positive rates compared to using a single general diagnostic model.
2Productivity
If general diagnostic models are used, then the diagnosis can be performed quickly, but the false positive rate is more than 10% particularly for atypical charging/discharging patterns
Solution Approach 1:
By segmenting test data into multiple charging/discharging state data sets and diagnosing each separately, the system can identify defects that manifest under specific conditions while maintaining efficient processing. The segmented approach ensures that atypical patterns are not missed, reducing false positive rates while keeping the overall diagnosis process efficient through parallel processing of multiple data sets.
Solution Approach 2:
The patent applies preliminary defect diagnosis to each charging/discharging state data set, which may seem like excessive action, but this partial diagnosis across multiple states ensures that defects under atypical patterns are detected. The final defect determination aggregates these partial results, achieving high reliability without requiring complete re-diagnosis, thus maintaining productivity while reducing false positive rates.
3Device complexity
If a single diagnostic model is used, then the system complexity is low, but the model performance cannot be guaranteed for atypical charging/discharging patterns
Solution Approach 1:
The patent segments the diagnostic process into multiple preliminary diagnoses for different charging/discharging state data sets, followed by a final defect determination step. This segmentation allows the use of a single preliminary diagnostic model structure while achieving robust performance across atypical patterns through multi-state evaluation, balancing system complexity with reliability.
Solution Approach 2:
The patent performs preliminary defect diagnosis for each charging/discharging state data set using a preliminary diagnostic model before making the final defect determination. This preliminary action across multiple states ensures that the system can handle atypical charging/discharging patterns effectively, guaranteeing model performance without requiring multiple complex diagnostic models, thus maintaining reasonable system complexity.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A battery diagnosis apparatus includes a sensor configured to measure battery test data from a diagnosis target battery and a processor configured to classify the battery test data into a plurality of charging/discharging state data sets based on a plurality of reference values related to a plurality of charging/discharging states, perform preliminary defect diagnosis for each of the plurality of charging/discharging state data sets based on a preliminary diagnostic model, and determine the diagnosis target battery as a final defect based on a result of the preliminary defect diagnosis.