Abnormality Detection Using Linear Relationship Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for detecting abnormalities in electric vehicle batteries, such as focusing on single-time series data like output voltage, suffer from reduced accuracy as the linearity between multiple measured items collapses in abnormal states, making it difficult to detect issues like deteriorated connections or imbalanced battery voltages.
Innovation Solution
A method that acquires operation log data from electric vehicles and determines abnormalities by identifying when measured data from multiple items no longer satisfy their linear relationship, using correlation analysis to set thresholds and detect deviations, thereby improving detection accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If single-time series data like output voltage is used for abnormality detection, then the detection method is simple, but the detection accuracy deteriorates when linearity between multiple measured items collapses in abnormal states
Solution Approach 1:
The patent transitions from monitoring single-time series data to multi-dimensional monitoring by analyzing linear relationships between multiple measured items simultaneously. This dimensional expansion enables detection of abnormalities that single-parameter methods miss, such as deteriorated connections or imbalanced battery voltages, thereby improving detection accuracy without excessive complexity increase
Solution Approach 2:
The patent changes the monitoring parameter from individual time series values to linear relationship characteristics between multiple parameters. By detecting changes in the linear relationship structure rather than individual parameter values, the system can identify abnormalities even when single-parameter thresholds are not violated, resolving the accuracy issue
2Measurement precision
If multiple measured items are monitored to improve detection accuracy, then abnormality detection accuracy improves, but the complexity of data processing increases
Solution Approach 1:
The patent extracts and analyzes only the linear relationship characteristics from the multi-dimensional data, rather than processing all raw data points. By focusing on the relationship structure between items rather than individual values, the system reduces processing complexity while maintaining high detection accuracy for abnormalities like deteriorated connections or imbalanced voltages
Data Source
AI summary
A method of detecting abnormality includes acquiring, by a computer, operation log data that include measured data of a plurality of items related to an operation of a mobile object that is electrically driven; and determining, based on the operation log data, that an abnormality occurs in the mobile object when measured data of two items among the plurality of items do not satisfy a linear relationship, the two items having the linear relationship in a normal state of the mobile object.