Battery Fault Detection via State Amplification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current battery fault detection methods often inaccurately identify noise as a fault or fail to detect faults, leading to premature aging and reduced performance due to the reliance on voltage comparison schemes that are sensitive to criteria settings.
Innovation Solution
A processor-implemented method that generates representative states from first states of detection targets in a battery, amplifies these states to create third states, and compares them to threshold states to accurately detect faults, distinguishing between noise and actual short circuits.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If voltage comparison schemes are used to detect battery faults, then fault detection capability is provided, but noise may be incorrectly detected as a fault or actual faults may not be detected
Solution Approach 1:
The patent transforms the detection approach by changing from direct voltage comparison to a multi-step parameter transformation process. First states (raw voltage values) are transformed into second states (normalized values with removed mean), then selectively amplified to generate third states. This parameter transformation chain enhances the distinguishability between noise and actual faults while maintaining reliable fault detection capability.
Solution Approach 2:
The patent introduces an additional processing dimension by applying amplification selectively to specific second states before generating third states. This dimensional transformation allows the system to emphasize certain voltage deviations while suppressing others, thereby improving the signal-to-noise ratio and detection precision without sacrificing reliability.
2Reliability
If voltage comparison criteria are set to detect faults, then fault detection is enabled, but noise may be detected as a fault leading to premature battery aging
Solution Approach 1:
The patent applies parameter transformation to convert raw voltage comparisons into a normalized domain where noise can be distinguished from actual faults. By transforming first states into second states (removing mean values) and selectively amplifying to create third states, the system achieves more accurate fault identification, reducing false positives that would otherwise trigger unnecessary protective actions and accelerate battery aging.
Solution Approach 2:
The patent converts the harmful effect of voltage noise into a beneficial detection mechanism. By amplifying specific second states (those representing actual faults) while suppressing others (noise), the transformation process turns what was previously a harmful interference into a useful signal enhancement strategy, improving fault detection accuracy and reducing false alarms that cause battery aging.
3Reliability
If voltage comparison is used for battery fault detection, then detection capability is provided, but detection accuracy is reduced due to sensitivity to criteria settings
Solution Approach 1:
The patent implements a systematic parameter transformation approach that converts raw voltage data through multiple stages. First states (original voltages) are transformed to second states (normalized values), then selectively amplified to produce third states. This multi-stage parameter change process enhances detection accuracy by emphasizing fault-related variations while reducing sensitivity to arbitrary criteria settings.
Solution Approach 2:
The patent introduces selective amplification as an additional processing dimension that operates on the normalized second states. This dimensional transformation allows the system to differentially amplify various voltage deviations based on their characteristics, thereby improving detection accuracy independent of the original voltage comparison criteria and reducing the system's sensitivity to threshold settings.
Data Source
AI summary
A processor-implemented method to detect a battery fault includes obtaining first states of detection targets in a battery, generating a representative state based on the first states, generating second states by applying the representative state to each of the first states, generating one or more third states by amplifying at least a portion of the second states, and detecting for a fault of the battery based on the one or more third states.