Battery Pack Fault Detection Using ML Voltage-Temperature Windows
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for detecting battery faults in rechargeable batteries, such as those used in electric vehicles, lack real-time accuracy and fail to account for variations in cell chemistries and applications, leading to operational disruptions and potential hazards during charging.
Innovation Solution
A computer-implemented method using machine learning models trained on voltage and temperature sensor data from battery packs, transforming measurements into a time-series sequential window format, determining scores, and predicting fault conditions to enable timely preventative actions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional fault detection methods are used, then the system structure remains simple, but the measurement precision and reliability of fault detection deteriorate
Solution Approach 1:
The patent introduces machine learning models as intermediary components between raw sensor measurements and fault detection decisions. These models process voltage and temperature measurements through complex feature extraction and pattern recognition, achieving high detection accuracy without requiring direct complex detection hardware at every level.
Solution Approach 2:
The patent replaces traditional mechanical or threshold-based detection systems with software-based machine learning models. Instead of using complex physical detection mechanisms, the system uses computational algorithms to analyze sensor data, substituting mechanical complexity with information processing capability.
2Reliability
If traditional fault detection methods are used, then the device complexity remains low, but the reliability of battery operation deteriorates due to undetected faults
Solution Approach 1:
The patent implements preliminary fault detection by continuously analyzing voltage and temperature measurements before actual faults occur. The machine learning models predict potential failures by detecting early signs of degradation in battery cells, modules, or packs, allowing preventive maintenance before reliability-critical failures happen.
Solution Approach 2:
The system establishes continuous feedback loops where sensor measurements are constantly fed into machine learning models, which then provide fault predictions that can trigger alerts or maintenance actions. This feedback mechanism ensures high reliability by maintaining ongoing monitoring and rapid response to developing faults.
3Adaptability or versatility
If generic fault detection methods are used, then the ease of implementation is high, but the adaptability to different cell chemistries and applications deteriorates
Solution Approach 1:
The patent achieves adaptability to different cell chemistries and applications by changing the parameters and training data of machine learning models rather than changing the physical detection system. The models can be retrained with data specific to different battery types, chemistries, and usage scenarios, allowing the same hardware platform to serve multiple applications with high adaptability.
Data Source
AI summary
In one aspect, computer-implemented method may include receiving, from one or more sensors associated with a battery pack, one or more measurements pertaining to voltage, temperature, or both. The method may include transforming the one or more measurements into a time-series sequential window format, determining, based on the time-series sequential window format of the one or more measurements, a voltage score and a temperature score, and predicting, based on the voltage score and the temperature score, whether the battery pack is experiencing a fault condition. The prediction is performed by one or more trained machine learning models. Responsive to predicting the battery pack is experiencing the fault condition, the method may include performing one or more preventative actions.