Invention:
This invention introduces a novel framework for predicting thermal runaway (TR) in lithium-ion batteries (LIBs) by combining strategically placed sensor arrays with machine learning models to predict TR events in lithium-ion batteries before they occur. TR is a phenomenon that occurs when the temperature in a LIB causes reactions to occur which produce more heat, which causes more reactions and is one of the leading causes of LIB fires. This technology uses an array of thermal sensors placed throughout a LIB module to collect real-time data. Then, a convolutional neural network-based machine learning model is used to compare the data to electrochemical models and predict whether TR will occur. By enabling the early detection of potential TR events, this technology identifies heat source locations and triggers real-time alerts, enhancing battery safety and thermal management.
Background:
Lithium-ion batteries (LIBs) power a wide range of technologies including electric vehicles, phones, laptops, and wearable electronics. They are widely used because of their efficiency, high energy density, long lifespan, and fast charging, among other factors. However, they carry the risk of thermal runaway (TR), a chain reaction that can cause overheating and battery fires. Managing TR risk is crucial in the safe development of environmentally friendly technologies. Existing methods focus on passive safety mechanisms or post-event mitigation, which do not always prevent catastrophic failures. By integrating sensors, multiphysics modeling, and advanced machine learning to proactively predict and manage thermal events, this technology provides a TR prediction model based on real-time sensor data to mitigate safety concerns in LIBs.
Applications:
- Lithium-ion battery safety
- Electric vehicles
- Consumer electronics
- Energy storage systems
Advantages:
- Improved prediction accuracy
- Real-time sensor data
- Enhanced safety
