A team of engineers led by a University of Arizona doctoral student has introduced a novel method for preventing EV batteries from overheating. The method uses AI algorithms to predict areas of concern before they become dangerous. Many see this study as a breakthrough in the industry, especially considering the growing demand for EVs. Here’s everything you need to know about AI’s future role in preventing thermal runaway.
Lithium-ion Batteries (LIBs)
There are some key points to understand to grasp the magnitude of this research. For one, lithium-ion batteries (LIBs) are the most commonly used battery in today’s EVs. These batteries integrate charged lithium ions to transfer energy across the unit, producing current for your electrical needs. What makes LIBs so popular is that they can be charged by flipping polarity temporarily and sending the ions back to the negative pole of the unit.
-
Source – Car Buyer
Today’s EVs rely on these devices for many reasons including that they have decent life spans, are relatively lightweight compared to alternatives, and provide exceptional energy density. Notably, it’s common for these batteries to use cells that are grouped to create the full EV pack. Notably, most EV battery packs have thousands of cells.
What is Thermal Runaway?
The current multi-cell structuring of LIBs helps the batteries charge faster and achieve longer lifespans. However, it can create hotspots within the battery pack that can result in catastrophic failure. When a single cell starts to malfunction, it can heat up quickly, causing the surrounding cells to experience increased temperature and potentially leading to more failure. This domino effect is called thermal runaway, and it’s one of the main problems faced by EVs today.
Thermal Runaway (TR) can reduce performance, cause battery decomposition, and even explosions. As such, it’s a real concern today for EV owners. Thermal Runaway can be caused by several factors including battery failures such as melting of the separator, decomposition of the cathode, or an adverse Li-electrolyte reaction.
Safety Concerns
These short circuits can occur quickly and result in nearby bystanders getting injured due to fire and explosions. There are plenty of stories of people waking up to house fires or other terrifying moments due to their EV battery igniting. As such, solving this problem has become a primary concern for researchers globally.
Rising Temperature
The need to reduce TR has become more important over the last few years due to multiple factors. The rise in both EV usage and global temperatures has made a dangerous scenario with more lives at risk than ever. These factors make keeping batteries cool essential to achieving a greener future.
AI Thermal Runaway Study
A study published in the Journal of Power Sources demonstrates how an advanced AI algorithm coupled with sensors could be the key to eliminating thermal runaway once and for all. The study, led by Basab Goswami, uses driver data simulations to mimic EV battery usage under daily driving conditions.
Multiphysics and machine learning models that leveraged thermal, electrochemical, and degradation sub-models were used to determine key moments when TR became noticeable. From there, the AI systems reinforced the data, allowing them to predict and identify overheating cells faster than any optical solution.
AI Thermal Runaway Test
Researchers sought to gain a better understanding of how a solid electrolyte interface degrades on a negative electrode under various conditions. The team used real driver data and battery states such as the constant charge/discharge and driving cycles to test the battery’s heat signature. To accomplish this task, the team created a battery that had special thermal sensors wrapped around it.
The temperature sensors provided detailed spatial and temporal temperature data that was then combined with historical data and fed to the AI algorithm. This data included key situations, environments, driver activities, and technical issues.
Goswami’s Algorithm
The Goswami Algorithm is unique in many ways. For one, it’s the first AI machine-learning model used to predict TR. This multiphysics model was only made possible thanks to new AI systems such as vector modeling. These advanced systems can analyze massive amounts of data and point out correlations or complex patterns far beyond human capabilities. Consequently, the modeling method enabled the team to create realistic data on EV-driving behavior.
AI Thermal Runaway Test Results
The study’s results are impressive. For one, the team was successful in its goal to accurately and precisely predict TR in LIBs consistently. The AI was very precise and could even determine where the thermal runaway began, alerting of the danger and preventing further damage. Now, the team seeks to expand its research, which could one day help create safer EVs for all.
AI Thermal Runaway Benefits
There are many benefits that this research brings to the market. For one, the AI algorithm is far less expensive than using other methods to prevent Thermal Runaway. In the past engineers, including the ones in this study, have put forth optical measurement methods. These tactics were effective but required expensive cameras and constant monitoring. AI frees up the team’s efforts and provides better results.
High Accuracy
The ability to predict thermocouple temperature and the location of hotspots precisely is a vital step in creating more durable and capable batteries in the future. Engineers can use this approach to help determine where their products can be improved and what parts need to be upgraded or altered to prevent future TR.
Achieving NetZero Emissions
Another benefit of this research is that it makes the global goal of reducing carbon and pollutants easier. EVs are one way to help reduce car emissions and any method of improving their efficiency and effectiveness will have a major impact on the auto market moving forwards. In 2023, EV vehicle sales increased 35% globally. Analysts predict this growth to continue, making TR safety a priority for most manufacturers.
Easily Integrated
This tech can be implemented into supply lines and manufacturing processes with minimal effort. The AI model will improve as more data, better sensors, and additional models are introduced into every electric vehicle’s battery management system.
Researchers
The study was conducted at the University of Arizona by Vitaliy Yurkiv, Farzad Mashayek, Yasaman Abdisobbouhi, Hui Du, and Todd A. Kingston. The project was made possible through a $599,808 grant from the Department of Defense’s Defense Established Program to Stimulate Competitive Research. Notably, Goswami and Yurkiv published another paper in January of this year demonstrating an optical method of detecting hot spots. This latest study is the culmination of that research and a desire to provide a more affordable and accessible option to them.
Two Companies that may Benefit from this Research
Several companies could see immediate benefits from this study. The demand for EVs is on the rise, and manufacturers are entering the market to fill this need. The ability to provide safer alternatives versus today’s options will help catapult these manufacturers to the top of the market. Here are two companies that could implement this tech today and see increased bottom lines.
1. Rivian 
Rivian was founded in 2009 as Mainstream Motors. After several years and name changes, the company officially became Rivian in 2011. Today, Rivian remains one of the largest EV manufacturers in North America. The company specializes in EV sports utility vehicles and is headquartered in California with manufacturing operations in Illinois.
Rivian listed $4.43B in revenue in 2023 with $16B in total assets. In recent years, the firm has put considerable effort towards achieving net zero carbon emission by 2040. These factors, plus the growing demand for its products, make Rivian a strong “hold” for traders seeking Tesla alternatives.
2. Xiaomi
Chinese-based EV manufacturer Xiaomi is seen as the biggest competitor to Tesla Motors in the EV race. The company continues to gobble up market share globally due to many factors. For one, the XIAMI brand is well-known and has been a powerhouse in the electronics sector for decades. As such, there is strong consumer loyalty and trust compared to newer options.
Xiaomi has a market cap of $462.55B and shows a 6.12% profit margin currently. The company continues to expand its EV market share and technology, leading to the stock remaining in high demand globally. Those seeking a fast-growing EV manufacturing stock should consider Xiaomi.
Thermal Runaway and Never Come Back
You have to commend this team for creating a reliable and accurate method of determining the precise steps involved in a thermal runaway scenario. Now, this data can be fine-tuned and used to improve detection even further. In the coming months, your EV may rely on a similar tech based on the hard work these engineers put forth. As such, it’s wise to keep tabs on this discovery moving forward.
Learn about other cool artificial intelligence projects now.
