College of Engineering and Computer Science

Is the generative AI hype bubble about to burst?

lblouin — Mon, 10 Mar 2025 13:08:28 +0000

Is the generative AI hype bubble about to burst? lblouin Mon, 03/10/2025 - 09:08

The release of ChatGPT in November 2022 sparked some of the broadest societal discussions about the promise and perils of artificial intelligence in recent memory. In the year after its debut, it was easy to find stories about the potential for large language models, the AI technology underlying ChatGPT and similar products, to totally restructure certain industries. Some looked out even further and worried that AI could eventually threaten human civilization.

In retrospect, the expectations were perhaps too tall for what was, at the time, a fascinatingly good chatbot that was still prone to lying and inexplicable hallucinations. Even with these limitations, investors have been bullish about the technology, with investments in generative AI technologies topping $1 trillion. But now, two years on, with super compelling use cases yet to materialize, some are starting to wonder whether the industry could be dangerously overvalued and overhyped. After all, the most commonly deployed uses — customer service chatbots, AI enhanced search (which has inspired an internet backlash and interesting workarounds), AI summaries of product reviews and help writing emails — aren’t life-changing. Even in the field of coding, where LLMs have arguably shown the most practical promise, applications are still limited. Moreover, there are huge concerns that ChatGPT and similar technologies are actually doing damage to society, by helping students cheat, erasing recent reductions in energy use and greenhouse gas emissions, and capitalizing on creators’ work without their permission.

Electrical and Computer Engineering Professor Paul Watta says these are all valid concerns. His own take is that it’s “going to be tough” for big bets on generative AI to pay off, and it’s completely possible the industry is heading for a reality check soon. (More on this below.) But he’s also not writing off the potential for some game-changing use cases to still emerge. His main reason for tempered optimism is that LLMs have undergone quite a lot of meaningfully technical evolution over the past two years, a story that often gets lost in the media’s coverage of generative AI. He describes the initial releases of ChatGPT and similar products as “chat tools,” whereas the new generations are “reasoning models.” Watta says the former were basically extremely powerful predictive text machines: Based on a text-based prompt, the model would use its knowledge of the patterns of human language to string together words that felt like an appropriate answer. Sometimes the outputs were really great. Other times, particularly when the technology failed to capture the full meaning of the prompt, its answers were frustratingly unhelpful. This is, for example, why early generations of LLMs generally failed as customer service chatbots. They simply couldn’t accurately interpret the nuances of people’s troubleshooting questions, let alone integrate with things like databases of relevant customer information.

Watta says reasoning models emerged to overcome these shortcomings. The main difference is that a reasoning model essentially thinks before it speaks. Rather than generating a quick text-based output based on a single computational line of thought, a reasoning model breaks a prompt down into its component parts to try to better understand the context of the task being asked of it. It then considers multiple options for its response and chooses an answer based on refined numerical parameters established during the model’s training that steer it toward more desirable results. The newer models can also quickly ingest and respond to new information. “That’s really one of the best use cases for it right now. It can take in all kinds of documents — like business prospectuses from a bunch of companies — and generate a summary report for you that’s really quite good,” Watta says. “So that’s something that might have taken an intern 30 hours to do, and now you can do it in minutes.” Because of this fundamentally different architecture, reasoning models also do a reasonably good job of showing how they arrived at a conclusion. Original LLMs were black boxes — not even their designers had a clue how they were coming up with their responses.

The other reason that Watta is still not counting out LLMs is that the technology is evolving very rapidly. Watta says that Sam Altman, CEO of OpenAI (maker of ChatGPT), boasts that every version they release is a 10-fold improvement and that the newest version, expected any day now, could surpass that pace. That’s fast even for the tech sector. There is at least some concrete evidence to demonstrate that progress. For example, Watta keeps an eye on competitive programming challenges on platforms like Kaggle, where companies post real-world unsolved programming problems for the world’s best developers to tackle. The earliest releases of LLMs were essentially non-competitors as coders. But Altman says internal benchmarks for their most recent model put it in the Top 50. Watta wouldn’t be shocked if the pending releases of ChatGPT or other models end up in the Top 10, or even land at No. 1.

Watta pulls up a website showing what ChatGPT actually looks like: huge data centers that use massive amounts of electricity. Watta says electricity consumption by AI-based technologies could become a hot-button policy issue in the future. Photo by Annie Barker

Even more interesting than the pace of development is that there now appears to be legitimate competition in the LLM space. Earlier this year, the Chinese startup DeepSeek released an LLM model that caused instant disruption. The big revelation wasn’t that DeepSeek was better than ChatGPT — though Watta says, by some benchmarks, DeepSeek slightly outperforms the best models out there. It was that this startup had managed to build this almost-just-as-good reasoning model using far less powerful technology, thanks to a U.S. trade policy which banned the highest-powered graphics processing units from being sold to Chinese companies. These chips, made by the U.S. company NVIDIA, were assumed to be essential to creating high-quality LLMs. “They appear to have proved that idea wrong,” Watta says. “What the DeepSeek team did is the classic startup story that Silicon Valley used to do. Startups never have enough resources, so you have to optimize what you have. The big players become lazy, they don’t look for new ideas, and so they overspend to solve a problem. A startup can’t do that. That’s how the little guy beats the big guy.” Moreover, Watta says it’s notable that DeepSeek made key features open source and published technical details about how they arrived at their performance breakthroughs. U.S. companies typically have just published benchmark performance but don’t share the details of how they got there.

Watta says, for him, the biggest takeaway from DeepSeek’s release wasn’t technological; it was its impact on the markets. The day after DeepSeek’s debut, NVIDIA’s stock fell by 17%. “The market lost half a trillion dollars. From one release. That’s scary,” he says. He argues that indicates we may be entering a period where venture capitalists become more discerning about their investments in LLM development. Moreover, that’s just one of many possible “bottlenecks” Watta and others foresee. There’s still the nagging question of how much more these models can be improved, given that they thrive on consuming human-produced data and that supply could run dry in the next few years. Newer models still lie and hallucinate, though not as much. In addition, some are growing increasingly worried about the electricity consumption of LLMs and AI-based technologies more generally. Currently, AI accounts for about 3% of global electricity use, but that number is expected to grow in the coming years. (By one estimate, a single ChatGPT query consumes about as much electricity as a light bulb does in 20 minutes and 10 times as much as a standard Google search.) Moreover, this growth in electricity demand from AI comes at a time when the push toward electrification in the transportation and heating sectors is already expected to strain the electric grid. Watta says if AI is seen as the reason for spikes in electricity cost, brownouts or an impediment in the fight against climate change, the public could sour on the industry.

Similarly, he says LLMs could run into hurdles with existing privacy law. Watta says one of the more hopeful applications for these new reasoning models could lie in solving complex medical problems. “But when you’re talking about people’s medical information, we have strict regulatory frameworks, like HIPAA, which are designed to protect people’s privacy. With something like that, there is no room for error,” he says. “People have already gotten a little uncomfortable with these technologies making decisions that impact their safety or their financial lives. Now what if someone uses the technology to make a virus that kills millions of people? That could create a major backlash. Even if it wasn’t something that catastrophic, if people’s views go negative, that could constrain the development and shake the markets, which could have large economic consequences. Because, at least for now, the markets are still betting on a big pay day.”

###

Story by Lou Blouin

Faculty and Staff

Opinion or Voices

Technology

College of Engineering and Computer Science

Electrical and Computer Engineering

Off

Mon, 03/10/2025 - 13:07

Two years after the launch of ChatGPT, generative AI has yet to produce a game-changing use case. Professor Paul Watta breaks down whether the trillion dollar bet on generative AI will pay off or go bust.

Electrical and Computer Engineering Professor Paul Watta. Photo by Annie Barker

��-Dearborn earns R2 research designation

lblouin — Wed, 26 Feb 2025 12:34:13 +0000

��-Dearborn earns R2 research designation lblouin Wed, 02/26/2025 - 07:34

By every big metric that Vice Provost for Research and Dean of Graduate Studies Armen Zakarian tracks, ��-Dearborn’s research enterprise is on a tear. The big top line stat: Since 2018, externally funded research expenditures have nearly tripled, increasing from $4.8 million to a projected $13.2 million for FY25. Earlier this month, that growth led to a reclassification of ��-Dearborn as an R2 institution, a designation from the Carnegie Foundation for the Advancement of Teaching and the American Council on Education for universities that award at least 20 research doctorates and have $5 million average annual research spending. It’s the first time the university has received this designation. ��-Dearborn Chancellor Domenico Grasso says the honorific is, without a doubt, an important new feather in the cap for the university. But, for him, the real achievement is all the work that it’s a reflection of. “This recognizes something which we already accomplished — and that is great and I’m very proud of that — but we were going to do it anyway. And honestly, we’re not even satisfied with where we are. There remains enormous untapped potential, so this is really just the start for us,” Grasso says.

So how did we get here? Zakarian and Grasso point to a number of contributing factors. Flash back to 2018 and Zakarian says you’d see an institution where research was “recognized but not prioritized.” The tone changed when Grasso came in and made growing the university’s research and scholarly output one of his top priorities. “In my view, distinguished universities are defined by a faculty comprising scholar-teachers. To my mind, the definition of a professor is someone who is a scholar at the forefront of their field, sharing their passion, knowledge and discoveries with their students,” Grasso says. “Even in the elite small liberal arts schools in New England, where they are known for their focus on educating undergraduate students, the faculty are active scholars.” Zakarian says the administration endorsed that model and supported faculty with larger investments in research development and pre- and post-award support, so researchers could get assistance securing and administering grants.

Aside from the messaging and support from leadership, Zakarian thinks three other things helped set the stage for the explosive research growth the university has seen over the past few years. First, he says faculty consistently cite a shortage of time as their biggest impediment to building their research careers. So some departments decided to lighten the teaching load from three to two courses per semester. The second big factor: the sanctioning of the College of Engineering and Computer Science doctoral programs by U-M’s Rackham Graduate School in 2019. That affiliation meant the programs had to meet the same rigorous standards as doctoral programs on the Ann Arbor campus, which boosted ��-Dearborn’s prestige and ability to recruit top PhD students. This helped faculty recruit valuable talent to power their labs, which further eased their time/labor burden. Finally, Zakarian says when filling faculty vacancies, colleges prioritized hiring ambitious, early-career, research-focused scholars who had the potential to become leaders in their fields.

Then, of course, there’s something to be said for momentum. As some of the above changes started to take root, a few faculty started landing bigger grants, and a larger share of the awards were coming from federal funders, like the National Science Foundation, Department of Energy and Department of Defense. “We used to get very few million dollar grants. But it just took a few teams to do that and then we started getting more,” Zakarian says. “You basically reach a point where I think people are looking at what each other are doing, and the mood shifts. Now, we are approaching $85 million in submitted proposals, which I think is just incredible. And the number of proposals hasn’t gone up by a whole lot. That means faculty are more confident going for bigger and bigger grants. Frankly, that’s what it’s going to take. It would be difficult for us to reach the next level, $200,000 at a time.”

Zakarian says the university has also reached a point where the research enterprise has a fairly broad base. “When I think about how sustainable this is, I’m basically looking at whether this is just a handful of people bringing in the big dollar amounts. But it’s much broader than that,” he says. Moreover, while engineering disciplines are responsible for the largest share of the research growth, Zakarian says other disciplines, particularly in the natural sciences, are starting to develop the same momentum. Last year, with engineering in a good spot, the Office of Research made it a priority to assist faculty in the natural sciences. That resulted in several big research grants in biology and chemistry from national funders, including the National Institutes of Health and Department of Energy. Looking ahead, Zakarian and Grasso say the most untapped potential lies in those disciplines, along with computer science, artificial intelligence, and health and human services — the latter of which could also be targeting NIH grants, one of the national funders with the biggest budget.

Grasso notes, however, that realizing that potential now faces some new political hurdles. The Trump administration has pursued several measures that directly disrupt university research, including attempting to severely cut the amount of indirect costs covered by NIH grants, layoffs at the National Science Foundation and pauses by grant review committees. “This is a significant threat,” Grasso says. “I understand the desire to be cost conscious and efficient. But research has been the fuel in the engine that has powered this country, whether you’re talking about revolutionary medications, transistors, AI or practically everything that has changed the world — it all has roots in research — and many of those roots found a home in American soil. To hamper this is incredibly reckless and foolhardy.”

Even within the climate, Zakarian says the R2 designation and what it represents should help keep the momentum going. “Obviously, a lot of this goes back to our ability to recruit and retain top faculty, and I think this designation really does help us communicate that if you are really ambitious and you want to build a strong research program, you can do it here. It says, ‘We’re open for business. We have the resources and the vision. Others have done it, you can do it too.’”

###

Story by Lou Blouin

College of Arts, Sciences, and Letters

College of Business

College of Education, Health, and Human Services

College of Engineering and Computer Science

Off

Wed, 02/26/2025 - 12:33

The recognition follows six years of rapid growth in which externally funded research spending at ��-Dearborn has nearly tripled. So how did we get here? And what’s next?

Associate Professor Dr. Zhi Zhang (not pictured) and Assistant Professor Jie Fan (pictured, middle) recently landed a $465,000 grant from the NIH. Large federal grants helped power ��-Dearborn to an R2 designation in 2025. Photo by Julianne Lindsay

For a ��-Dearborn student and professor, the influence goes both ways

lblouin — Mon, 24 Feb 2025 13:06:36 +0000

For a ��-Dearborn student and professor, the influence goes both ways lblouin Mon, 02/24/2025 - 08:06

In 2020, Electrical and Computer Engineering Professor Wencong Su was looking for an undergraduate student to assist with a project focused on adding innovative energy efficiency measures to the already LEED-certified Engineering Lab Building. At the time, Su was teaching all graduate courses so he didn’t have many undergrads in his orbit. So Su asked for a little help from one of his graduate assistants, who emailed the job posting to the students in a 200-level lab. LaRico Andres, a sophomore and recent transfer from Henry Ford College, thought it sounded interesting and reached out to Su with his resume. Su remembers Andres didn’t exactly fit the description of the type of student who usually applies for these kinds of opportunities. Andres was a few years older than Su, already had a full-time job working as an instrument tech at the Great Lakes Water Authority, and his GPA at the time wasn’t knocking anyone out. But after the two chatted for a little while, Su offered him the job. Andres was personable, seemed excited to learn, and though he was a research novice, it turned out his background in instrumentation was actually quite relevant to the types of things they would be investigating in the ELB.

Andres excelled on the project and it led to many more opportunities to work alongside Su. On one project supported by the National Science Foundation’s I-Corps — a program that helps researchers translate their laboratory work into marketable products — Andres traveled to conferences in Orlando, San Diego and Austin, completing a flurry of more than 100 interviews with industrial professionals in seven weeks about a product Su’s lab was developing. On another project, he and Su created an app that used the ELB’s building management software and the utility’s fuel profile to give users an idea of what being a building occupant added to their personal carbon footprint. As you’d expect, Andres picked up a lot of skills working with Su. But as often happens when two people spend a lot of time together, Andres and Su also started talking about things other than their work. Andres jokes that they initially bonded over the fact that they “both like to eat.” But the conversations gradually grew more personal. Both men had daughters, so they talked about their kids. Su shared some of the challenges he faced when he first came to the United States for college and he knew very little English. Andres reciprocated with stories about some challenges he faced growing up in Detroit, not always having enough money to pay all the bills at once, and some things that happened when he was younger that tripped up his plans to go to college earlier in his life.

Su says he enjoyed Andres' perspective — in part, because Su is someone who’s interested in people, and the ups and downs of Andres’ life made for a compelling personal story. In addition, Su says he’s always been interested in Detroit, and through Andres, a lifelong Detroiter, he enjoyed learning about a side of the city that “you don’t hear about on the TV news.” Interestingly, at a certain point, both men noticed that many of their personal conversations weren’t actually that separate from their field of study: energy systems. For example, Andres shared stories about how, in his family, when the power would go out, that meant you didn’t open the fridge so his grandfather’s insulin wouldn’t spoil. Or how, if the power was out long enough, they’d start barbequing the meat they had in the freezer so it wouldn’t go to waste. Or how he and his relatives used to take turns refilling his grandmother’s boiler when she couldn’t afford to fix it right away. “The biggest burden a lot of my people have is energy, whether it’s fuel, gassing up your car going to work, or just trying to keep your house warm,” Andres says. “I know what it’s like to struggle to pay a utility bill and choose between paying that bill on time or buying food. I know people whose furnaces go out and can’t afford to get them fixed. That’s the impact energy has. Energy helps. But if you don't have a lot of money, energy can be a burden.”

Andres’ testimony left an impression on Su, and eventually the two of them started kicking around ideas for projects focused on energy justice, an emerging field that focuses on equity, affordability, accessibility and participation in the energy system and the ongoing transition to new energy technologies. That lens became the cornerstone of their 2023 Summer Undergraduate Research Experience project, in which Andres helped create an interactive mapping tool that explores the connection between historical real estate redlining and disparities in energy access in metro Detroit. Using Andres’ dashboard, a user can quickly toggle through different layers and explore whether a neighborhood is formerly redlined, where current utility outages are and their causes, locations of renewable energy installations and EV charging stations, as well as census data that provides demographic information about the people currently living in the neighborhood. One of the big reveals from Andres’ work is that metro Detroit neighborhoods that were historically redlined tend to be lower-income, non-white, have power outages that tend to be longer, and have less access to renewable energy and EV charging infrastructure. Their work on the app led to an IEEE conference publication in 2023.

Andres shows off an interactive mapping tool he helped create that explores the connection between historical real estate redlining and disparities in energy access. Photo by Annie Barker

After the project, Andres was nearing the end of his undergraduate career, and Su began asking him a little more frequently and persistently if he’d ever thought about grad school. Andres says that honestly wasn’t on his radar when he initially transferred to ��-Dearborn. At that time, his thinking pretty much revolved around the straightforward calculus that getting an engineering degree would help him get a higher-paying job and lead to a more comfortable life for him and his family. But after working in Su’s lab, particularly on the energy justice-focused mapping project, he says his perspective began to change. “If I did the safe thing and got a job, yeah, that would help me. But working with Professor Su helped me see what it is that I really wanted to do. And what I really wanted to do was help people,” Andres says. To do that, Andres decided to take a chance and continue his education, enrolling in the department’s doctoral program, where he’s now continuing his work alongside Su. Andres is still just in his first year of the program, but the pair have already started discussing ideas for Andres’ dissertation research. Not surprisingly, energy justice topics are rising to the top of their list. One idea they’ve been kicking around recently is creating a machine learning-based app that would use weather and outage data to forecast the likelihood of power outages, sort of like the snow day prediction models that have become super popular with parents and kids. “A technology like this could have multiple applications,” Su says. “It could help a utility identify where they need to upgrade their infrastructure to make it more resilient, or where they might need to deploy the most crews ahead of a storm. But it could also directly help people. Like, if you saw that there was a 90% chance you were going to lose your power for five hours tomorrow, people could charge their batteries or make a trip out to get food.”

Andres says it’s still a little wild to think he’s in a PhD program, and the future he’s contemplating for himself now is much different than the one he was thinking about a few years ago. For the time being, he’d love to continue to do research in the energy justice field and develop innovative ways to help people. He also wants to start a nonprofit that gets young people excited about things like sustainable energy technologies and smart cities. Ultimately, if all goes well, he’d love to be a college professor just like his mentor. “It’s funny, the opportunities that come up in your life,” Andres says. “Like, what if I had ignored that email? What if I hadn’t stepped out and emailed professor Su? He saw something in me that I didn’t even know about myself. And he probably doesn't even know how many other people that’s going to touch. My daughter and little cousins, they all get to see that, ‘Hey, Uncle Rico is gonna be a doctor.’ Because of Professor Su, I’m setting a new example for the people in my life. My wife, sister and other family members, everybody is, like, ‘Wow, you was always smart.’ I just went the long way. And I’m just glad I have an opportunity now to make an impact in the community.”

That a college professor would have such a profound impact on a student isn’t super surprising. It’s what many faculty shoot to achieve every day. But Su says Andres has also had a big influence on him. On a personal level, he finds Andres' life so inspiring, and admires his work ethic, the personal sacrifices he’s making to be part of the PhD program, and his vision for using research to help people in his community. Their relationship has also had a big influence on Su’s own work. “For example, losing power for one hour in a wealthy neighborhood compared to a disadvantaged neighborhood, it has totally different impacts. Or if you lose power, then that suddenly becomes a food issue for some people. I never really thought about stuff like that before,” Su says. “Now, I'm thinking more about how our research has the potential to help people. Because, otherwise, maybe you’re just publishing a paper with some very fancy solutions and complicated equations. But how are you helping average people?”

###

Story by Lou Blouin

College of Engineering and Computer Science

Electrical and Computer Engineering

Off

Mon, 02/24/2025 - 12:56

Unlikely collaborators who met through a job posting, first-year PhD student LaRico Andres and Professor Wencong Su have forged a partnership that’s left a lasting mark on both men.

Electrical and Computer Engineering Professor Wencong Su (left) and doctoral student LaRico Andres have teamed up on a number of research projects focusing on energy justice. Photo by Annie Barker

Researchers prep for landmark field test of ‘second-life’ EV batteries

lblouin — Mon, 10 Feb 2025 15:30:35 +0000

Researchers prep for landmark field test of ‘second-life’ EV batteries lblouin Mon, 02/10/2025 - 10:30

Connecting large batteries to the electric grid is an idea that has multiple benefits. Most fundamentally, an electricity system powered by an increasingly large share of solar and wind will almost surely require storing energy in some fashion, given the variability of the wind and sunshine at any particular moment in the day coupled with our expectation that power will always be available. Indeed, large battery storage systems are already being deployed across the world in sophisticated ways. For example, a Massachusetts town of about 27,000 people is building a 15 MWh, grid-connected battery storage system that can help power the town during power outages, stock up on less-expensive energy during off-peak hours and feed electricity back to the grid during times of peak demand, when utilities generally have to fire up their most carbon-intensive diesel-powered backup systems. Grid-tied battery storage systems, however, come with a few caveats. For one, it takes a lot of batteries to store enough energy to actually make a significant difference, which means battery storage systems are expensive. Moreover, if you consider the climate emissions involved in manufacturing the batteries and the mining for the lithium and cobalt needed to make them, battery storage systems are far from being climate neutral.

Throw at this predicament a seemingly too-good-to-be-true solution: Right now, electric vehicles represent the largest use of large batteries, and EVs are steadily gaining market share across the globe. However, manufacturers estimate that batteries in new EVs will likely need to be replaced after eight to 12 years, or when they can hold about three quarters of their original charge. These slightly diminished batteries, however, could still be useful for other things, like — you guessed it — grid-tied storage. Now add into the mix the fact that EV sales have been strong for the past few years and we could be a decade or so away from having millions of used batteries ready for a “second life” in the electric grid. Importantly, this approach could reduce the costs and climate impacts of building large storage systems compared to those that use new batteries.

Undoubtedly, it’s an idea with tons of promise. The catch, at least for now, is that we don’t really know how well it will work, according to Associate Professor of Electrical and Computer Engineering Xuan Zhou. Zhou says our understanding of the performance of new EV batteries is pretty solid and getting better all the time. However, we simply don’t have the same level of data about what’s really going on inside used EV batteries to say for sure they’ll be as reliable an option for grid storage as we hope they’ll be. The reason is that charging and discharging a battery hundreds of times — and subjecting it to thousands of hours of road vibrations — changes its underlying chemical and physical properties. That’s ultimately why it loses capacity over time. And many details about a battery’s potential second life inside a grid storage system — like how many years it would last, how safe it would be and how cost-effective it would be — are still largely unknown.

Zhou, along with his Electrical and Computer Engineering Department collaborators Associate Professor Mengqi Wang and Professor Wencong Su, hope they’re about to provide the world with some of the best data-backed answers to date. With a new $1.48 million grant from the Michigan Department of Environment, Great Lakes, and Energy, the largest state grant in the university's history, the team of researchers is set to build a 500 kW grid-tied storage system that will use actual used EV batteries.

Su and Wang say some of the best work in this field is currently being done at Stanford University. But those experiments involve taking new EV batteries and subjecting them to hundreds of charging and discharging cycles in order to intentionally deplete them in a way that mimics what happens in an EV under certain driving conditions. “For the emulation, however, they might not be able to emulate the true environment the batteries would be operating in,” Wang says. “We can imagine, in an EV, there will be vibrations. There will be accelerations and decelerations. The devices around the pack might also generate heat. And there will be variations in the temperature and all these environmental conditions for the battery pack — they will not be constant at all. This all impacts the battery. So by using real used EV batteries, we hope our data will end up being more practical.”

From left, Associate Professor Xuan Zhou, Associate Professor Mengqi Wang and Professor Wencong Su.

Zhou says they are still working out many details for the design and build of the project. One of the unexpected challenges was simply locating used EV batteries. “Initially, it looked like we were going to have to get them from Europe, where EVs have been more popular,” Zhou says. “But now it looks likely that we will be able to get about eight to 20 battery packs from used GM vehicles.” After they secure the batteries, Zhou says they’ll perform extensive tests to get a sense of their chemical and physical properties. Meanwhile, Wang and Su will work out how the batteries can be connected and tied to the grid. Because the team is planning to test several types of use cases, this will require a carefully planned system of custom power converters, controllers and control algorithms. For example, they want to see how well the batteries perform in a “peak shaving” role. That’s when the batteries would feed electricity back to the grid during times of peak demand, which helps utilities avoid using their expensive, carbon-intensive backup systems. They also want to test “load shifting,” which is where the batteries would charge during non-peak hours and then release that energy during peak hours, when electricity is more expensive. Another variable: the partner site that will host the storage system has a solar array. Wang says connecting all those components and systems in a way that doesn’t produce harmonics on the grid, damage equipment or compromise power quality takes careful planning. Many of the components will have to be custom built, too, since this kind of system is not an off-the-shelf technology.

Another big hurdle the team might encounter is testing the system. At 500 kWh, this project is so big that the current testing facilities at ��-Dearborn — or any other university the team knows of — isn’t equipped to handle a test of the full system. Zhou says the new battery lab in Ann Arbor may be able to provide single pack-level testing, and they may need an industry partner for tests beyond that. Wang says the size and testing requirements also make it more challenging to design other aspects of the system. “I think the piece of equipment that has the highest power rating is our battery simulator in our IAVS lab, which has 100 kW testing capacity. So we’ll start with a smaller system as a proof-of-concept design. And then we’ll scale up to the real-world test. But this makes it much more challenging. The components, like the semiconductors or power converters, can’t just be scaled up. They’ll have to be swapped out with new hardware with higher power ratings. So we’re going to be busy!”

One other noteworthy feature of the project: After the study of the storage system is complete, the research team’s Genesee County-based industry partner on the project, ReCharge ReCycling, which is helping design and build the storage system, will help prepare the battery pack for a third life. “That’s why we’re calling this project ‘Closing the loop,’” Zhou says. “They’ll be dismantling the battery down to the cell level and then trying to recover the precious metals so they can be used to make a new battery. So we’ll really get a full picture of what it’s going to take to take a battery through this full cycle.”

###

Story by Lou Blouin. Photos by Annie Barker.

College of Engineering and Computer Science

Electrical and Computer Engineering

Off

Mon, 02/10/2025 - 15:30

Connecting millions of used electric vehicle batteries to the grid is an idea that holds tons of promise. But how well will this strategy work in the real world?

Associate Professor Xuan Zhou (right), Associate Professor Mengqi Wang (left) and Professor Wencong Su are heading up a new project that will tie used EV batteries to the electric grid.

‘��-Dearborn is a respected institution . . . I feel very lucky to study here’

stuxbury — Sun, 02 Feb 2025 18:08:09 +0000

‘��-Dearborn is a respected institution . . . I feel very lucky to study here’ stuxbury Sun, 02/02/2025 - 13:08

Sambathchatovong Na is ��-Dearborn’s first Fulbright Scholar studying in a doctoral program. This point of pride can be added to other experiences he’s had. Na, who goes by the first name Vong, is a volunteer researcher with the United Nations Educational, Scientific and Cultural Organization, commonly known as UNESCO, through their Asia-Pacific Regional Center of the CultureLink Network. Na has also heard President Barack Obama speak — and shook his hand.

Na, who has a reserved, humble nature, appreciates the experiences that he’s had. He sees them as milestones on his path to help others.“President Obama spoke about how we need to work together to effectively solve problems and he encouraged us to not sit back, but to get involved,” says Na, a cybersecurity professional who was selected as a Young Southeast Asian Leaders Initiative fellow in 2015. This is a still-active program created by the Obama administration that aims to build leadership capabilities and cross-border cooperation for solving global challenges. “I want to be a part of advancing cybersecurity technology — which changes second by second — while also helping shape how it is used through public policy.”

Na is learning to do this through his latest endeavors — studying in the College of Engineering and Computer Science’s Computer Information Systems doctoral program as a Fulbright Scholar, with CIS Professor Di Ma as his advisor. Ma is a cybersecurity and information assurance expert.

Na wants to strengthen his country’s education and legislation in the always-evolving industry by seeing how other nations fight hackers and enforce the safety measures in ways that work for the citizens.“Cambodia, my country, has not passed cybersecurity laws to protect our infrastructure yet. It’s in the process. I’m hoping, with my combined experiences, that I’m a person who can assist with this,” he says. “I am very appreciative that I am here to earn my degree and Dr. Ma is helping me meet my research goals.” His goals include drafting a cybersecurity framework for Cambodia that balances government and citizen needs when it comes to security and privacy, along with appropriate enforcement. Na also hopes that his degree will assist him in contributing to Cambodia’s foreign policy direction in cyber diplomacy.

Na has worked in the cybersecurity field in Cambodia for nearly a decade, both in the private and public sector. He established information technology security and IT governance departments for a bank in his country, Na has served as a cybersecurity-advising official for Cambodia's deputy prime minister cabinet. And Na teaches at two universities — American University of Phnom Penh and Royal University of Phnom Penh — to develop the next generation of Cambodia's cybersecurity experts and grow interest in the field.

“I started with an interest in the technology aspect, but the more we learn and things change — like AI and data-science mining — the more I’ve become interested in the public policy surrounding it,” he says, noting that the U.S. is a good place to study because there are many U.S. protection laws related to cybersecurity.

When applying for the Fulbright award, Na had his sights set on ��-Dearborn after reading about the university’s faculty expertise in the computer science field and seeing the number of research publications. “I looked for the best programs in the United States. ��-Dearborn is a respected institution in my field with professors who have published research in prestigious journals and have connections to industry. I feel very lucky to study here,” he says.

Na earned his undergraduate degree from Royal University of Phnom Penh and a master’s degree in cybersecurity from the University of Southampton in England, where he attended as a Chevening Scholar, which is a fully funded award to study in the United Kingdom.

“I had dreams to study in other countries for programming skills, networking and learning how other countries do things so I can help my country build a framework for cybersecurity. The resources for education in cybersecurity are limited in Cambodia, so I kept looking for ways to go out and bring what I learn back with me so I can teach others,” Na says. “It’s hard to study abroad when you don’t have the money. I kept applying even when I didn’t have success at first. You don’t give up on your dreams.”

Na’s parents must have known about their son’s commitment and drive from the beginning — as evidenced by his name. “My first name, ‘Sambathachatovong,’ is three words put together,” Na says. “‘Sambath’ symbolizes a richness or a wealth. ‘Chato’ is because I am the fourth born. ‘Vong,’ which is short for ‘Vong Trokaul', is for family tree. It’s about working together to bring about a richness. It doesn’t need to be about money, it can be about knowledge or community.”

Now in his second semester at ��-Dearborn, Na says he has met friends, enjoys the campus and has the best of all worlds by being a part of the U-M system. “Both campuses are beautiful and provide very good, but different, things. Going to Ann Arbor is great for being in a busy place and meeting up with friends. Staying in Dearborn is nice because it is peaceful and welcoming and your professors and colleagues look out for you.”

In addition to traveling to Ann Arbor — Na first took the train, which is walking distance from campus, but now has a car — he’s planning to see other parts of Michigan. He especially wants to travel to Michigan’s Upper Peninsula to see its forests, rock formations and waterfalls.

Na shares his fun Michigan adventures and plans with his family and friends in Cambodia and looks forward to the day when he can share his completed research work too.“I’ve been building layers of education for many years to help me become a qualified person to help advise my country,” he says. “I am grateful for all of my experiences, but my time at ��-Dearborn is especially important. Earning my PhD here is a big step forward when it comes to reaching my goals.”

Story by Sarah Tuxbury

Graduate Research

International

College of Engineering and Computer Science

Graduate Studies

Off

Sun, 02/02/2025 - 18:07

Meet Sambathchatovong Na, ��-Dearborn’s first doctoral Fulbright Scholar. He’s a cybersecurity professional from Cambodia who’s in CECS’ computer information systems program.

Sambathchatovong Na, who goes by Vong, is ��-Dearborn's first doctoral Fulbright student. Photo by Annie Barker

How can we make driver assist features play nice together?

lblouin — Mon, 27 Jan 2025 13:14:53 +0000

How can we make driver assist features play nice together? lblouin Mon, 01/27/2025 - 08:14

The periodic cycles of excitement and disappointment over the emerging future of fully autonomous vehicles sometimes overshadow the reality that our vehicles are steadily getting more intelligent. Blind spot monitoring, which is often standard in new cars, will give you an alert anytime a driver (and maybe even a pedestrian) enters an area that generally represents the vehicle’s blind spot. Adaptive cruise control automatically maintains a preset distance to the car in front of you. Lane keep assist varies by vehicle, but in vehicles with electronic power steering, it can nudge your vehicle back between the lane markers if you start to drift too far. Autonomous emergency braking forcefully engages the brakes if your car detects an imminent frontal collision. And a total of eight 2024 models available in the U.S. offer some version of hands-free driving under certain conditions. This is on top of more convenience-focused technologies, like cars that parallel park for you and those with optional one-pedal driving, where the driver brakes and accelerates with a single pedal.

Assistant Professor of Industrial and Manufacturing Systems Engineering Areen Alsaid says these driver assist features have the ability to reduce driver workload and make driving safer. But as an expert in human-technology interactions, she’s also been wondering if the learning curve associated with this growing list of advanced features could limit their benefits. “For example, I have a vehicle that has both lane keep assist and adaptive cruise control, and I was under the impression that you couldn’t have them both engaged at the same time, but I tried it and I totally could,” she explains. “That sort of made it like an AV, because the car was controlling both its speed and lateral orientation within the lanes. I definitely did not expect that.” She says sort of like an AV, of course, because, in her vehicle, she discovered these features only remain engaged as long as she has her hands on the steering wheel. But for Alsaid, the experience underscores an important caveat about new driver assist features: When drivers are purchasing a vehicle equipped with several of these advanced technologies, they don’t automatically understand how they work, what their limitations are, and if features will interact with each other in ways they find intuitive. It’s only once drivers become more familiar with how these technologies function that they can fully reap their benefits.

Doctoral student Duha Alkurdi is helping Alsaid with a new project that explores how drivers respond to using multiple driver assist features simultaneously. Photo by Julianne Lindsey

Alsaid says there is a growing body of research on individual driver assist technologies, but researchers are just scratching the surface on understanding the various scenarios when multiple features are operating simultaneously. As part of a new project funded by Toyota, Alsaid is hoping to build some foundational knowledge in this space by studying how drivers behave as they use multiple longitudinal features — technologies that, as contrasted with lateral features like lane keep assist, control the forward momentum of the vehicle. These include things like adaptive cruise control, one pedal driving, forward collision warning, autonomous emergency braking and hands-free driving at low speeds. In the first stage of her study, she’ll observe 10 participants as they engage multiple driver assist features in real vehicles on real roads, monitoring their behavior with cameras and real-time questions, as well as biometrics, like eye tracking and heart-rate monitoring. Using some of the initial insights from this experiment, she and her student research assistants will then design a full driving simulator experience, where dozens more drivers will be confronted with challenging driving situations.

Alsaid says after this initial study, the team will review the results and try to identify areas where the interactions between features and the user interfaces can be made more intuitive for drivers. Her hope is that their work in this area can help drivers adapt more smoothly to these new features so they can enjoy the full potential of their smarter new cars.

###

Story by Lou Blouin

College of Engineering and Computer Science

Industrial and Manufacturing Systems Engineering

Off

Mon, 01/27/2025 - 13:14

Assistant Professor Areen Alsaid says the growing suite of semi-autonomous technologies have the ability to make driving safer, more convenient and less stressful. But we’ll only fully realize these benefits if the learning curve doesn’t confuse drivers.

Assistant Professor Areen Alsaid (right), pictured here with doctoral student Shiva Rasouli, has been using a driving simulator to study how drivers respond to interactions between driver assist features. Photo by Julianne Lindsey

Planning for EV infrastructure is complicated. Here’s a better way to do it.

lblouin — Wed, 22 Jan 2025 13:45:53 +0000

Planning for EV infrastructure is complicated. Here’s a better way to do it. lblouin Wed, 01/22/2025 - 08:45

Electric vehicles still represent less than 1% of registered vehicles in the U.S. and comprised just over 8% of new U.S. vehicle sales in 2024. But government officials across the country are already planning for a future in which EVs are the go-to mode of private transportation — as evidenced by increasingly large public investments in EV charging infrastructure, like the $7.5 billion pledged under the Biden administration. When it comes to actually making a plan for building out that infrastructure, however, the devil is in a complex web of largely unknown details, says Associate Professor of Industrial and Manufacturing Systems Engineering Jian Hu. Ideally, Hu says, you’d want to build a charging system that serves people’s needs for the next 10 years. The tricky part is that involves correctly anticipating a host of trendlines which are basically impossible to predict. “There are so many uncertainties,” Hu says. “We don’t know how fast the EV market will grow, which means we don’t know what the charging demand will be. Will the public want fast or slow charging? Will people charge at home or at work? During peak hours, or non-peak hours? Will utilities adopt dynamic pricing so it’s cheaper to charge at certain times of day? And how will this affect charging behavior? And what if there are new technologies that emerge? There are so many things we don’t know.”

Typically, when government agencies are planning for things like this, Hu says they might use what statisticians call a deterministic model. When you’re making this kind of model, you take account of what data you have, plug it into a formula that defines the relationship between the variables, and then you typically get a single answer, which then guides your plan of action. A simple retirement calculator, which takes your current age, retirement age, anticipated rate of return, the rate of inflation, and monthly savings target and tells you how much money you’ll have when you retire is a simple kind of deterministic model. But Hu argues that these kinds of prediction models aren’t the best tools when you’re dealing with complex situations with lots of uncertainty. “What if your rate of return on your investments isn’t the 7% you’ve put into that calculator? What if it is less and then you don’t have the money you expect when you retire? If you only use this kind of model, you’ll have this single prediction and you base everything on that. But you’re not preparing for a possible worst-case scenario.”

Because of this, statisticians often use different tools for forecasting complex, highly uncertain phenomena, like investment performance. One of the most popular is called stochastic modeling, which, instead of giving you just one answer, provides a range of results represented by thousands of different simulated scenarios in which the values and probabilities of the underlying variables are allowed to change. A popular type of stochastic model is a Monte Carlo simulation, which is used in financial planning. Unlike a simple retirement calculator, which assumes fixed values for things like the rate of inflation or a rate of return, a Monte Carlo analysis will run thousands of simulations in which all the variables you have in your model have different values and assigned probabilities. As a result, you get a much wider representation of possible outcomes, including best-case, worst-case and most likely scenarios. A Monte Carlo simulation might tell you, say, that in 98 percent of its simulated scenarios, which account for a wide range of possible economic conditions, your savings plans are going to get you to the end of your life with money in the bank. That might give you much more confidence in your investment strategy. You might even decide to take a more aggressive approach. Or, as Hu frames it, stochastic modeling gives us a much more complete way to understand risk.

Because stochastic models are so much better at helping us understand risk of uncertain events, Hu says they could be an excellent approach for planning EV infrastructure, given that reliable data on important details, like consumer charging preferences and future electricity demand, are sparse or lacking altogether. As part of a National Science Foundation-funded project, and in collaboration with Argonne National Laboratory, which is itself partnering with the Chicago Transit Authority, Hu will be creating novel approaches to stochastic modeling that he hopes can profoundly improve this planning process. To get just a little bit technical for a moment: Hu says one of the big challenges of stochastic modeling is coping with “distributional ambiguity,” which is where you’re lacking a lot of reliable information about the probability of key phenomena. Currently, one of the ways statisticians deal with this is a method called distributionally robust optimization, but Hu says DRO has a few limitations. One, it’s computationally intensive, which means as models get more complex, it becomes increasingly cumbersome and time consuming to run them and adapt them to new data. Second, DRO tends to hedge against worst-case scenarios and can therefore lead to overly conservative outcomes. Third, it's a machine learning-based process that is a “black box,” meaning it does not show how it arrives at its answers. It therefore might not be a great option for publicly funded agencies, who, in spending large sums of taxpayer dollars, could certainly benefit from transparent or interpretable models. Hu says his novel stochastic modeling strategy aims not only to improve on these limitations, but also provide more overall usability for public agencies, including models that can be more easily updated as better data emerge.

After developing the model, Hu is looking forward to a big field test toward the end of the project: helping the Chicago Transit Authority as it embarks on electrifying their massive fleet of public buses. Current electric bus battery technology generally provides a range of around 60 miles. But many of CTA’s current scheduled vehicle route blocks are much longer; nearly a third of weekday blocks are longer than 100 miles. CTA therefore envisions building out a bus-charging infrastructure that’s integrated with their garages and bus routes, which will also require enhancements to the local electric grid to accommodate the new electricity load. Figuring out where these grid modifications need to be made and the ideal placement for the charging stations in a way that harmonizes with the complex bus routes — which are likely to change in the future — is exactly the kind of complicated, ambiguous planning task Hu says his model can help with. Most importantly, it gives a government agency its best chance of spending limited public resources for EV charging as effectively as possible.

###

Story by Lou Blouin

College of Engineering and Computer Science

Industrial and Manufacturing Systems Engineering

Off

Wed, 01/22/2025 - 13:40

Given a web of uncertainties, like future electricity price, consumer charging behavior and EV market growth, Associate Professor Jian Hu says a novel type of stochastic modeling can minimize the risks of public investments in EV infrastructure.

Associate Professor of Industrial and Manufacturing Systems Engineering Jian Hu is testing a new approach to planning charging infrastructure with the Chicago Transit Authority. Photo by Annie Barker

Explore Dearborn (2.17.25)

Anonymous — Tue, 21 Jan 2025 21:52:12 +0000

Explore Dearborn (2.17.25) Anonymous (not verified) Tue, 01/21/2025 - 16:52

At the President's Day Exploration Event and Tour, you will learn about the exciting opportunities and programs available at ��-Dearborn. You will have the opportunity to learn about our practice-based learning approach, research opportunities, and unique offerings in our various degree programs. Interact with staff and current students to hear first-hand the incredible experience ��-Dearborn provides. This event also features an optional 45-minute walking tour of campus. This event is designed to showcase our academic programs and see what ��-Dearborn has to offer. For a complete list of majors, visit: /academics/undergraduate-studies/majors-and-minors.

2025-02-17T10:00:00-0500 to 2025-02-17T12:00:00-0500

United States

Office of Undergraduate Admissions

313-593-5100

umd-admissions@umich.edu

Prospective students

College of Engineering and Computer Science

Info Session

The challenges with younger drivers and autonomous vehicles

lblouin — Mon, 13 Jan 2025 15:28:25 +0000

The challenges with younger drivers and autonomous vehicles lblouin Mon, 01/13/2025 - 10:28

Autonomous vehicles promise a future of almost unimaginable driver convenience. After all, who wouldn't want to be able to watch a movie during a tedious stop-and-go rush hour commute? But proponents of AVs think the technology’s even more meaningful impact will be in the realm of safety, given that more than 90% of traffic accidents are thought to be the result of human error. Indeed, we’re already seeing some of the safety benefits as semi-autonomous driver assist features, like lane keep assist, blind spot monitoring and emergency braking, become commonplace on newer vehicles. Even so, the next big developmental leap to so-called Level 3 vehicles, which are capable of doing all the driving for us in certain situations and are now starting to enter American and European markets, could introduce a whole new set of safety and driver behavior challenges, says Associate Professor of Industrial and Manufacturing Systems Engineering Bochen Jia. That’s because Level 3 AVs still require full driver intervention when conditions are beyond the system’s capability — like when lane markers aren’t clearly marked or are obscured by weather; or in the event of system malfunction. That introduces a whole new type of driver experience, which the AV research community calls “takeover.” That’s the tricky liminal moment when a driver has to abruptly shift their focus from, say, watching their favorite Netflix series, to dealing with a potential road emergency.

Intuitively, takeover is a fraught, psychologically complex space, which is why Jia says a lot of research has been devoted to studying how drivers respond to this type of demand. For example, early research has indicated, somewhat counterintuitively, that the risk of driver fatigue actually increases when operating a Level 3 vehicle. It seems the mental task of constantly monitoring a vehicle doing the driving for us can actually be more draining than doing the driving ourselves. And in Jia’s lab, he’s been focusing on a neglected area of takeover research: younger, inexperienced drivers. From a research perspective, he says it’s a fascinating population. First, young people’s brains are still developing, particularly when it comes to attention resources and situational awareness, which can influence things like fatigue development and how well they respond to potentially hazardous situations, like takeover. In addition, younger people tend to, in general, be more trusting of technology, which could lead to less intense monitoring of the vehicle and undesirable takeover outcomes. Moreover, Jia says we might not be too far away from an era where young drivers’ only driving experience is on Level 3 vehicles. “That would mean they actually aren’t getting much experience doing actual driving, which could influence how well they respond to situations when the vehicle asks them to take over,” Jia says.

In focusing on younger drivers, Associate Professor of Industrial and Manufacturing Systems Engineering Bochen Jia has been exploring a neglected area of AV research. Photo by Annie Barker

For one of his latest projects, Jia has been digging into the particular issue of fatigue onset in younger, novice drivers. For this study, he recruited both drivers under the age of 19 who had limited driving experience and middle-aged drivers who had years of experience. Using a high-fidelity driving simulator, both sets of participants engaged in extended periods of autonomous driving and were instructed to take control of the vehicle when the autonomous mode was not safe or had failed. During their driving experience, Jia then introduced several types of takeover events to see how drivers would react. By combining physiological measures and self-reported assessments collected periodically during their driving task, Jia then monitored each set of participants for fatigue onset. The results were pretty interesting: Novice young adult drivers developed both physical and mental fatigue earlier than their middle-aged counterparts during extended periods of autonomous driving. "This is somewhat surprising to people because they think young people naturally have more energy," Jia explains. "However, it appears that the mental demands of autonomous driving, or the tendency of younger drivers to be more engaged in non-driving tasks during autonomous operation, may accelerate the onset of fatigue in younger drivers." Moreover, the takeover outcomes for fatigued drivers were often far from desirable. “Essentially, we were looking at how often participants would fail to prevent a collision,” Jia says. “And it happened more frequently than we had hoped.”

Jia says the results point to a need for highly effective fatigue warning systems for Level 3 AVs that have special considerations for younger, inexperienced drivers. Interestingly, Jia envisions this system would be most effective if it wasn’t just part of the vehicle, but integrated with other readily available wellness monitoring technologies, like wearable fitness trackers that monitor heart rates and sleep quality. “I don't think you can have a system that’s confined to the car — that’s not enough,” Jia says. “It should be part of your life monitoring — an overall system — because they don’t suddenly become a different person when they get in the car. With all the personal data monitoring technologies we have nowadays, we could easily identify younger drivers, and we likely have a lot of the data we would need.”

Next up, Jia says he plans to continue his research on fatigue onset with a study that puts participants in real automated vehicles within a closed testbed, such as the Mcity test facility. In addition, he has a similar project that is studying differences in distraction levels between younger, novice drivers and middle-aged experienced drivers. He hopes both studies will provide critical insights into improving the safety and usability of autonomous vehicle systems for drivers of all experience levels.

###

Story by Lou Blouin. Want to learn more about Jia's work on driver fatigue? You can read the full-text of his recent article on the subject in the “Journal of Safety Research,” which was co-authored by Jia's student Yourui Tong, ��-Dearborn Associate Professor of Industrial and Manufacturing Systems Engineering Shan Bao, Tsinghua University Professor of Industrial Engineering Changxu Wu and ��-Dearborn Associate Professor of Psychology Nitya Sethuraman.

College of Engineering and Computer Science

Industrial and Manufacturing Systems Engineering

Off

Mon, 01/13/2025 - 15:22

Associate Professor Bochen Jia’s latest research is revealing some surprising things about how fatigue can impact younger, inexperienced drivers when they're operating semi-autonomous vehicles.

Associate Professor of Industrial and Manufacturing Systems Engineering Bochen Jia (right) and graduate student Darlington Egeonu check out the driving simulator Jia used for his most recent project on driver fatigue. Photo by Annie Barker

‘Healing’ batteries with ultrasonics

lblouin — Wed, 08 Jan 2025 17:59:20 +0000

‘Healing’ batteries with ultrasonics lblouin Wed, 01/08/2025 - 12:59

High-capacity batteries have emerged as an essential building block of the clean energy future, but it likely won’t be today’s lithium-ion batteries alone that get us there. That’s because, as powerful and ubiquitous as they are, lithium-ion batteries have some major limitations. For starters, they are very heavy — a fundamental quality that curbs the range of today’s electric vehicles and basically rules them out as an option for powering commercial aircraft. The world’s growing demand for the minerals needed to manufacture lithium-ion batteries also requires ever expanding mining operations, which have environmental, climate and geopolitical consequences. And on the safety front, the liquid electrolytes used in lithium-ion batteries are highly flammable, which is one reason why electric vehicle battery fires can be so catastrophic.

Because of these issues, researchers are developing new kinds of batteries, including solid-state batteries, which are lighter, can hold more energy for their weight and require fewer materials. They are also considered safer, since, as the name suggests, they don’t require a liquid electrolyte. But mechanical engineering postdoc research fellow Yaohong Xiao, who works with Associate Professor Lei Chen, says that last quality is both a blessing and a curse. Inside a typical lithium-ion battery, the liquid electrolyte is responsible for transferring charged particles back and forth between the cathode and anode sides of the battery, which is necessary both for charging and discharging. Because this electrolyte is a liquid, Xiao says it creates a nearly uniform interface between the two sides of the battery, since liquids naturally fill tiny nooks and crannies that might exist between the anode and cathode materials. In a solid-state battery, however, the electrolyte separating the anode and cathode sides is a strong, stiff, solid piece of material, and Xiao says that makes it more difficult to get really good contact between the two sides. “Essentially, you have solid on solid on solid,” Xiao explains. “Right after manufacturing, you might have no problem with this interface, but after a few cycles, or after the battery has experienced vibrations from being on the road, you start to get little voids in the interface so the electrolyte is no longer making perfect contact with both sides of the battery.” As a result, Xiao says a battery can develop zones of higher than usual current, which reduces its overall effectiveness.

Typically, Xiao says researchers have tried to solve this solid-state battery interface problem in a couple of different ways. First, they use pressure to basically squish all the components back together. Or they can use an “interlayer” material, which, kind of like double-sided tape, keeps the anode and cathode materials snug up against the solid electrolyte. But each approach has its limitations. Xiao says introducing interlayers to a process is typically expensive. That might not be a huge deal for experiments in a lab, but it would increase the already high costs of solid-state batteries when they’re being manufactured at scale. On the other hand, Xiao says using pressure to smooth out the interface is sort of like trying to squish two pieces of wood together. The rigidity of the lumber and any imperfections along the surfaces mean that pressure alone often isn’t enough to create perfect contact.

A couple years ago, Xiao began kicking around a sort of unconventional solution for this interface challenge. Xiao’s background is actually in metallurgy, not battery chemistry, and he saw this issue primarily as a question of how to get different metals to stick together. As it turns out, this is actually a pretty common problem in the metals universe, and one with a variety of well-established solutions. Specifically, Xiao was thinking that a technique called ultrasonic welding might be able to “heal” the voids in the interface, restoring uniform contact with the electrolyte. Ultrasonic healing, which is commonly deployed in a wide range of industries to weld pieces of plastic or thin pieces of metal together, uses high-frequency acoustic vibrations instead of high temperatures or solders to join materials together. The result is a highly uniform bond, but the process doesn’t use a whole lot of energy and can be done very quickly. Xiao worked with his postdoc colleague in Chen’s lab, XinXin Yao, who helped provide the theoretical feasibility for the welding concept, to set up an experiment, which revealed that ultrasonic welding could indeed nearly completely restore the uniformity of the electrolyte interface in as little as a minute. Moreover, the process required temperatures barely warmer than the water in your hot water heater and used about as much energy as an old-fashioned incandescent light bulb.

Xiao says the effectiveness of this metals-based approach came as a surprise to many of his colleagues, most of whom are experts in battery chemistry. Now, he’s hoping the results, which were recently published in “Advanced Energy Materials,” will help inspire other researchers to further explore the technique’s full potential. He says one possible application would be to use this as a tune-up strategy for eventual solid-state battery-powered EVs, which could come in for a quick ultrasonic healing treatment and get their optimal range back. And one other surprise finding from his experiment is that ultrasonic welding actually increased the conductivity of the electrolyte, albeit temporarily. That result was particularly intriguing to some colleagues at ��-Ann Arbor, who recently visited the newly established Battery Manufacturing and Testing Lab for a firsthand look. “They were very interested in this point,” Xiao says. “The conductivity would slowly recover after treatment, but they were thinking, what if we can fix that? What if we can figure out a way to not let it recover? So it’s very exciting to see this research already inspiring further work in this area.”

###

Story by Lou Blouin

College of Engineering and Computer Science

Mechanical Engineering

Off

Wed, 01/08/2025 - 17:58

Two mechanical engineering postdoc research fellows have discovered a novel way to solve a pesky challenge with next-generation solid-state batteries.

In Associate Professor of Mechanical Engineering Lei Chen's lab, postdoc research fellow Yaohong Xiao has been working on developing solid-state batteries, a safer, more energy dense alternative to today's lithium-ion batteries. Photo by Annie Barker

College of Engineering and Computer Science

Is the generative AI hype bubble about to burst?

������-Dearborn earns R2 research designation

For a ������-Dearborn student and professor, the influence goes both ways

Researchers prep for landmark field test of ‘second-life’ EV batteries

‘������-Dearborn is a respected institution . . . I feel very lucky to study here’

How can we make driver assist features play nice together?

Planning for EV infrastructure is complicated. Here’s a better way to do it.

Explore Dearborn (2.17.25)

The challenges with younger drivers and autonomous vehicles

‘Healing’ batteries with ultrasonics

��-Dearborn earns R2 research designation

For a ��-Dearborn student and professor, the influence goes both ways

‘��-Dearborn is a respected institution . . . I feel very lucky to study here’