Elaine Liu leads the charge in new project.

MIT senior Elaine Siyu Liu does not possess an electric vehicle or any vehicle. However, she views the impact of electric cars (EVs) and renewables on the grid as integral parts of an energy puzzle she aims to unravel.

According to the U.S. Department of Energy, the count of public and private EV charging points has nearly doubled in the past three years, with many more currently in development. Users anticipate the convenience of plugging in at their leisure, charging up, and hitting the road. But what if the grid struggles to cope with this demand?

Electricity demand in the United States, which has long remained steady, has surged as a result of EVs, data centers powering artificial intelligence, and industrial activities. Grid planners are projecting a rise in electricity demand ranging from 2.6 percent to 4.7 percent over the next five years, as per data presented to federal regulators. The entire spectrum of stakeholders, from EV charging-station operators to utility-system operators, requires assistance in navigating a system experiencing rapid transformation.

This is where Liu’s expertise becomes critical.

Studying mathematics and electrical engineering and computer science (EECS) at MIT, Liu is particularly interested in distribution — the mechanism of delivering electricity from a centralized location to end-users. “I find power systems a suitable platform for exploring theoretical research as a practical tool,” she states. “I’m intrigued by this field because of my familiarity with the optimization and probability techniques applied to address such problems.”

Having grown up in Beijing, Liu relocated with her parents to Canada after middle school and enrolled in a prep school situated in Oakville, Ontario, about 30 miles from Toronto.

By chance, Liu got involved in a regional math competition and eventually initiated a math club. Nevertheless, the prevailing math culture at the school was a revelation for her. Reflecting on some students’ apparent aversion to math, she remarks, “I didn’t alter my perspective on math; rather, my perception of how others perceive math changed.”

Bringing her mathematical passion to MIT, Liu embarked on the first of two Undergraduate Research Opportunity Program projects with electric power system expert Marija Ilić during the summer following her sophomore year. Ilić, a joint adjunct professor in EECS and a senior research scientist at the MIT Laboratory for Information and Decision Systems, guided Liu in her research endeavors.

Envisioning Grid Scenarios

Since 2022, with support from the MIT Energy Initiative (MITEI), Liu has collaborated with Ilić on investigating the challenges confronting the grid.

Among these challenges is the integration of renewables into the energy mix. Shortages in wind or sunlight could lead to a delay in power generation. Should such delays occur during peak demand periods, the grid could face strain due to severe weather conditions and other unexpected events.

Conceiving the grid as an interconnected network comprising numerous components, any disruption in the network — for instance, a tree felling a transmission line — necessitates the redirection of electricity previously flowing through that line. This rerouting might overload other lines, potentially triggering a cascade failure.

“These scenarios unfold rapidly and have far-reaching consequences,” Liu explains. “Sudden blackouts could affect millions of individuals.”

In addition to managing a single downed line, Liu highlights the challenge posed by renewables’ unpredictability. “The crux is that we now have a substantial share of renewables, which are inherently less predictable. Gaps in wind or sunlight cannot be forecasted. In such events, where power generation falls short and demand outstrips supply, similar failures occur, but on a much larger and uncontrollable scale,” she elaborates.

Moreover, the fluctuating output from renewables introduces voltage variations. “When we plug in our devices, we expect a voltage of 110. However, due to oscillations, achieving that precise voltage is elusive,” Liu notes. “Therefore, providing adequate electricity is only effective when it matches the specific voltage level required, which poses a significant challenge.”

Liu and Ilić are developing a model to anticipate instances of grid failure. Working with European industry data and test cases provided to universities due to restricted access to proprietary data, Liu conducts experiments using a simulated power grid. “I run my simulations on a fictitious power grid,” she discloses. “The same tool could be adapted for deployment on an actual power grid.”

Liu’s model anticipates the evolution of cascade failures. For instance, a wind generator’s output could sharply diminish within an hour. The model assesses which substations and households will be impacted. “Once the need for intervention becomes evident, this predictive tool empowers system operators to strategize and intervene proactively,” Liu affirms.

Impacting Pricing and Power Dynamics

Last year, Liu shifted her focus to EVs, which present distinct challenges compared to renewables.

According to S&P Global’s 2022 report, legislators raised concerns that the U.S. Federal Energy Regulatory Commission’s (FERC) wholesale power rate structure unfairly disadvantaged EV charging station operators.

Aside from being charged based on kilowatt-hour consumption, some operators face heightened costs for electricity during peak demand periods. Even a few EVs charging during these peak hours could inflate operator expenses, regardless of their overall energy usage being modest.

Forecasting the power demand from EVs is more intricate than predicting energy requirements for activities like heating and cooling. Unlike fixed structures, EVs are mobile, complicating the estimation of their energy consumption at any given moment. “If users are dissatisfied with a charging station’s price or lengthy queues, they will seek alternatives,” Liu observes. “Determining the allocation of EV chargers is an issue many are grappling with currently.”

One potential solution involves FERC imposing directives on EV users regarding when and where to charge and the corresponding pricing structure. However, Liu does not find this a preferable strategy. “People generally dislike being instructed what to do,” she asserts.

Liu is exploring an optimized market-driven approach that would garner acceptance from major energy producers like wind and solar farms, as well as nuclear plants, down to municipal aggregators responsible for securing competitively-priced electricity and overseeing its distribution to consumers.

By analyzing the locations, movements, and behavioral patterns of daily-traveled EVs within Boston and other key energy hubs, demand aggregators could ascertain optimal charger placement and pricing for consumers, akin to how Walmart determines the price mark-up for wholesale eggs across diverse markets.” Last year, she presented her research at the MITEI annual conference and plans to submit a paper on market optimization analysis to a journal of the Institute of Electrical and Electronics Engineers this coming spring

Reflecting on her exposure to varying attitudes towards STEM during her formative years, Liu remarks, “The (prep) school had a prevalent notion that ‘math was for nerds,’ particularly girls, with a sentiment that questioned why one would shoulder additional academic burden. Over time, I learned to disregard these views.”

Post-graduation, Liu, the sole undergraduate researcher in Ilić’s MIT Electric Energy Systems Group, aims to pursue fellowships and advanced programs in EECS, applied mathematics, and operations research.

Based on her analysis, Liu posits that the market could effectively govern the pricing and availability of charging stations. Offering incentives for EV owners to recharge during peak hours instead of nighttime, when demand is at its zenith, could prevent grid strain and alleviate additional operator costs. “Consumers would still retain the freedom to select an alternative charging station if desired,” Liu emphasizes. “I advocate that this approach is feasible.”