Energy UROPs

These students’ projects are generously supported by the Friends of MITEI UROP, which includes MITEI members Shell, ExxonMobil, and Chevron, as well as MITEI donors.

Filter by

2024

Catherine Tang

Electrical Engineering and Computer Science

Headshot of Catherine Tang

Research projects

Transportation is the largest contributing sector to carbon emissions in the United States, highlighting the critical need for effective roadway interventions to mitigate its environmental impact. Evaluating the effectiveness of these interventions requires models that accurately capture human driving behaviors. However, modeling human driving presents significant challenges due to the complex interplay of factors such as inter-vehicle dynamics and human decision-making processes. We seek to build realistic human driver models using real-world human driving data and generative machine learning techniques, formulating the challenge as a next-token prediction task. Our next-token prediction model combines interactions between vehicle trajectories, road topology, and upstream traffic dynamics to achieve realistic and scalable agent-based traffic flow simulations. With a more realistic simulation of human driving behavior, researchers can study roadway interventions such as eco-driving to reduce the impact of emissions from transportation.

Advisor

Cathy Wu, Associate Professor, Civil and Environmental Engineering

Direct supervisor

Vindula Jayawardana, Graduate Student, Electrical Engineering and Computer Science
2024

Jordan Tierney

Materials Science and Engineering

Headshot of Jordan Tierney

Research projects

Hydrogen has a wide variety of important industrial applications, from fuel to chemical production, however the current dominant production method results in climate-warming carbon dioxide emissions. We are developing a reactor for methane pyrolysis, a pathway which utilizes high-temperature chemical reactions to turn methane into hydrogen and solid carbon. Solid carbon is a preferable byproduct to carbon dioxide, as it can be collected and sold for use in other applications, rather than being released into the atmosphere.

Advisor

Asegun Henry, Associate Professor, Mechanical Engineering

Direct supervisor

Seiji Engelkemier, Graduate Student, Mechanical Engineering
2024

José Vargas

Physics, Electrical Engineering and Computer Scien

Headshot of José Vargas

Research projects

Energy forecasting is crucial for addressing grid instability, as inaccuracies in predicting demand and generation can lead to power outages or energy overproduction. This results in inefficiencies, higher costs, and increased environmental impact. Inaccurate forecasts can force reliance on backup power sources, often fossil fuel-based, which increases greenhouse gas emissions. Additionally, overproduction may lead to unnecessary energy waste, further straining natural resources. Using Python packages and probabilistic machine learning models, such as AutoTS and Bayesian neural networks, this research improves energy demand and generation predictions. The models address uncertainties like randomness in energy generation and errors in forecasting by validating predictions with real-world data. Accurate forecasting helps integrate renewable energy, reduce fossil fuel reliance, and minimize energy waste. It ensures more reliable and sustainable energy grids, preventing outages and supporting a just energy transition.

Advisor

Anuradha Annaswamy, Senior Research Scientist, Mechanical Engineering

Direct supervisor

Vineet Jagadeesan Nair, PhD Student, Mechanical Engineering
2024

Vandell Vatel

Artificial Intelligence and Decision Making

Headshot of Vandell Vatel

Research projects

New England aims to be net zero in terms of carbon emissions by 2050 to aid in a global effort to curb the effects of global warming. The MIT Energy Initiative’s GenX model is a tool for identifying cost-efficient portfolios of generation capacity. My project focuses on fine tuning GenX so that the model it generates more accurately mirrors New England’s electric grid. The more accurate the model, the more we can trust GenX’s investment decisions, and the more reliable of a resource it becomes.

Advisor

John Parsons, Senior Lecturer, Sloan School of Management; Deputy Director for Research, CEEPR
2024

Annika Vivekananthan

Artificial Intelligence and Decision Making

Headshot of Annika Vivekananthan

Research projects

It’s challenging to efficiently route vehicles through complex, ever-changing city road networks, but applying transformer-based machine learning models to non-Euclidean transportation data can help overcome this difficulty. By refining and adapting neural combinatorial optimization techniques, our work aims to deliver more reliable and cost-effective routing solutions. Ultimately, this research could streamline urban logistics, reduce congestion and emissions, and pave the way for more sustainable, data-driven transportation systems.

Advisor

Cathy Wu, Associate Professor, Civil and Environmental Engineering

Direct supervisor

Yining Ma, Postdoctoral Associate, Laboratory for Information and Decision Systems
2024

Audrey Wei

Artificial Intelligence and Decision Making, Math

Headshot of Audrey Wei

Research projects

The unpredictable rheological properties of soft solids, such as biomass fuel slurries, impede efficient renewable energy conversion and pose challenges for modeling geophysical flows like lava and mudslides. We utilize neural networks to extract and simulate the physical properties of soft solids from video footage, integrating physics-based simulations for real-time adaptability and enhanced predictive accuracy. Leveraging video-based techniques such as 3D reconstruction methods and physics-based models, our approach enables the identification of material properties like viscosity, elasticity, and flow dynamics. These properties are then incorporated into standard constitutive models, ensuring accurate simulations of real-world behaviors under diverse conditions. This research supports scalable renewable energy by enabling optimized biomass energy systems that are both efficient and reliable. Additionally, the insights gained from modeling geophysical flows contribute to improved disaster preparedness, mitigating risks and safeguarding vulnerable communities. By advancing our understanding of material behavior under diverse conditions, this work addresses critical challenges in renewable energy and environmental resilience.

Advisor

Crystal Owens, Postdoctoral Fellow, Computer Science and Artificial Intelligence Lab

Direct supervisor

Wojciech Matusik, Professor, Electrical Engineering and Computer Science
2024

Jansen Wong

Computer Science and Engineering

Headshot of Jansen Wong

Research projects

Mapping data-driven coal retrofitting solution with geospatial synthetic power system datasets for India

Without strategic interventions, operational coal power plants in India hinder progress toward ambitious climate goals, including a national net-zero target by 2070. Leveraging Geographic Information Systems (GIS), development of a comprehensive database, and machine learning, this research project focuses on visualizing effective retrofitting strategies for 284 operational coal power plants. Our interactive web platform and database could provide policy makers and stakeholders involved in India’s energy transition with plant-level strategies as they design retrofitting initiatives.

Advisor

Christopher Knittel, Professor of Applied Economics, MIT Sloan School of Management

Direct supervisor

Yifu Ding, Postdoctoral Associate, MIT Energy Initiative
Mapping data-driven coal retrofitting solution with geospatial synthetic power system datasets for India

India’s 806 coal-fired power units, characterized by aging boilers and low thermal efficiency, impede progress toward the country’s climate goals. This is compounded by incomplete documentation regarding their characteristics, including thermal efficiency. Employing a machine learning methodology, we construct a comprehensive database detailing the operational features of all extant coal plants in India, integrating environmental considerations like water stress and coal price, alongside the development of a visualization tool to enhance accessibility and understanding of our datasets. The ultimate aim is to furnish policy makers with accurate data for computing carbon emissions from coal plants, thereby enabling informed decisions regarding decarbonization strategies.

Advisor

Christopher Knittel, Professor of Applied Economics, MIT Sloan School of Management

Direct supervisor

Yifu Ding, Postdoctoral Associate, MIT Energy Initiative
2024

Katherine Zhou

Mechanical Engineering, Concentration in EECS

Headshot of Katherine Zhou

Research projects

The place of provincial and international trade in electricity to maximize the value of Quebec’s hydro resources

Hydro generation can play a valuable role as a flexible, renewable energy resource. This project uses MITEI’s GenX program to identify the lowest-cost method of allocating investments towards various renewable energy technologies, energy storage, and transmission lines across the northeast region of the US and Canada. My contribution was to dive deeper into mapping and analyzing Ontario’s current hydroelectric resources—generating stations, capacities, and major river sources—so that this more detailed data can be used to garner a more accurate depiction of efficient renewable energy technology implementation.

Advisor

John Parsons, Senior Lecturer, Sloan School of Management; Deputy Director for Research, CEEPR
2024

Aaron Zhu

Business Analytics, EECS

Headshot of Aaron Zhu

Research projects

Perovskite solar cells depend on a wide variety of factors (such as temperature, humidity, chemical concentrations, etc.) that are incredibly difficult to precisely measure and control for. With the creation of autonomous, enclosed testing chambers that allow for quick and accurate collection of environmental data, we are able to derive accurate results and form sound conclusions to understand what factors truly affect perovskite fabrication, ultimately advancing perovskite research at a quicker pace.

Advisor

Tonio Buonassisi, Professor, Mechanical Engineering

Direct supervisor

Alexander Siemenn, Graduate Student, Mechanical Engineering

unfiltered: all results