Behind the Climate Action Plan: You Lin, Postdoc at MIT Laboratory for Information and Decision Systems (LIDS)

MIT's operational staff and academic community are key partners to MITOS and in reaching MIT's climate goals. To highlight this, we will be featuring the contributions of operational, academic, and administrative partners.

How did you get involved in the AI building management systems project, and how does it relate to your current work?
I am a postdoctoral researcher at MIT's Laboratory for Information and Decision Systems (LIDS), with a background in machine learning applications for energy systems analysis and optimization. When I joined MIT in September 2022, I was fortunate to become involved in the AI Building Management Systems project. This project aims to reduce greenhouse gas (GHG) emissions and energy consumption in MIT campus buildings through HVAC optimization, particularly temperature setpoint optimization. The project had already achieved significant milestones before my involvement. The previous team successfully implemented setpoint optimization using a physics-based thermal model, model predictive control, and perception models focused on single spaces. To accelerate progress and scale up to cover more spaces, the AI Smart Building team decided to leverage AI to enhance campus energy efficiency.

Building on the team's prior work, my role involves analyzing the current setpoint control sequences and temperatures to identify additional energy-saving opportunities and developing advanced AI models to determine optimal multi-space HVAC system control with limited sensor. I am creating machine learning-based thermal models to predict internal temperatures based on external weather conditions and setpoints, and reinforcement learning models to determine optimal setpoints for each space, considering thermal interactions between different zones. The ultimate goal of the AI setpoint optimization model is to automatically provide customized and optimal setpoints for each individual space on campus, minimizing the energy consumption, GHG emissions, and the need for frequent manual checks.  The AI Smart Building team has also established an automated system that reads data from building management systems, trains AI models, and pushes optimal setpoints to MIT buildings. Using this system, I am currently testing the AI model’s capabilities in the classrooms of MIT's Building 66. Preliminary results have shown a significant reduction in heating and cooling energy consumption. We are now working to expand these tests to more classrooms across other MIT buildings and exploring potential applications in office spaces.

My PIs and colleagues have provided critical support in developing and refining these models and conducting verification tests, offering valuable perspectives on decision-making, building technology, sustainability goals, and facility management.

What is a typical day like for you at MIT?
As a researcher at MIT focused on both academic research and real-world applications, my typical day involves reading papers and reports in my research areas to stay up to date on the latest developments. I also attend seminars at MIT to learn from others about their academic research and how they implement their work to create real-world impact. Meetings with my supervisors and PIs are a regular part of my day, where we discuss my progress and incorporate their feedback and suggestions.
The majority of my time is dedicated to coding, where I translate academic concepts into functional algorithms through data analysis and optimization. I then verify these models using real-world data to ensure their effectiveness and applicability.

What do you do to live more sustainably outside of your work?
Sustainability is important to me not only in my professional work but also in my daily life. I make a conscious effort to avoid using disposable plastic bags, opting instead for reusable paper boxes when grocery shopping, especially at places like Costco. I also separate recyclable and non-recyclable waste in my daily routine and donate unused clothing to charities for recycling. Before leaving home, I usually turn off all lights and the air conditioning system, although I sometimes forget, even with a remote-control app.
In addition to applying smart building concepts in my work, I am also trying to make my home more sustainable. I am currently developing an automation system to manage my smart thermostat and robotic vacuum cleaners, aiming to reduce both energy consumption and costs at home.