PIRE: Building Decarbonization via AI-empowered District Heat Pump Systems

Increasing concerns about climate change, clean energy, and energy security demand that our society transition to a net-zero carbon economy that serves the triple bottom line of planetary health, societal well-being, and economic prosperity. This PIRE project is focused on innovating Artificial Intelligence (AI) techniques with an understanding of human need and behaviors to enable an efficient, human-centered, resilient, and socially justifiable operation of district- and community-scale heat pumps systems that promote a regional scale adoption of building decarbonization. In an increasingly urbanized world, there is a pressing need to address the critical challenges of climate change through the built environment because the building sector accounts for nearly 40% of the primary energy use in the U.S. and associated greenhouse gas and CO2 emissions, with about 50% of that energy dedicated to heating, cooling, ventilation, and lighting. In addition, people spend more than 90% of their time indoors which means that addressing their needs and comfort in a sustainable way is critical for climate resilience planning. Electrification of heating and cooling systems is widely acknowledged as a core and non-negotiable strategy for decarbonization. Many major U.S. metropolitan areas have put the adoption of electric heat pump technologies on the roadmap reaching building decarbonization in the next decade. Taking advantage of the wide adoption of district heating/cooling heat pump systems in Nordic countries, this project seeks to leverage the data and testbeds provided by our core international partners from Sweden and Denmark and the AI innovations provided by the U.S. team to catalyze the readiness to support the scaling up and adoption of human-centered and equity focused AI empowered district system operation strategies at a regional and global scale. Findings from this project will be disseminated through two International Energy Agency’s Annex teams (81 and 84) which will reach to researchers from more than 20 countries. Through their outreach activities, including focus group discussions and workshops, the team will work closely with partners from government and community stakeholders to promote community-focused and equitable district heat pump adoption, implementation, and operation. Two new cross-institutional education programs are designed to promote convergent international education in human centered sustainable built environment: a) Summer International Graduate Bootcamp and Exchange Program; and b) Smart Built Environment Certification Program. Leveraging partner institutions’ other existing flagship programs ranging from K-12, undergraduate, and workforce training, the project will train a diverse, convergent workforce well-versed in science, technology, engineering, arts, and mathematics (STEAM), AI, and socioeconomics to tackle global challenges of climate change. This project contributes to: 1) data science – the development of new human-interactive AI tools to help facility managers and building owners (i.e., users) to make timely decisions and to help incentivize a faster and wider adoption of building decarbonization; 2) building science – the development of novel and scalable models and algorithms for occupant-centric and socially equitable controls driven by occupant, environmental, and community needs; and 3) social science – engaging communities and decision-makers in the design of AI systems to incorporate their environmental, social, economic, and equity needs in the models. Specific anticipated engineering/science contributions include: 1) novel causally-informed Bayesian network with dynamic causal discovery and thermography-based big data analysis for a scalable performance monitoring, energy diagnosis, and prognosis of district heat pump systems, 2) a comprehensive AI learning based occupant needs and behavior modeling framework to bridge the gaps within human-building-community interactions and promote energy equity across the district, 3) new physics- and data-informed learning models for forecasting and optimization under uncertainty, and 4) holistic data sets collected from pilot sites in Sweden/Denmark and the U.S. that use an energy justice lens and are developed through a community-informed approach. The project team's international industry partners, including a heat pump manufacturer and district energy system management companies, will guide the project to streamline the potential technology transfer process. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.