AI-Driven Vehicle-to-Building Optimization Strategies for Sustainable Smart Energy Systems

National Taiwan University's push toward a carbon-neutral campus is organized as an eight-investigator core research group — Designing Carbon-Neutral Infrastructures for a Sustainable NTU Campus — led by Prof. Christina W. Tsai and spanning three research thrusts: intelligence-driven infrastructure, extreme-event-resilient infrastructure, and environmentally sustainable infrastructure. Running 2024–2026, it pairs civil-engineering subprojects on water demand, building-energy behavior, rechargeable concrete batteries, green hydropower, and low-carbon mobility. The E3 Center leads Subproject 5, with Prof. I-Yun Lisa Hsieh as PI: an AI-enhanced vehicle-to-building (V2B) smart energy management system for net-zero energy buildings.

The subproject builds on the E3 Center's earlier campus V2B work but pushes it from a rule-based dispatch toward an AI-driven one. Parked electric vehicles, rooftop solar, and building loads all fluctuate, and the payoff from coordinating them depends on anticipating those swings rather than reacting to them. So the work develops machine-learning forecasts for three moving targets — building energy consumption, solar power output, and campus traffic patterns — and feeds them into the V2B optimization. The system can then schedule EV charging and discharging against what demand, generation, and vehicle availability are about to do, aiming for efficient energy distribution among EVs, solar, and the building that holds up under real, uncertain campus conditions.

As one thrust of the wider program, the work plugs into the group's shared net-zero-by-2050 vision: its building-energy and mobility forecasting sits alongside subprojects modeling water–energy–ecology interactions and testing energy-storing construction materials, so that campus-scale demonstrations feed a common picture of how smart, resilient infrastructure gets to net zero. For the E3 Center, it extends the group's nearly-zero-energy-building line — the same V2B and renewable-integration ideas, now sharpened with AI forecasting and scaled toward whole-campus energy management.