Mitigating cold-climate performance decline of air-source heat pumps through data center waste heat recovery: A dynamic techno-economic assessment

Air source heat pumps suffer a coefficient of performance penalty in cold climates due to low evaporator inlet temperatures and outdoor-coil frosting, while data centers reject large amounts of low-grade heat at temperatures suitable for source upgrading. This study proposes and evaluates a building-scale air-side integration in which cooling-loop water from a data center is circulated through a crossflow water-to-air heat exchanger to preheat the ambient air entering the evaporator of a residential air source heat pump, addressing the source-temperature deficit at its origin without an indoor hydronic retrofit. A dynamic TRNSYS co-simulation framework couples the data center cooling system, the heat exchanger sized through a coupled MATLAB–TRNSYS optimization, and a four-storey multi-unit residential building. The integrated system is evaluated in four Canadian cities (Vancouver, Toronto, Montreal, Edmonton) using provincial grid carbon factors and commercial tariffs. Heat recovery raises the seasonal heat pump COP by 8.5% in Vancouver and up to 55% in Edmonton, and all four cities reach a similar COP of about 3.0 regardless of climate. Combined building and data center electricity savings range from 1253 to 1357 MWh per year, corresponding to operational CO₂ reductions of 1.6 to 638 t CO₂eq per year. The heat exchanger also maintains the indoor comfort range that the baseline heat pump cannot sustain, acting as a virtual capacity expansion beyond the standard 99th-percentile sizing. A 20-year discounted cash flow analysis gives a positive net present value in every city, from CAD 1.06 million in Montreal to CAD 2.27 million in Toronto, with internal rates of return between 27% and 49% and discounted payback periods of 2.3 to 4.4 years. A sensitivity analysis on six economic and environmental assumptions confirms that the positive economic outcome is robust across all four cities.