Relatively few “dominant” zones drive energy savings in setpoint adjustment experiments: Implications for energy efficiency

• Conducted cooling setpoint experiments in 10 buildings and 88,000 m 2 of floorspace. • Observed significant savings driven by relatively few (14%) responding zones. • Introduced new class of zones, dominant zones (16%), to explain zonal response. • Dominant zone phenomenon driven by zonal heterogeneity and overcooling. • Retrofits can be prioritized in dominant zones to streamline efficiency programs. We report on cooling setpoint adjustment experiments conducted in the summer of 2022 in seven office buildings and three lab buildings on Stanford University’s campus, spanning ∼ 88,000 m 2 and ∼ 1,600 zones in total. We find significant energy savings, up to 52 % depending on the building, but driven by relatively few responding zones: 14 % of zones across the testbed. To explain this behavior, we introduce a new class of zones we call “dominant zones”, defined via the Lorenz curve as the smallest set of zones responsible for 85 % of cooling requests. A total of 16 % of zones are flagged as dominant, with substantial overlap to zones that responded to CSP adjustments, providing experimental support for the definition. We explain why few zones responded to setpoint changes and why so few zones are dominant, linking the phenomenon to 1) “zonal heterogeneity” – the idea that zones have innately different cooling needs, and 2) “overcooling” – an inefficiency endemic to U.S. buildings where zones are cooled well below setpoint. In light of observed zonal heterogeneity and overcooling, we highlight solutions to reduce consumption in high-demanding “dominant” zones and overcooling in low-demanding “dominated” zones. Ultimately, our results suggest that solutions can be prioritized in relatively few zones to attain significant energy savings with a fraction of the time and cost.