Simulation-Based Evaluation of Predicted Mean Vote (PMV) Distributions Across HVAC Systems and Climate Zones

Abstract

Predicted Mean Vote (PMV) is widely used in building performance simulation to evaluate thermal comfort; however, outcomes can vary by HVAC system type, climate zone, and calculation workflow. This study evaluates PMV outcomes in a simulated medium-sized open-plan office across seven ASHRAE climate zones using two HVAC configurations: Variable Air Volume with Reheat (VAV-Reheat) and Radiant with Dedicated Outdoor Air System (Radiant-DOAS). PMV results were calculated using Ladybug PMV Comfort (LB-PMV) and Honeybee PMV Comfort Map (HB-PMV). Hourly occupied PMV distributions, comfort-hour percentages within the −0.5 to +0.5 PMV range, and LB/HB differences were compared. Results show that HVAC system type and climate zone had a stronger influence on PMV outcomes than differences between LB-PMV and HB-PMV workflows. Radiant-DOAS generally produced narrower PMV distributions and higher comfort-hour percentages than VAV-Reheat. Across all climate zones, HB-PMV frequently exceeded LB-PMV, especially for Radiant-DOAS, where this occurred in approximately 95–100% of occupied hours. This shift is likely related to differences in the underlying calculation procedures, particularly HB-PMV’s sensor-grid-based comfort mapping and Radiance-based longwave and shortwave mean radiant temperature calculations. These findings show that PMV simulation results should be interpreted with attention to HVAC system type, climate zone, and calculation procedure.