Strategies to characterize thermal exposure to convection- and radiation-based personal comfort systems: a comparative study

• Suitable sensor strategies to assess PCS effects on thermal sensation were identified. • A desk fan and a radiant panel were tested with three measurement approaches. • HVAC manikin captured detailed thermal gradients missed by conventional sensor setups. • Three-level sensor strategy approximated manikin data for vertical gradients. • Findings suggest revising ISO 7726 and ASHRAE 55 for heterogeneous PCS environments. Personal comfort systems (PCSs) show promise in terms of enhancing individual thermal comfort while reducing energy consumption in buildings. However, their inherently heterogeneous thermal environments pose challenges to accurate evaluation using conventional environmental sensing and modelling approaches. This study presents an integrated framework to quantify the impact of PCSs on human thermal sensation by combining environmental measurements with thermophysiological simulation and thermal sensation models. Two representative PCS scenarios, involving a desk fan for cooling and a heating panel for warming, were tested in a climatic chamber. Human thermal exposure was characterised using three input strategies: a multi-sensor, anatomically formed HVAC manikin with 46 distributed sensors; environmental sensors placed at manikin location across three heights; and mean values from three-level sensors close to manikin. These environmental inputs were processed using a validated thermophysiological model to generate physiological responses, which were then used to predict global and local thermal sensations using four global and two local thermal sensation models. The results show HVAC manikin data consistently yielded larger and more realistic differences in thermal sensation between PCS on and off conditions, particularly for localised and asymmetric effects. The strategy of using three-level sensors at manikin location performed reasonably well, especially in approximating vertical gradients, while sensors close to manikin failed to capture critical exposure differences. The sensitivity of thermal sensation predictions was also model-dependent, highlighting inconsistencies between different models. This study provides a practical methodology that can support improved evaluation, design, and standardisation of PCS technologies in spatially heterogeneous indoor environments.