Modelling spatial variations in thermal comfort in indoor open-plan spaces using a whole-building simulation tool

Abstract

The building space air temperature and mean radiant temperature are known to vary spatially. Lumped models, which are the basis for various building energy simulation tools, treat the thermal properties of the air in the entire space to be uniform. Using such tools, modelling spatial non-uniformity in the temperatures is a challenge. In the present work, a novel method is proposed that customises a building energy simulation tool for such applications. The proposed method divides the room into several lumped nodes and models the cross mixing between the nodes. The spatial variation in mean radiant temperature is estimated using a combination of analytical and numerical methods. The proposed approach is validated using real-world controlled experiments. Further, the paper demonstrates the usefulness of modelling the spatial variations in a room with multiple air-conditioning units. It is observed that representing the space as a single lumped node is inadequate for predicting spatial variations in both the air temperature and the glazing surface temperature. The proposed approach effectively models the spatial variations in the air, mean radiant, and operative temperatures. The proposed method outperforms the baselines that model air temperature as a single node and compares well with the measurements with a maximum error of approximately 1.3%. Further, it is shown that the proposed approach helps in achieving a better uniformity in operative temperature in a test-bed that is served by two AC units.