Center for Environmental Building & Design

  • 2021 Best Paper Award

    Building and Environment best paper award 2021. Selected 3 articles our of 4,500.

  • Graphs of thermal preferences

    Distribution of three thermal preference classes over the selected variables.

  • Adaptive behavior and different thermal experiences of real people

  • Sankey plot of thermal preferences

    A Sankey plot to depict the number of correct and incorrect predictions
    classified by the BNN model.

Adaptive behavior and different thermal experiences of real people

2021

A Bayesian neural network approach to thermal preference prediction and classification

Winner of the Building and Environment 2021 Best Paper Award

Abstract: Various observed and unquantifiable factors affect the thermal comfort of occupants in indoor environments and can lead to high uncertainty in the prediction and classification of their thermal preferences. The behavioral adaptation of occupants, by operating window systems for example, changes their thermal experience and expectations and therefore contributes to even higher prediction uncertainty. In this study, we applied a Bayesian neural network (BNN) algorithm to build a predictive model for occupant thermal preference using the ASHRAE Global Thermal Comfort Database II. The Bayesian method allows us to synthesize prior knowledge and available measurements into a unified modeling framework. It also offers a way to express and quantify uncertainty. Here we have performed a systematic study to test the efficiency and robustness of different BNN model configurations. The results show that the BNN model outperforms conventional thermal comfort models such as Predicted Mean Vote (PMV) and adaptive comfort model. The BNN model tends to produce more confident “prefer cooler” predictions with high possibility and low uncertainty. In contrast, the BNN model produces less certain predictions for “prefer no change” and “prefer warmer” across all occupants. Our findings suggest that  linking occupants’ subjective evaluation measures and window opening/closing behavior to thermal comfort modeling effectively improves predictive performance.

For more information