Poster session 8

In building environment optimization, thermal sensation and comfort evaluations rely on the prediction models. The UC Berkeley comfort model (the CBE comfort model), an advanced multi-nodal prediction model, is developed for the prediction of both local and overall thermal sensations and comforts under non-uniform and transient thermal conditions. In this model, cooling changes in three dominant body parts — back, chest and pelvis — are emphasized as the parts which dominate overall sensation under non-uniform cases. However, this prediction logic can lead to prediction errors in real-world scenarios. To quantitatively assess this issue and propose a feasible correction strategy, we conducted field experiments in semi-outdoor spaces with typical tropical conditions surveying the participants’ thermal sensations. Cooling devices were targeted at the back to lower its sensation, aiming to reach the ‘dominated cold’ threshold. According to the results, the cool sensation of the dominant body part, i.e., the back, can affect overall sensation but does not always dominate it. To improve alignment with actual thermal sensation votes, when the cool sensations of the dominant parts reach the threshold but do not fully dominate overall sensation, a modified method is proposed which treats the whole body as feeling cool, while accounting for opposing warm sensations in other areas. These opposite sensations generate a corrective modifier, shifting the overall sensation towards a warmer side. The proposed revision of the overall sensation calculation in this study can be incorporated into the original CBE comfort model to increase accuracy in predicting thermal sensation.

Evaporative heat loss is crucial to the human body's energy balance. Various thermal comfort models have been established to analyze the heat and mass transfer between the body and the environment during human sweating, yet the effects of sweat-soaked clothing on heat and moisture transfer remain insufficiently evaluated. This study investigates the thermal and moisture properties of different summer clothing and measures the moisture adsorption and desorption curves. The results revealed significant differences in these curves at both low and high moisture content levels. Additionally, the heat loss of the human body in different sweating states was compared using a sweating thermal manikin. It was found that clothing moisture content and clothing coverage significantly influence the human body's heat loss, with overall heat loss increasing alongside moisture content. However, as the sweating rate increased, the evaporative cooling efficiency of the body was suppressed by the soaked clothing. This implies that sweat-soaked garments contribute additional resistance to moisture transfer, adversely affecting skin temperature. Consequently, we empirically assessed and predicted the clothing's thermal and evaporative resistance, along with temperature and vapor pressure at different clothing moisture contents. Furthermore, we observed the dynamic variations in thermal and evaporative resistance of clothing ensembles at different moisture contents, establishing empirical equations for the modification coefficients of thermal and evaporative resistance for sweat-soaked clothing. Finally, we developed a dynamic heat and moisture transfer model that incorporates clothing nodes in the thermal comfort model in detail. This model quantifies the energy exchange between the skin, soaked clothing, and the environment under heavy sweating conditions, highlighting the importance of considering real clothing physical parameters when evaluating thermal comfort in hot and humid weather conditions.

Solar photovoltaic (PV) systems are essential for promoting nearly zero energy buildings (nZEB). Accurately estimating the rooftop solar irradiance simulation is critical for the effective deployment of PV systems in urban environments. However, complex shading from nearby structures and various rooftop obstacles presents significant challenges to this assessment. This research seeks to evaluate rooftop solar irradiance potential in urban areas by analyzing the shading effects of surrounding buildings and obstacles, ultimately providing insights for optimizing PV system installations within the nZEB framework. We proposed an integrated approach that combines deep learning with a novel 3D geographic information system (GIS) to accurately measure solar energy potential at the neighborhood scale, factoring in shading and obstacle effects. By utilizing LiDAR data to assess inter-building shading, we developed a deep learning framework to identify available rooftop areas. Using this hybrid approach, we simulated the available rooftop area for 48 buildings in Montreal, Canada, resulting in a detailed map of PV potential. The estimated annual rooftop solar irradiance potential for the study area is 1.48×10^8 kWh. Ignoring the effects of rooftop obstructions and shading would lead to a 12.87% overestimation of the rooftop PV capacity. Our methodology provides comprehensive simulations of rooftop solar irradiance on an hourly, monthly, and annual basis, offering valuable insights for optimizing solar panel placement and aiding urban energy planning. This work enhances the accuracy of renewable energy advancement and contributes to eco-friendly urban development strategies.