The rapid emerging of urban air mobility (UAM) in recent years necessities real-time and high-fidelity prediction of low-altitude urban wind environments, mainly to support the operational safety regulation and path planning. State-of-the-art approaches including computational fluid dynamics (CFD) cannot meet the requirement of instant prediction. This study proposes a graph neural network (GNN)-based large-time-step surrogate model to provide real-time and high-fidelity flow field prediction to support UAM operation safety. High-precision urban wind environment datasets, generated via CFD, were first validated against wind tunnel experiments and then expanded to train the GNN model. The prediction performance of the surrogate model was comprehensively evaluated in both idealized and real-world urban building settings. The results demonstrate that the proposed model overcomes the constraints of the Courant-Friedrichs-Lewy (CFL) condition, and the time step is 3 orders of magnitude larger than that of CFD. The single-step (0.5 s) inference of flow field consumes millisecond level computational time, achieving 3 to 4 orders of magnitude speedup and faster-than-real-time high-fidelity prediction. However, its accuracy declines as the iterative steps increase due to the accumulation of recursive errors, resulting in insufficient precision for long-term prediction. Spatial analysis indicates that prediction errors are primarily localized in the leeward regions of buildings, manifesting as the underprediction or smoothing of velocity gradients within complex vortex structures. This research demonstrates the potential of proposed model to realize faster-than-real-time urban wind environment prediction to support UAM safety.

Fast flow simulation is imperative in the design of pulsating ventilation, which is potentially efficient in indoor air contaminant removal. The execution of the conventional CFD method requires considerable amount of computational resources. In this study, five different numerical schemes were proposed based on fast fluid dynamics (FFD) and fractional step (FS) methods, and were evaluated to achieve quick simulation of airflow/contaminant dispersion. One of these numerical schemes was identified with the best overall computing efficiency for investigating the performance of pulsating ventilation. With this numerical scheme at hand, the air contaminant removal effectiveness of different ventilation types was evaluated. Two kinds of pulsating ventilation and one kind of steady ventilation were tested upon a benchmark isothermal mixing chamber. The effect of adjusting supply velocity parameters on the ventilation performance was also investigated. CO₂ concentration, airflow pattern, and vortex structure of different ventilation types were illustrated and analyzed. The results reveal that the FS method is more suitable for transient simulation of wall-bounded indoor airflow than the FFD method, and 34%–51% of computing time could be saved compared to the conventional CFD method. Regarding the choice of ventilation type, steady ventilation might result in short-circuit airflow and stagnant zones; alternatively, pulsating ventilation has greater potential in air contaminant removal due to its ever-changing vortex structure.

The air supply terminal located at the floor level attached to side-wall is widely used in large space buildings, leading to potential energy saving as well as significant vertical thermal stratification. The cooling load calculation of such system is challenging, especially the calculation of the load gained from unoccupied zone. This paper adopts experiment and computational fluid dynamics (CFD) methods to study the heat transfer upward and downward across the stratified surface in large space building with floor-level side wall air-supply system. Five experimental cases with different heat source power and exhaust airflow ratios are performed to study their effects on the indoor thermal environment. We investigate the same cases in CFD and verify the result of vertical temperature distribution and cooling load components. As a critical parameter in evaluating the thermal stratification environment of large space building, the inter-zonal heat transfer coefficient C_b is emphatically discussed. By comparing the C_b value obtained through the two methods, the accuracy of the microscopic method is verified by the heat balance method. The results show that the C_b value is mainly affected by the zonal division and air distribution, but less prominently by exhaust airflow ratio and heat source in the occupied zone.

Prompt identification of an indoor air pollutant source location is important for the safety of building residents in gaseous contaminant leakage incidents. Using the computational fluid dynamics (CFD) method in such inverse modeling is time consuming, especially for naturally ventilated residential buildings, which have multiple rooms and require consideration both indoor and outdoor environments. This paper compares the results of the pollutant source location identification and the simulation time based on two different inverse modeling methods: the CFD method and joint modeling of the multizone and CFD methods, to discuss the consumption of the computing time, as well as the accuracy of the location identification result. An instantaneous airborne pollutant source is assumed in a typical residential apartment that utilizes natural ventilation. A CFD model with the computational domain of the whole apartment and surrounding environment is built, for which the adjoint probability method is applied to simulate the source location probability from limited sensor readings. Meanwhile, a multizone model of the apartment is built to simulate and identify the room in which the source is located using the adjoint probability method. The CFD method is applied to the identified room afterwards to identify the exact location of the source within that room. The joint simulation of CFD and the multizone model is verified by a scaled model experiment of the apartment. It is found that the joint simulation method can significantly reduce the computing time and provides a good alternative for real-time inverse tracking of the indoor airborne pollutant.