Search for: All records

Abstract The concept of Advanced Air Mobility involves utilizing cutting-edge transportation platforms to transport passengers and cargo efficiently over short distances in urban and suburban areas. However, using simplified atmospheric models for aircraft simulations can prove insufficient for modeling large disturbances impacting low-altitude flight regimes. Due to the complexities of operating in urban environments, realistic wind modeling is necessary to ensure trajectory planning and control design can maintain high levels of safety. In this study, we simulate the dynamic response of a representative advanced air mobility platform operating in wing-borne flight through an urban wind field generated using Large Eddy Simulations (LES) and a wind field created using reduced-order models based on full-order computational solutions. Our findings show that the longitudinal response of the aircraft was not greatly affected by the fidelity of the LES models or if the spatial variation was considered while evaluating the full-order wind model. This is encouraging as it indicates that the full LES generation of the wind field may not be necessary, which decreases the complexity and time needed in this analysis. Differences are present when comparing the lateral response, owing to the differences in the asymmetric loading of the planform in the full and reduced order models. These differences seen in the lateral responses are expected to increase for planforms with smaller wing loadings, which could pose challenges. Additionally, the response of the aircraft to the mean wind field, the temporal average of the full order model, was misrepresentative in the longitudinal response and greatly under-predicted control surface activity, particularly in the lateral response. 

This works aims to generate realistic wind data in urban spaces, which is essential in developing, testing and ensuring the safe operations of Small Unmanned Aerial Systems (sUAS) using Deep Learning (DL). This provides an alternative to existing turbulence models, specifically aimed at urban air spaces. We devise and utilize a Non-Intrusive Reduced Order Model (NIROM) approach to replicate and realistically predict wind fields in urban spaces. The method uses Large Eddy Simulation data from well-established computational fluid dynamics solvers like OpenFOAM to devise the NIROM. High-fidelity data generated from OpenFOAM is decomposed using Proper Orthogonal Decomposition (POD) into its orthogonal modes and basis. These orthogonal modes obtained over time are trained on specialized Recurrent Neural Networks like Long-Term Short Memory (LSTM) to complete the NIROM formulation. This method combined the traditional reduced order modeling approach with deep learning techniques to devise a framework for easy building and application of Machine Learning (ML) based Reduced Order Models (ROMs). A typical urban morphology subject to the wind is chosen and considered a test case for demonstrating the method. 

Results of a previous aerodynamics study conducted over a wing that exhibits the Prandtl Bell Spanload were implemented into a simulation environment with the intent of studying unique flight characteristics that are theorized to be presented by this spanload. However, early simulations over the dynamics show that the yawing moment due to roll rate is of higher effect than the yaw moment due to aileron deflection angle. This over-prediction of the roll-yaw coupling term has been called into question. A new method is to be tested, which implements a compact vortex-lattice (CVLM) formulation, to show the difference between the flight dynamics predicted by this new method and the stability derivative method currently in use. The analysis utilizes two initial conditions to test the differences as the dynamics propagate through time. The first, a large initial bank angle, leads to the stabiltiy derivative method diverging while the CVLM results show this to not be the case. The second condition, a wind-field representative of a stable nocturnal boundary layer over the ground, leads to much more agreement between methods before divergence occurs due to a velocity higher than that of the stability derivative linearization point. It is then agreed that, since CVLM cannot predict stall effects and other nonlinear flight regions, a hybrid approach is proposed that takes advantage of the roll-yaw coupling prediction of the CVLM and the range of condition available to the stability derivative method.