Surrogate models that combine dimensionality reduction and regression techniques are essential to reduce the need for costly high-fidelity computational fluid dynamics data. New approaches using β-variational autoencoder (β-VAE) architectures have shown promise in obtaining high-quality low-dimensional representations of high-dimensional flow data while enabling physical interpretation of their latent spaces. We propose a surrogate model based on latent space regression to predict pressure distributions on a transonic wing given the flight conditions: Mach number and angle of attack. The β-VAE model, enhanced with principal component analysis (PCA), maps high-dimensional data to a low-dimensional latent space, showing a direct correlation with flight conditions. Regularization through β requires careful tuning to improve overall performance, while PCA preprocessing helps to construct an effective latent space, improving autoencoder training and performance. Gaussian process regression is used to predict latent space variables from flight conditions, showing robust behavior independent of β, and the decoder reconstructs the high-dimensional pressure field data. This pipeline provides insight into unexplored flight conditions. Furthermore, a fine-tuning process of the decoder further refines the model, reducing the dependence on β and enhancing accuracy. Structured latent space, robust regression performance, and significant improvements in fine-tuning collectively create a highly accurate and efficient surrogate model. Our methodology demonstrates the effectiveness of β-VAEs for aerodynamic surrogate modeling, offering a rapid, cost-effective, and reliable alternative for aerodynamic data prediction.

This paper shows a series of tools that help in the research of morphing micro air vehicles (MAVs). These tools are aimed at generating parametric CAD models of wings in a few seconds that can be used in aerodynamic studies, either via CFD directly using the model obtained or via wind tunnel through rapid prototyping with 3D printers. It also facilitates the analysis of morphing wings by allowing for the continuous parametric deformation of the airfoils and the wing geometry. In addition, one of the tools greatly simplifies the purely experimental design of this type of vehicle, allowing the transfer of experimental measurements to the computer, generating virtual models with the same deformation as the physical model. This software has two fundamental parts. The first one is the parameterization of the airfoils, for which the CST (Class-Shape Transformation) method will be used. CST coefficients can be modified according to the actuator variable that changes the wing geometry. The second part is the generation of a three-dimensional parametric model of the wing. We used OpenCASCADE technology in its Python version called PythonOCC, which enables the generation of geometries with good surface quality for typical and non-standard wing shapes. Finally, the use of this software for the study of a morphing aircraft will be shown, as well as improvements that could be incorporated in the future to increase its capabilities for the design and analysis of MAVs.

Purpose: This research paper explores the potential application of ChatGPT, a natural language processing-powered language model, in the efficient analysis of aerodynamic data within the context of aeronautics. Recognizing the critical role of aerodynamic data in aircraft performance and design, we investigate a novel approach by integrating ChatGPT into the analysis pipeline. Utilizing a well-known dataset with the aerodynamic features of a NACA0012 airfoil, the study demonstrates the model’s ability to discern complex patterns and articulate nuanced relationships within the data. Beyond showcasing the effectiveness of ChatGPT in aerodynamic data analysis, we delve into the interpretability of its outputs and discuss the model’s potential to automate certain aspects of the analysis, allowing researchers to focus on high-level interpretation and decision-making. The research also addresses challenges and considerations associated with integrating a language model into the aeronautics domain. The findings suggest that ChatGPT holds promise as a valuable tool for aeronautic engineers and researchers, providing a new dimension to aerodynamic data analysis. The synergy between artificial intelligence and aerodynamics presented in this paper opens avenues for future research, with implications for improved design processes, performance optimization, and a deeper understanding of aerodynamic phenomena. Design/methodology/approach: The integration of ChatGPT into aerodynamic data analysis is designed to simplify and expedite various aspects of the process, making it more accessible and efficient for researchers and engineers in the aerospace industry. The approach involves several key steps: data interaction interface, automation of data preprocessing, code generation, exploratory data analysis (EDA), predictive modelling support, interpretability, user-friendly reports, and an iterative feedback loop. These steps allow researchers to interact with aerodynamic datasets using natural language queries and receive responses that facilitate data exploration and analysis. Findings: ChatGPT significantly enhances aerodynamic data analysis by automating complex tasks, improving predictive modelling, and reducing the time and effort required for these processes. It provides an intuitive interface for data interaction, streamlines data preprocessing, and generates code for various analysis tasks. ChatGPT aids in exploratory data analysis by revealing trends, patterns, and anomalies within datasets. Additionally, it supports predictive modelling by assisting in model selection, hyperparameter tuning, and offering explanations for model predictions. Overall, ChatGPT demonstrates its potential to transform aerodynamic research by accelerating analysis procedures and making data analysis tools more accessible to researchers and engineers. Originality/value: The originality of this research lies in its pioneering integration of ChatGPT into the aerodynamic data analysis domain. This study is one of the first to explore and demonstrate the potential of using a natural language processing-based language model to enhance the efficiency and effectiveness of aerodynamic analysis. By automating data preprocessing, code generation, and exploratory data analysis, ChatGPT offers a novel approach that complements traditional methods. The focus on the NACA0012 airfoil configuration provides practical relevance, showcasing how ChatGPT can be applied to real-world aeronautical scenarios, addressing a gap in current research and opening new avenues for future studies.