From high fidelity CFD to Reduced Order Modeling for launcher aerodynamics using nDAI
The evaluation of aerodynamic loads is a key aspect in the design and optimization of space launch vehicles. In early design phases, multiple geometrical configurations must be assessed to understand aerodynamic interactions and support design trade-offs. However, conventional approaches based on high-fidelity Computational Fluid Dynamics (CFD) simulations are often time-consuming and computationally expensive, making large parametric studies difficult to perform.
This study presents the development of a Reduced Order Model (ROM) to rapidly estimate the aerodynamic forces acting on space systems. The objective is to replace repeated high-fidelity simulations with a fast predictive model capable of capturing the main aerodynamic trends across a defined design space.
The entire workflow was implemented using nDAI and VOLTA, both developed by ESTECO. The platform was used to automate the execution of CFD simulations, orchestrate the parametric study by distributing the execution across different computing environments, manage the generated dataset, and train surrogate models. The CFD analyses were performed using CT Ingenierie in-house tool called CPS_C, specifically developed and optimized for space applications. Through its design of experiments (DoE), data analytics, and model training capabilities the platform enabled the efficient construction and validation of the reduced order model while significantly simplifying the management of the simulation campaign. Furthermore, VOLTA's democratization capabilities allow the obtained surrogate model to be shared seamlessly with interested stakeholders for instantaneous performance predictions.
The methodology is illustrated through a case study focusing on the aerodynamic interactions between launch vehicle boosters. The ROM predicts the evolution of aerodynamic loads as a function of the separation distance between boosters, allowing rapid evaluation of different configurations. The results demonstrate that the surrogate model accurately captures the main aerodynamic behaviors while drastically reducing computational time compared to traditional CFD-based approaches.
This work highlights how an AI engineering platform such as nDAI in collaboration with VOLTA can significantly streamline the development of reduced order models and support faster engineering decision-making during early-stage launch vehicle design.
