This conference paper was submitted for presentation at the NAFEMS World Congress 2025, held in Salzburg, Austria from May 19–22, 2025.

Abstract

Timetooth Technologies has developed landing gears for drones and UAVs, having optimum energy absorption capabilities. In conventional methods, physical tests are conducted to validate the performance, and then the test results are used to fine-tune the design. However, such tests are time-consuming and costly. The current study focuses on the design of landing gear by leveraging advanced Computer Aided Engineering (CAE) simulations, and predictive modeling approaches to accurately capture the entire mechanics involving structure and fluids. A judicious combination of CAE tools along with application of machine learning algorithms, the intricate landing gear system was designed efficiently and effectively, in a simplified manner. The efficient dissipation and absorption of energy during landing rely primarily on two key properties: stiffness and damping. Stiffness property is achieved by pre-charged gas while damping phenomena is achieved by restricted flow of fluid across orifice during stroke. The multibody dynamics model incorporated essential landing parameters such as the vehicle's weight, vertical sink rate, horizontal landing speed, and allowable ground reaction factor, while capturing the structural deformation of critical components. To optimize the performance of the landing gear, variables such as gas stiffness, orifice radius, and tire stiffness were considered and treated as adjustable parameters. For modeling damping characteristics, a novel approach that combines Computational Fluid Dynamics (CFD) simulations with machine learning techniques is introduced. Approximately 500 CFD simulations were performed for various orifice radii to obtain the corresponding force-velocity characteristics, which were then used to train a machine learning model. The trained and validated model was employed to predict the orifice radius corresponding to a desired set of force-velocity characteristics. This method provides a more efficient way to design damping systems, significantly reducing the need for multiple iterations of CFD simulations. Dynamic energy absorption characteristics of the landing gear were validated through the physical drop tests. The correlation between the design parameters and test data was found to be remarkably good, indicating the accuracy and reliability of the ML augmented CAE simulations. This high level of correlation demonstrated the effectiveness and efficiency of the design process that captures multiple physical phenomena. This approach demonstrated several advantages, including convenience, cost-effectiveness, and time efficiency. The findings of this study make valuable contributions to the advancement of landing gear design and the augmentation of machine learning with CAE simulations in the field of aerospace engineering.