Enhanced Computational Performance and Stability & Control Prediction for NATO Military Vehicles

Abstract

Numerical simulations are a principal tool across the lifecycle of NATO military air and sea vehicles. Traditional low-fidelity methods such as potential flow or Euler codes are often used in the early stages of the design cycle due to their lower computational cost and time requirements. However, these methods may lack the required accuracy for certain design scenarios, and the use of higher-fidelity methods such as computational fluid dynamics and experiments is often limited by the tremendous cost and timescales necessary for providing comprehensive data sets. A feasible means to overcome these limitations is reduced-order modeling. Over a four-year period the NATO Science & Technology Organization Research Task Group 351 of the Applied Vehicle Technology panel developed various data-driven reduced-order models. Those were applied on four different air and sea vehicle use cases, which demonstrated nearly real-time data set population and maneuver or course-keeping prediction capabilities. A summary of the Research Task Group is presented in terms of the air and sea use cases, methods, and results. Additional work is required to understand the real extent of cost reduction achievable with these approaches. Open questions remain on combining data sets with different fidelity levels or quantifying uncertainty in models.