Small, inexpensive flying drones can serve as the soldier’s eyes, ears, and nose in situations that are hazardous or that require 24 x 7 attention, but small flying vehicles face challenges, such as air turbulence and viscosity properties, that larger flyers do not. Bird-sized micro-aerial vehicles (MAV) have roughly one-tenth the aerodynamic efficiency (lift to drag ratio) of larger aircraft. This factor limits the range and endurance of MAVs.

Nature has provided several effective design examples in the form of birds and insects, but imitating these flyers is a complicated task, involving flexible wing structures, complex wing motions, and unsteady viscous flow. Unlike their natural counterparts, drone vehicles cannot spend most of their time in a search for food. Drone vehicle wings must achieve the most lift and propulsion with the least expenditure of energy, so that the drones can carry their load of sensors, communications devices, and fuel.

At present, computing a three-dimensional unsteady viscous flow solution requires on the order of 10 CPU hours, and flapping wing optimizations require thousands to tens of thousands of such solutions. Creating realistic models of deformable wings in periodic motion requires the use of massively parallel computational systems and software that is optimized to make the most efficient use of the available hardware.

AHPCRC researchers are using massively parallel HPC simulations to create pressure distribution models that optimize airfoil shapes for maximizing lift and minimizing drag. Computer simulations have already generated promising airfoil shapes with non-intuitive characteristics such as supercritical wing forms (flattened upper surfaces and highly curved aft surfaces), expanding the realm of possibility beyond the imagination of the engineer. Large-scale three-dimensional turbulent flow simulations are being used to verify the fast two-dimensional shape optimization process for micro aerial vehicle airfoils.

Read more about this project...