Engineers frequently approach Tyto Robotics regarding computational fluid dynamics (CFD) versus physical testing. Many organizations contemplate eliminating experimental validation, questioning whether digital simulations alone suffice for propeller and motor development.
Tyto Robotics maintains that laboratory testing remains essential despite technological advances in modeling capabilities. While their team extensively utilizes CFD analysis, these digital tools primarily serve to forecast potential performance characteristics rather than deliver definitive data.
Computational modeling proves particularly valuable when evaluating modifications to established systems where baseline parameters have been thoroughly documented. However, the accuracy of simulated results depends heavily on multiple technical factors: turbulence model selection, mesh architecture, boundary specifications, and numerical processing methods.
This analysis examines the correlation between predicted and measured performance metrics for a 60-inch unmanned aerial vehicle propeller, demonstrating key disparities between virtual and physical testing environments.
Summary of the Report
Tyto Robotics’ report summarizes tests and simulations the company completed to determine the relationship between numerical simulation results and experimental data for a 60″ UAV propeller.
The Ansys Fluent software was used to generate a CFD model of our 60” propeller, which was then experimentally tested with the Flight Stand 150 thrust stand.
What Tyto Robotics discovered was that during tests from 1000 – 2000 RPM, the CFD simulation underpredicted thrust by 69%.
Procedure
Simulations and experimental tests were run in parallel over the course of several weeks.
The Ansys simulation was designed to mimic the experimental lab space, including the proximity of walls, obstacles and the overall size of the room. The thrust stand and propulsion system were also modeled to the exact dimensions of their live counterparts.
Each Ansys simulation took an average of 10 days to complete using a computer with a latest generation GPU. Occasionally, the simulation would fail to converge and the parameters would have to be adjusted before restarting the simulation. Overall, the dozen simulations Tyto Robotics completed took about 20 weeks to run. Preceding the simulations, its team completed 16 hours of training with a consultant.
In parallel, experimental tests were run using the Flight Stand 150 thrust stand. The company used a 60 kW power supply to power the stand, mounted with a TA150 motor and 60” propeller.
Click this link to read the full article and see the specifications and results.
