If computational fluid dynamics (CFD) tools are so expensive, so computationally-hungry, and so darn slow to run, why not just stick with the other types of tools we’ve covered?
I like to think of these 3D analyses with viscous effects as a detailing step. You’d likely use the other tools for the initial design stages of your aircraft, but you reach a point where you start to need more precision in your calculations.
An example is when you start to need more precise drag estimates. You can do good buildups of skin friction, base, and other types of drag from textbook methods. But those are all estimates based on approximations of fuselage shape and other factors.
If you want to get a drag number that actually takes into account your UAV’s unique fuselage shape and the various pucks and other bump-outs for payloads, you’ll need to use CFD. Panel codes can’t support the level of geometry detail you would need to represent this. But CFD tools can.
When the problem you’re solving involves anything to do with air pressure, CFD is also probably the best choice. In the past I’ve needed to derive pressure corrections for an air data system vastly different from the one that came standard on the airplane. By running a model of the payload this system was integrated with, I could see how the air flowed around it, and measure the air’s physical properties right where the pressure ports would be on the real thing.
Comparing these measurements to the “true” air pressure that I had set in the simulation gave me a coefficient to put into the UAV’s autopilot to correct the pressure readings. Because the autopilot uses air pressure to calculate airspeed, this made sure that the airplane was using the correct pressure measurement for that calculation, and therefore would always report an accurate airspeed.
I have way more examples I could talk about than just these couple. Really, because of how it simulates the true viscosity of air, you can think of CFD as a way to answer all the more complex questions that have popped up during development.
How do the hinge moments for our elevators change if the flow separates around them?
Does our propeller being mounted just above the wing impact its “normal” airflow?
Which part of our aircraft’s wing stalls first, the inboard or outboard? Can we change that?
Being able to set up a good CFD simulation is certainly a skill unto itself; I won’t overwhelm you with all the myriad settings you can use to best replicate airflow. And it takes a lot longer to get really good at it than it does for an inviscid tool.
But once you have a solid level of competence and familiarity with your chosen software, there is a heck of a lot you can explore with that modeling power.