I want to know the difference between : Panel codes and Potential flow codes and Euler codes and Navier Stockes codes?

And which of them, is the fastest code in order to resolve a external subsonic flow (aircraft) ?

bobby

In increasing order of complexity/cost:

Panel methods: only surface of aircraft is represented by a series of source/doublet panels, and the code solves for the strength of the sources/doublets such that no flow through the aircraft is enforced. The method assumes inviscid, incompressible, irrorational flow. This is the fastest of the methods you asked about as only the surface is modeled and solved. A popular method here is the PMARC code.

Potential flow codes: here the full potential equation is solved for each point in a domain surrounding the aircraft as well as the aircraft surface itself. Only a single scalar for velocity potential is solved. More expensive than panel methods but faster than Euler solvers. I believe the assumption of incompressible, irrotational, inviscid flow still holds.

Euler: full nonlinear inviscid flow equations are solved for mass/momentum/energy conservation. As with potential flow solvers you set up a grid around the aircraft. More expensive than potential flow, less than full Navier-Stokes. No assumption made on irrotational or incompressible flow, but flow is still inviscid.

Navier-Stokes: now viscous diffusion terms are added and usually one also solves a turbulence model, so the number of transport equations (and variables) goes up again. Most expensive option and the most work, as gridding/resolution requirements are greater compared to Euler/potential flow.

Thanks pete for yours responses.

I'm reading Panel methods are very much expensive compared to Navier codes (commercial codes of course), but panel are pretty good in comparaison of experiments results (only if flow angle is less 15° for a external subsonic aircraft!!!) I think VSAero is panel code leader, someone know an other panel code ?

bobby

Bobby,

I'm very surprised you're finding materials stating Navier-Stokes solvers are cheaper than panel methods! I find that completely counterintuitive.

Panel methods are dramatically more efficient because: 1) You only have to grid the surface of the aircraft. With a Navier-Stokes solver you have to generate a 3D grid, with proper boundary layer clustering, around the entire vehicle plus some domain at least 2-3 body lengths upstream/downstream. Lots more work involved up front. 2) You are only solving for the aircraft surface, not the whole 3D field. You will solve for a few thousand panel sources versus hundreds of thousands of cells for a N-S solver. 3) The panel solution is not as intensely iterative. The influence coefficient matrix is formed once and solved, and that's it. There may be a few steps involved to account for wake rollup, but still, panel methods run pretty quick.

I've heard of VSAero but never used it. PMARC is a NASA code. Some panel codes have integral boundary layer routines coupled in which are good up to mild separation, such as the angle of attack you state.

Good luck.