"Hi, there I'm considering on purchasing a shareware code PHOENICS to get started with it. If anybody has any experience with the shareware version of it, please give me an insight. Is it worth of spending $99 (or $199)? Which level (out of 5) would you recommend? I need to solve an unsteady turbulent swirling flow in a pipe and conical diffuser. Is it good for that? Thank you"

(1). I would say "yes", because you are not buying a code, you are buying the ticket to CFD. (2). The company is one of the oldest commercial CFD company in the whole world, so the access to the experts behind is worth the ticket of a couple of hundred dollars. (3). Last year, I was interested in "ball-point pen". And I tested out various pens such as Montblanc, Waterman, Parker, Cross, etc. These are all in the couple of hundred dollars range. (4). I hope you are not saying that the solution of unsteady turbulent swirling flow in a pipe and conical diffuser worth just a couple of hundred dollars. As long as a code can handle swirl flow through a sudden expansion dump combustor configuration (cold flow), it should be all right for your problem. (5). By the way, it is the ticket to the experts behind the scene. So, make sure that you are buying the access right to them. That is you can talk to them and get questions answered later on.

I used PHOENICS V3.0 about 2 years ago. It is not an easy program to use even with support. I believe the shareware version is V1.5 and does not come with support, or documentation, I think the documentation is extra, and then it is not that helpful. If your geometry is very simple then PHOENICS is probably ok, but I would not like to attempt it without any previous cfd experience. The learning curve is very steep with PHOENICS. I think you can get PHOENICS as freeware, try looking at www.cranfield.ac.uk

From my experience, PHOENICS is not easy to use. I am using PHOENICS v 3.1. As you see, PHOENICS with version older than 3.3 do not support BFC(body fitted coordinate)in the VR environment, it only support the PIL environment. As you are in the PIL case, then it is very tedious work in making grid meshing or etc. It is totally not user friendly. Even if you are not using BFC, you will encouter problem in inserting or creating objects in the cartesian or polar grids. I think the version of PHOENICS you mentioned is only v1.5. For me, even with the documentations, it is hard to use PHOENICS.

Hello, Do you use PHOENICS to do simulation on airfoil?

From your experience, as the solutions are converge but the %error still remain quite high and cannot be reduced anymore, what will you do to cope with this problem? Do you accept the solutions under the above circumstances if nothing can be done?

From my experience, PHOENICS is not easy to use. I am using PHOENICS v 3.1. As you see, PHOENICS with version older than 3.3 do not support BFC(body fitted coordinate)in the VR environment, it only support the PIL environment. As you are in the PIL case, then it is very tedious work in making grid meshing or etc. It is totally not user friendly. Even if you are not using BFC, you will encouter problem in inserting or creating objects in the cartesian or polar grids. I think the version of PHOENICS you mentioned is only v1.5. For me, even with the documentations, it is hard to use PHOENICS.

--------------------------------------------------------------------------------

Yes, you can renormalise the equations based on the inlet fluxes and when the residuals fall to with 1% of these fluxes the one can say the solution is converged.

yes I did use PHOENICS to simulate flow over a turbine blade, I supplemented the code to predict boundary layer transiton

(1). I have used two user-friendly commercial CFD codes, and I was not able to get satisfactory results. (good pictures, yes; qualitatively correct results, maybe. some wrong answers, yes. In the real world situation, you have to be 100% correct and sure in order to win the one million dollars or any dollars at all.) (2). The basic problems are: mesh and boundary layer, numerical methods, turbulence models. (3). To make the CFD codes useful, one must be able to modify it in the above areas(2), and fine tune it with a lot of test data. (4). This is the reason why it is essential to have the ability to write the CFD code and also knows how to interpret the test data. It is possible to modify someone's source code, but it is still very limited and takes much longer to understand the code. (5). People are in business to sell the codes and services, they are not in school doing research and writing papers. (6). Every CFD problem , if it has to be done right, is a PhD dissertation topic. And the reason why people are using CFD is because they have many years of real design experience and test data. In this way, CFD is just another extra method to check the design concepts. (The need to use commercial codes? basically, it is one way to buy the time for engineer, so that better methods and solutions can be found. )