Hi, folks: I had some limited experience in the numerical analysis of some engineering problems. now i am gonna to deal with a problem concerning of the mechanical artificial heart valve cavitation. my question is: do you experts think it is better for me to employ a commercial code (if so, any recommendation is there) or do it by myself from the all beginning. thanks for your advices in advance. Rick

(1). I would say that you need to understand the fluid dynamic aspect of the mechanical artificial heart valve cavitation problem first. (whether it is flow separation or flow cavitation) (2). To use a commercial code, it is going to take you a while to learn how to use the code properly, such as geometry and mesh generation. If you are talking about a moving heart valve, then the problem becomes a transient problem. The mesh has to be a function of time also. Once this is done, the control of the solver normally is rather straightforward. But this does not mean that you will obtain the result right away. It is just the begining. (3). To do it on your own is actually more interesting. There you can dig up some related papers and figure out how to write a CFD program. It is not really difficult at all once you understand how to do it. In this approach, you can divide the effort into several phases, such as simple 2-D problem to figure out the programming strategy, simple 2-D transient problem to animate the mesh and the solution, a more realistic 3-D model, and finally a full 3-D program. In this way, you will be able to find out what you didn't know. That is the first step to learn. (4). At some point in the process-3, you may want to explore the possibility of using the commercial code when you are not sure whether you can handle the problem or not. (5). Once you have learned how to write a CFD program, it does not matter whether you use a commercial code or your own code to solve your heart valve problem. (6). If you are mainly interested in the final solutions, then getting help from a professional consulting firm is also a good alternative.

John, thank you for your suggestions. Accoding to your experience, how much time I have to spend on a 2-d mechanical heart valve CFD ( I think you know this topic) on my own. actually I have finished the mesh generation, discretization. but I am afraid I will stick it too much time.

In terms of computing time, a 2-D laminar flow problem will take CPU time of order 5 minuites for incompressible flow case. A 2-D turbulent flow will take 10 minuites. A 3-D laminar incompressible flow will take 2 hours of CPU time and a 3-D turbulent flow will take 3 hours of CPU time. For the transient flows, it will depend on the number of time steps used in the computation. That takes care of the CPU time. ( the CPU time for compressible flow is similar to that of the transient flow computation because most compressible flow formulation is based on the transient flow formulation. Sometimes, it takes twenty thousands time step to converge. In this case the CPU time will be several days to weeks.) As you see that, the CPU time required to give a 3-D steady,incompressible, turbulent flow solution is less than one day. So, you can spend the rest of the time to study and develop your program. To write a 2-D code for incompressible flow, it will take, say a couple of weeks. ( one day for equations, two days for writing the code, three days for debuging the code, that's about a couple of weeks) For a 3-D incompressible flow code, it will take, say a couple of months. ( but remember that a couple of months can easily become six months or a year.) For a PHD student, a couple of years to write a 3-D code is practical.( not including the time to learn the method, etc.) By the way, running a commercial code without proper training and certification can be very dangerous to your career. ( not mentioning the patient's heart ) Someone has to spend time so that life ( irreversible time ) can be saved. You see that is " the law of conservation of time ".But that is outside the scope of the CFD.