Hello, In the velocity inlet conditions, we can define turbulence specifications as: (1) Intensity & Hydraulic Diameter (2)Intensity and length scale, (3) Intensity & Viscosity Ratio, (4) Intensity & Hydraulic Diameter. Which method am I to choose? Of course I know only the inlet velocity and diameter of inlet pipe.

What about the turbulence intensity? How can I determine an accurate value for the turbulence intensity before I can proceed with the solution?

If I choose method (4), I know the hydraulic diameter, which is the actual diameter for a pipe inlet, but I am confused about the turbulence intensity. Can I assign any arbitrary value of 5%? What if I assign 1% or 10%? I fail in my logic.

(1). You have to know what you are eating, and the inlet condition is the place where the flow enter the system. (2). You have to tell the system the inlet condition(period)you are going to use. You can not say that you don't have the inlet condition. (3). This is because in most cases one is trying to simulate the real flow problem. (there exists the test data) (4). To run a calculation, you can assume a condition at the inlet, say 1%, 3%, 5%, 8%, 15%,etc. If you do it this way, you will never know the real inlet condition. As a parametric study, it is all right. To simulate a real flow problem, you will need the test data to guide you. (garbage-in-garbage-out, some problems are very sensitive to the inlet non-uniform conditions.)

That means, if I know the velocity at the inlet, (1) It might be in the transition regime, (2) It might be in the turbulent regime, I cannot interprete these turbulence parameters theoretically (with some correlation of course), until and unless I have pure experimental data. (Theoertically, these parameters have been obtained from the transient terms in the velocity field, and these transients are pure experimental outputs).

So if I am confirmed that the flow is in the laminar regime (e.g. Re=10), I will switch on the laminar model. But what if I am not sure about the regime (at a higher Re, for example). If I know that a particular Re corresponds to the transition zone, and I turn on the laminar model, definitely it will have some error. I will switch on some turbulence model, but still then I will not be able to feed these turbulence parameters. The only way out for such simulations will be experimental determination.

Is it right what I think?

(1). A flow problem must be defined first so that solution in the interior computational domain can be obtained through cfd analysis. (2). The boundary conditions such as the inlet, the exit, the walls conditions are places where one can specify conditions to define the flow problem. That is the solution will depend on the boundary conditions one specified. (3). If some of the input parameters are not specified in the boundary conditions, it is not possible to solve the problem. (4). If one is solving for the laminar flow, he does not need the k-epsilon equations, and so he does not need to provide the boundary conditions for the k and epsilon. (5). For turbulent flow, if the k-epsilon model is used, then the boundary conditions for k and epsilon must be provided. In this case, k and epsilon at the inlet must be given. (6). In many cases, one can assume a % value based on the inlet mean velocity to calculate the k, and a % value based on the inlet geometry or diameter to estimate the length scale. These estimate will be good enough to specify the inlet boundary condition in many cases.(depends on the problem you are trying to solve.) (7). If the inlet location has highly non-uniform flow, it is a good idea to relocate the inlet to an upstream location to avoid specifying an unknown inlet condition. In some cases, one needs to extend the inlet further upstream to reach the place where one can easily specify the inlet condition. (8). You can either use the measured and derived values or use the assumed parametric values at the inlet. In either case, the final results computed must be validated against the know results or data.