Unit of Density in Navier Stokes Equation

Hi. I am a CFD beginner and I need help from experts. I am trying to simulate a simple 2D water flow in a straight pipe. This may be a dumb question, but I am having a hard time to understand what the right unit of density is in the Navier Stokes equation in FEA Multiphysics. Here is my problem. I created a geometry of the pipe using the metric system. For example, 0.05 meter diameter and 10 meter length. Water density is 1000 kg/m3 and viscosity is 0.001 in Pa•s. The inlet velocity is 1 m/s. They are in SI units. When I run the simulation using these inputs, the model does not provide the right answer. When I set the density value to 1, it seems to provide the right answer. But I don’t know if I should use the different unit for only density while other parameters are in SI units. Thanks in advance.

It is quite hard to say what is wrong without more information. What is the answer you get, what is the answer you are expecting, and why?

Here is what I did.

1. Open FEATool Multiphysics and select 2D Navier Stokes
2. Go to Geometry and create 0.1 x 0.5 (I tried to use SI units. So these are 0.1 m diamter and 0.5 m length.
3. Create 0.001 Meshes both for Subdomain and Boundary Grids
4. Go to Equation and set density to 1,000 and mu to 9.47E-05 Pa s. These are Si units.
5. Go to Boundary and set inlet velocity, u0, to 2 for the left side (Domain 4), set Outflow/pressure and p0=0 for the right side (Domain 2), and set Wall/no-slip for the top and the bottom (Regions 1 and 3).  
6. Go to Solve and run by hitting “=”.

Result 1.png is what I got. The velocity is obviously too high and it is a wrong result. Result 2.png is the result I got when I ran analysis with density of 1, not 1,000. Result 2 is the result what I am looking for.  The difference between result 1 and result 2 is created by the difference in densities. When I use 1,000 for density to meet the SI unit system, the answer is obvisouly wrong. However, when I use 1 for density the result seems right to me. So I suspect there is something wrong the units for density I used.

Result1.png
Result2.png

Ah, although you didn't include the solver convergence history in "Result1.png" it is typical of when the solver has failed to converge. Mostly likely your flow is in a turbulent regime, and hence you will need to use a turbulence model to solve it. Please have a look at the tutorials/examples with turbulent flows.

Thank you so much. I understand. I have another question that I will ask in a new thread.