Discussion Forum

What are your boundary conditions ? A typical BC for flow problem is to specify zero pressure at the flow exit which yields stable, consistent solutions. If the units are set up in SI, then the pressure output will be in units of Pascals (Pa) and the pressure is relative to the static pressure at that boundary (zero). If you want to solve the problem in absolute units of pressure, you can try to set the exit pressure to the actual pressure at that point (if known), however, it will be less stable in solving. If you evaluate the fluid properties as a function of pressure, you must be careful to adjust the pressure arguments to the fluid property evaluation to the correct units (especially if solving in terms of relative pressure).

Thanks James for the reply..
I have a constant flow at the outlet equal to 0.0426 m/s and at the inlet the flow varies with time as (9E-12*t^2 - 2E-08*t + 3E-05)*4/.089/22*7/(0.05)^2..
The amount of hydrogen entering the cylinder is more than that exiting.. So the pressure in the cylinder should rise.. I have used SI units.. Can I specify both Pressure and flow rate at inlet and outlet?

Vishal, a mathematically well-posed boundary-condition setup is also what COMSOL is using be default for the flow problem. That is, flow should be specified at the inlet, and pressure at the outlet. Under some conditions, you can specify flow at the outlet instead of inlet, but basically you want to peg the pressure at the outlet.

The problem you are solving sounds quite interesting, but is inviting instability.

I might suggest running the same problem with much simpler setup to start, then gradually add your complications. You will need to understand what level of meshing is required among other things.