[michaelhoo]

Hi Everyone,

I am trying to simulate an incompressible fluid flow in a pump turbine.

I wanted to study the effect of micro-compression of water on hydraulic machinery. I added the density change with pressure through the CEL expression, without considering the temperature change, and set it to isothermal.
But the result is obviously wrong. The head is very high and the torque is wrong. The torque on the hub and rim is very large, accounting for about a quarter of the blade torque.The turbulence model is sst.

Below is the cel expression for the water.
LIBRARY:
&replace MATERIAL: water com
Material Description = Water (liquid) COM
Material Group = Water Data,Constant Property Liquids
Object Origin = User
Option = Pure Substance
Thermodynamic State = Liquid
PROPERTIES:
Option = General Material
ABSORPTION COEFFICIENT:
Absorption Coefficient = 1.0 [m^-1]
Option = Value
END
DYNAMIC VISCOSITY:
Dynamic Viscosity = 8.94E-4 [m^2 s^-1] *Density
Option = Value
END
EQUATION OF STATE:
Density = 997 [kg m^-3] + (Absolute Pressure-101325 [kg m^-1 s^-2])/(2.15e9 [m^2 s^-2])
Molar Mass = 18.02 [kg kmol^-1]
Option = Value
END
REFERENCE STATE:
Option = Automatic
END
REFRACTIVE INDEX:
Option = Value
Refractive Index = 1.0 [m m^-1]
END
SCATTERING COEFFICIENT:
Option = Value
Scattering Coefficient = 0.0 [m^-1]
END
SPECIFIC HEAT CAPACITY:
Option = Value
Specific Heat Capacity = 4181.7 [J kg^-1 K^-1]
Specific Heat Type = Constant Pressure
END
TABLE GENERATION:
Maximum Absolute Pressure = 1.0E6 [Pa]
Minimum Absolute Pressure = 1000.0 [Pa]
END
THERMAL CONDUCTIVITY:
Option = Value
Thermal Conductivity = 0.6069 [W m^-1 K^-1]
END
THERMAL EXPANSIVITY:
Option = Value
Thermal Expansivity = 2.57E-04 [K^-1]
END
END
END
END


Would you please let me know why this way of making the fluid slightly compressible causes serious problems?

Many thanks in advance!

[Opaque]

Which equation of state is that one?

What boundary conditions are you using? Once it is not compressible, pressure boundary conditions level matters. That is not the case for incompressible fluid where the "delta" pressure is only thing that matters.

List to check:
1 - Correct way to access pressure for material properties usage?
2 - Domain reference pressure?
3 - Since density is an expression, table generation settings?
4 - Boundary conditions levels with respect to the domain reference pressure

[ghorrocks]

That is a bulk modulus EOS. It uses the bulk modulus of the material to slightly modify the density as a function of pressure. This is approach is typically used in things like water hammer.

Be aware that this is no longer an incompressible simulation. Obviously the density varies, so it is now compressible. This means that you get acoustic effects (pressure waves etc), and in water the speed of sound is around 1500m/s, so this frequently causes a severe restriction on the time step size. You usually need to use far finer time steps then incompressible simulations, and even finer than if you used an ideal gas model for air (the acoustic velocity of air is around 330m/s at atmospheric conditions).

I would strongly recommend you to model this with a constant density fluid first, and get the simulation working accurately with constant density first. Then add the bulk modulus density function - you are probably going to have to make the time steps much finer, and possibly other convergence tricks will be required. But unless the basic simulation works and is accurate with an incompressible fluid you have no hope of getting it to work.

[michaelhoo]

Originally Posted by Opaque

Which equation of state is that one?

What boundary conditions are you using? Once it is not compressible, pressure boundary conditions level matters. That is not the case for incompressible fluid where the "delta" pressure is only thing that matters.

List to check:
1 - Correct way to access pressure for material properties usage?
2 - Domain reference pressure?
3 - Since density is an expression, table generation settings?
4 - Boundary conditions levels with respect to the domain reference pressure

Thank you for your reply!

I am newer to the setting for the CFX software besides the basic setting.

The boundary is shown below:
Inlet: mass flow rate, outlet: static presure( 1atm)
Domain reference pressure

For TABLE GENERATION, only set the max/min pressure:
Maximum Absolute Pressure = 7e+05 [Pa]
Minimum Absolute Pressure = 4000 [Pa]

From my limited experience, I thing static presure ought to 0 atm.Have a try.

[ghorrocks]

As the EOS is linear table generation is normally OK, you just need to make sure you have enough range to cover the highest and lowest pressure you will ever encounter + a tolerance for numerical convergence.

What is the highest and lowest pressure you expect to see in this model?

Is this a steady state or transient model?

And as I said before - the most important thing is that this model runs OK and is accurate with a constant density material model before you apply the variable density model. Have you done this?

[michaelhoo]

Originally Posted by ghorrocks

That is a bulk modulus EOS. It uses the bulk modulus of the material to slightly modify the density as a function of pressure. This is approach is typically used in things like water hammer.

Be aware that this is no longer an incompressible simulation. Obviously the density varies, so it is now compressible. This means that you get acoustic effects (pressure waves etc), and in water the speed of sound is around 1500m/s, so this frequently causes a severe restriction on the time step size. You usually need to use far finer time steps then incompressible simulations, and even finer than if you used an ideal gas model for air (the acoustic velocity of air is around 330m/s at atmospheric conditions).

I would strongly recommend you to model this with a constant density fluid first, and get the simulation working accurately with constant density first. Then add the bulk modulus density function - you are probably going to have to make the time steps much finer, and possibly other convergence tricks will be required. But unless the basic simulation works and is accurate with an incompressible fluid you have no hope of getting it to work.

Yes, you are right, water hammer needs to modify the density as a function of pressure.

The work set as constant density fluid is over, I wana discuss how the fluid slightly compressible will effect on internal flow field and pressure pulsation.

Try finer time steps first.

Thank you for your advice.

[michaelhoo]

It"s teady state.

The result set as constant density material model is ok.

The max/min pressure is set as below and the pressure range is much wider than the real from the simulation of constant density.
Maximum Absolute Pressure = 7e+05 [Pa]
Minimum Absolute Pressure = 4000 [Pa]

[ghorrocks]

On the constant density simulation, what are the maximum and minimum pressures the model predicts?

Please attach an output file from a variable density simulation.