Diffusion Constants Changing with Temperature

wadeellis · July 11, 2016, 16:59

Hello,

I am simulating a flow of hot (350 C) helium into room temperature (25 C) air. When defining how the diffusion works, I have not been able to find a way to let the diffusion constant change with temperature. I can define it as a constant, but that is not accurate for the fairly wide temperature range I am using. Does anyone know how to let the constant change?

Thanks in advance.

marmot · July 12, 2016, 03:18

I guess you are using a Multi-Component Gas, for Molecular Diffusivity the default is Schmidt Number but if you change it to say ComponentMolecularDiffusivities and switch from constant to composite than you can create a field function for each component.

The same with BinaryDiffusionCoefficients, selecting composite gives you the fieldfunction choice.

wadeellis · July 12, 2016, 11:19

So if I am understanding correctly, once I have selected field function, I then choose Temperature and it will vary the diffusion coefficient with temperature? How does it know what initial value to give for the coefficient?

marmot · July 12, 2016, 12:03

No you need to create a new user field function, give it a name, myDiffCoef and say something like 0.2*$Temperature, some function of your choice that is dependent on temperature. The initial value will be whatever your function says, so if your initial temperature is 20C than it will use that value for your coefficient

wadeellis · July 19, 2016, 13:56

Thank you for that tip. Now that I have defined a field function (I am using the Fuller Schettler Giddings equation to define the diffusion constant), when I try to run the simulation, it tells me the thermal wall function intersection is not found. I am not sure what that means. If it is helpful, the equation I have put in is 0.001*T^1.75 *(1/Ma+1/Mb)^0.5/(va^(1/3) + vb^(1/3))^2 where M is the molar mass of a given constituent and v is related to the cross section of the species in question.

marmot · July 20, 2016, 02:37

After initializing the solution are you able to view your field function in a scalar scene? Is it defined properly? Perhaps you were just giving the equation and not the actually field function, for example 0.001*pow($Temperature,1.75)...

Does your simulation run when using a constant diffusion coefficient and the problem occurs when using the field function. Check that each one of the variables in your field function is defined and working properly

wadeellis · August 2, 2016, 17:41

It is defined properly at each point, however the temperature of the gas is increasing at some points in a nonphysical way. The helium flowing into the air is at 750 K, but the simulation is saying the temperature gets up to over 5000 K. Does defining the diffusion throw off the calculation of temperature?

Also, if I simply use the Schmidt number for the diffusion, does it account for temperature changes? I might be able to get around this whole thing that way.

marmot · August 3, 2016, 04:09

I think that is a good idea, start with default Schmidt number and see what kind of answers you get, is the temperature physical how does the He spread in the room, based on these results you would move toward your diffusion coefficient.

If your viscosity and density are a function of temperature it does change the Schmidt number,

July 11, 2016, 16:59	Diffusion Constants Changing with Temperature	#1
wadeellis New Member Wade Join Date: Jul 2015 Posts: 24 Rep Power: 11	Hello, I am simulating a flow of hot (350 C) helium into room temperature (25 C) air. When defining how the diffusion works, I have not been able to find a way to let the diffusion constant change with temperature. I can define it as a constant, but that is not accurate for the fairly wide temperature range I am using. Does anyone know how to let the constant change? Thanks in advance.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Static Temperature / Opening Temperature	JulianP	CFX	12	April 10, 2019 19:00
Moving mesh	Niklas Wikstrom (Wikstrom)	OpenFOAM Running, Solving & CFD	122	June 15, 2014 07:20
diffusion effect opposite to convection	hasan karabay	Siemens	0	December 5, 2006 07:59
changing model constants in k-e turbulence model	Sunil	CFX	3	October 3, 2006 13:12
CFX4: temperature is changing	Miroslav Babic	CFX	1	March 2, 2004 19:58

July 12, 2016, 03:18		#2
marmot Senior Member kevin alun Join Date: Sep 2011 Location: Germany Posts: 106 Rep Power: 15	I guess you are using a Multi-Component Gas, for Molecular Diffusivity the default is Schmidt Number but if you change it to say ComponentMolecularDiffusivities and switch from constant to composite than you can create a field function for each component. The same with BinaryDiffusionCoefficients, selecting composite gives you the fieldfunction choice.

July 12, 2016, 11:19		#3
wadeellis New Member Wade Join Date: Jul 2015 Posts: 24 Rep Power: 11	So if I am understanding correctly, once I have selected field function, I then choose Temperature and it will vary the diffusion coefficient with temperature? How does it know what initial value to give for the coefficient?

July 12, 2016, 12:03		#4
marmot Senior Member kevin alun Join Date: Sep 2011 Location: Germany Posts: 106 Rep Power: 15	No you need to create a new user field function, give it a name, myDiffCoef and say something like 0.2*$Temperature, some function of your choice that is dependent on temperature. The initial value will be whatever your function says, so if your initial temperature is 20C than it will use that value for your coefficient

July 19, 2016, 13:56		#5
wadeellis New Member Wade Join Date: Jul 2015 Posts: 24 Rep Power: 11	Thank you for that tip. Now that I have defined a field function (I am using the Fuller Schettler Giddings equation to define the diffusion constant), when I try to run the simulation, it tells me the thermal wall function intersection is not found. I am not sure what that means. If it is helpful, the equation I have put in is 0.001T^1.75 (1/Ma+1/Mb)^0.5/(va^(1/3) + vb^(1/3))^2 where M is the molar mass of a given constituent and v is related to the cross section of the species in question.

July 20, 2016, 02:37		#6
marmot Senior Member kevin alun Join Date: Sep 2011 Location: Germany Posts: 106 Rep Power: 15	After initializing the solution are you able to view your field function in a scalar scene? Is it defined properly? Perhaps you were just giving the equation and not the actually field function, for example 0.001*pow($Temperature,1.75)... Does your simulation run when using a constant diffusion coefficient and the problem occurs when using the field function. Check that each one of the variables in your field function is defined and working properly

August 2, 2016, 17:41		#7
wadeellis New Member Wade Join Date: Jul 2015 Posts: 24 Rep Power: 11	It is defined properly at each point, however the temperature of the gas is increasing at some points in a nonphysical way. The helium flowing into the air is at 750 K, but the simulation is saying the temperature gets up to over 5000 K. Does defining the diffusion throw off the calculation of temperature? Also, if I simply use the Schmidt number for the diffusion, does it account for temperature changes? I might be able to get around this whole thing that way.

August 3, 2016, 04:09		#8
marmot Senior Member kevin alun Join Date: Sep 2011 Location: Germany Posts: 106 Rep Power: 15	I think that is a good idea, start with default Schmidt number and see what kind of answers you get, is the temperature physical how does the He spread in the room, based on these results you would move toward your diffusion coefficient. If your viscosity and density are a function of temperature it does change the Schmidt number,