temperature gradient

sahoo · June 13, 2018, 14:51

hello everyone,
Actually I am trying to solve the transport equation, as given in the attachment. So, I have made 2 UDS which has also been attached. I just want to know how to include gradient with respect to temperature.

The code for UDS are as follows:-
UDS0 :-

DEFINE_DIFFUSIVITY(teg_diffusivity, cell, thread, i)
{
real res;
real temp = C_T(cell, thread);
res = -3.982 * pow(CON,-5) + 2.98 * pow(CON,-7) * temp - 4.685 * pow(CON,-10) * pow(temp,2) + 2.51 * pow(CON,-12) * pow(temp,3);
real res1 = 1 / res;
return res1;
}

UDS1 :-

DEFINE_SOURCE(teg1_source, c, t, dS, eqn)
{
real temp = C_T(c,t);
dS[eqn] = grad((8.439 * pow(CON,-4) - 3.81 * pow(10,-6) * temp + 6.736 * pow(CON,-9) * pow(temp,-9) - 3.934 * pow(CON,-12) * pow(temp,-12))*grad(temp));
return (8.439 * pow(CON,-4) - 3.81 * pow(10,-6) * temp + 6.736 * pow(CON,-9) * pow(temp,-9) - 3.934 * pow(CON,-12) * pow(temp,-12))*grad(temp);
}

Please help me, I am a beginner and I need it desperately for my project.

`e` · June 13, 2018, 21:35

The temperature gradient is retrieved with C_T_G(c,t) which is a vector with components in each direction, and the magnitude can be calculated with NV_MAG(C_T_G(c,t)).

Further reading: UDF temperature gradient

sahoo · June 14, 2018, 04:13

Thanks 'e' for replying ,
Can you just tell me, if I have a property as a function of temperature(like the one I have attached) and I need gradient of that function then can I simply use grad(function) ?
And also tell if you know the code for finding differential of function with respect to different variables.

`e` · June 14, 2018, 04:59

First of all, determine which equations you would like to solve, and provide some further details. Remember that the nabla symbol means that you are differentiating the variable in each spatial direction, for example in three dimensions:

$\nabla T = \frac{\partial T}{\partial x}\mathbf{i} + \frac{\partial T}{\partial y}\mathbf{j} + \frac{\partial T}{\partial z}\mathbf{k}$

Each term corresponds to C_T_G(c,t)[0], C_T_G(c,t)[1] and C_T_G(c,t)[2]. For details, read section "3.2.3.7. Gradient (G) and Reconstruction Gradient (RG) Vector Macros" (for R15.0) in the UDF manual.

Quote:

Originally Posted by sahoo

And also tell if you know the code for finding differential of function with respect to different variables.

For the example you provided with $\alpha(T)$ , you could differentiate with respect to

analytically. What different variables are you referring to?

sahoo · June 14, 2018, 05:24

Hey thanks again,
Actually I was trying to include the R.H.S of the equation (attached) in source of UDS0 and I know the alpha in the equation as a function of temperature. So, I made the code as follows :-

static real CON = 10;

DEFINE_SOURCE(teg1_source, c, t, dS, eqn)
{
real temp = C_T(c,t);
dS[eqn] = grad((8.439 * pow(CON,-4) - 3.81 * pow(10,-6) * temp + 6.736 * pow(CON,-9) * pow(temp,-9) - 3.934 * pow(CON,-12) * pow(temp,-12)) * C_T_G(c,t));
return (8.439 * pow(CON,-4) - 3.81 * pow(10,-6) * temp + 6.736 * pow(CON,-9) * pow(temp,-9) - 3.934 * pow(CON,-12) * pow(temp,-12)) * C_T_G(c,t);
}

But it wont work. So, I was a bit confused between gradient and differential. But now I understood what you explained. Can you just post the correct code for the above equation, it will be very helpful.

`e` · June 14, 2018, 07:47

If your source term is $\frac{\partial}{\partial x} \left( \alpha(T) \frac{\partial T}{\partial x}\right)$ then you could create a user-defined scalar (UDS) for $\alpha(T) \frac{\partial T}{\partial x}$ using:

Code:

C_UDSI(c,t,0) = alpha_function(C_T(c,t))*C_T_G(c,t)[0];

within a loop, and creating a function for alpha. For example:

Code:

real alpha_function(real T) {
	return 8.439e-4 - 3.81e-6*T + 6.736e-9*pow(T,-9.) - 3.934e-12*pow(T,-12.);
}

where e-4 means *10^-4 or *pow(10,-4). Lastly, the source term can be evaluated with C_UDSI_G. Read through the examples in "8.2.5. User-Defined Scalars" (for R15.0) in the UDF manual for further guidance.

Note: you would need two other UDS for the y and z components (remember to take the gradients in the y and z directions respectively).

sahoo · June 14, 2018, 09:05

Thank you very much I got it.

June 14, 2018, 07:47		#6
`e` Senior Member Join Date: Mar 2015 Posts: 892 Rep Power: 18	If your source term is $\frac{\partial}{\partial x} \left( \alpha(T) \frac{\partial T}{\partial x}\right)$ then you could create a user-defined scalar (UDS) for $\alpha(T) \frac{\partial T}{\partial x}$ using: Code: C_UDSI(c,t,0) = alpha_function(C_T(c,t))C_T_G(c,t)[0]; within a loop, and creating a function for alpha. For example: Code: real alpha_function(real T) { return 8.439e-4 - 3.81e-6T + 6.736e-9pow(T,-9.) - 3.934e-12pow(T,-12.); } where e-4 means 10^-4 or pow(10,-4). Lastly, the source term can be evaluated with C_UDSI_G. Read through the examples in "8.2.5. User-Defined Scalars" (for R15.0) in the UDF manual for further guidance. Note: you would need two other UDS for the y and z components (remember to take the gradients in the y and z directions respectively). ray_lo and FelixS like this.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
How to create a temperature gradient in thin wall and two dimensions	leggman7	Main CFD Forum	0	October 30, 2013 00:21
How to get temperature gradient in dat.file	jason feng	FLUENT	0	September 30, 2013 00:39
Wall correction for the temperature gradient	Hagen	OpenFOAM	0	March 7, 2011 21:04
temperature normal gradient on a boundary	Sandrine	Main CFD Forum	2	June 10, 2009 17:34
Direct calculation of temperature gradient	J.W.Ryu	FLUENT	5	December 27, 2001 07:39

June 13, 2018, 21:35		#2
`e` Senior Member Join Date: Mar 2015 Posts: 892 Rep Power: 18	The temperature gradient is retrieved with C_T_G(c,t) which is a vector with components in each direction, and the magnitude can be calculated with NV_MAG(C_T_G(c,t)). Further reading: UDF temperature gradient

June 14, 2018, 09:05		#7
sahoo New Member Animesh Sahoo Join Date: May 2018 Posts: 20 Rep Power: 8	Thank you very much I got it.