This example can help.

#include "udf.h"

DEFINE_PROFILE(inlet_x_velocity, thread, position)

{

real x[ND_ND]; /* this will hold the position vector */

real y;

face_t f;

begin_f_loop(f, thread)

{

F_CENTROID(x,f,thread);

y = x[1];

F_PROFILE(f, thread, position) = 20. - y*y/(.0745*.0745)*20.;

}

end_f_loop(f, thread)

}

For 2D case, the X[ND_ND] has X[0] for axis x and X[1] for axis y.

For 3D case, the X[ND_ND] has X[0] for axis x, X[1] for axis y and X[2] for axis z.

You should just substitute the F_PROFILE definition. Use the current time 

real t = CURRENT_TIME;

to set up the time dependency.

http://www.arc.vt.edu/ansys_help/flu_udf/flu_udf_udf_bc_appl.html

Hope it helps.

Franco

Thanks Franco for your help.  I have a time dependent triphasic pulsatile velocity profile in blood flow case.  The inlet face is perpendicular to the z-axis so the z-velocity will depend on x- and y-coordinates and the parabolic profile will be in 3D.  I plan to run the simulation with the udf c-code below.  Any suggestion such as using any of the CFD-Post features for me to check whether if it works?  

#include "udf.h"//file that contains definitions for define functions and fluent operations

#define PI 3.141592654

DEFINE_PROFILE(inlet_velocity,th,i)

{

        float x[3]; /* an array for the coordinates */

        float xx;

 float y;

 face_t f; /* f is a face thread index */

 begin_f_loop(f,th)

  double t = (CURRENT_TIME*2-floor(CURRENT_TIME*2))/2; //t is the local time within each period

                F_CENTROID(x,f,th);

  xx = x[0];

                y = x[1];

 {

  if(t > 0.1 && t 0.175 && t 0.425 && t

I think I have resolved my problem with the udf c-code as outlined below.  Any one with similar problem, please feel free to use this code:

#include "udf.h"//file that contains definitions for define functions and fluent operations

#define PI 3.141592654

DEFINE_PROFILE(inlet_z_velocity,th,i)

{

        float x[ND_ND]; /* an array for the coordinates */

        float xx, y, d;

 face_t f; /* f is a face thread index */

        d = 0.02; /* inlet diameter in m */

 begin_f_loop(f,th)

        {

  double t = (CURRENT_TIME*2-floor(CURRENT_TIME*2))/2; //t is the local time within each period

                F_CENTROID(x,f,th);

  xx = x[0]; /* x coordinate */

                y = x[1];   /* y coordinate */

  if(t > 0.1 && t 0.175 && t 0.425 && t 0.5 && t

:-)

Hi . . . I corrected two minor mistakes.  Periodicity is now 0.8 second previously 0.5 second.  Parabolic profile should have a numerator as the sum of square of x and y-coordinates while denominator is the square of radius of blood vessel.  

Final version is below and I have checked that it is running as per the udf, feel free to try it out.

#include "udf.h"//file that contains definitions for define functions and fluent operations

#define PI 3.141592654

DEFINE_PROFILE(inlet_z_velocity,th,i)

{

        float x[ND_ND]; /* an array for the coordinates */

        float xx, y, r;

 face_t f; /* f is a face thread index */

        r = 0.01; /* inlet radius in m */

 begin_f_loop(f,th)

        {

  double t = (CURRENT_TIME*1.25-floor(CURRENT_TIME*1.25))/1.25; //t is the local time within each period

                F_CENTROID(x,f,th);

  xx = x[0]; /* x coordinate */

                y = x[1];   /* y coordinate */

  if(t >= 0 && t 0.1 && t 0.175 && t 0.425 && t 0.5 && t

Hi,

Can any body please give me the udf for pulsating air jet in which velocity is a sinusoidal function of time

Hi Siddiqui, What is the amplitude and period of your sinusoidal function?  The udf below is 0.8 second period and amplitude of 10.  Try it and you can adjust according to your situations and see if it works for your case.

#include "udf.h"//file that contains definitions for define functions and fluent operations

#define PI 3.141592654

DEFINE_PROFILE(inlet_velocity,th,i)

{

 face_t f;

 begin_f_loop(f,th)

 }

    F_PROFILE(f,th,i) = 10*sin(2*PI/0.8*current_time);

 }

 end_f_loop(f,th);

}

Hi sir,

Thank you for reply 

Sir my velocity profile is as attached in the file

Thank you once again

Sir,

Above all while defining udf and selecting the transient state(unsteady) i am not getting the effect of pulsating impingement of air on the hot flat plate.  the temperature of plate is remains constant.

Sir,

pulsating velocity programme is correct it is imported succesfully

can u pls tell me where i may be wrong in fluent

as far as amplitude and time is concerened i have ajusted that

Oh, I see, it is a square wave periodic not sinusoidal. Did you select the udf inlet_velocity as inlet velocity of your boundary condition in the solution set-up?  

Please try this and see how it goes:

#include "udf.h"//file that contains definitions for define functions and fluent operations#

define PI 3.141592654

DEFINE_PROFILE(inlet_velocity,th,i)

{ 

face_t f;

        begin_f_loop(f,th)

        {  

        double t = (CURRENT_TIME*0.01-floor(CURRENT_TIME*0.01))/0.01; //t is the local time within each period

        if(t >= 0 && t < 0.0062)

            F_PROFILE(f,th,i) = 20;

        else

            F_PROFILE(f,th,i) = 0; 

       end_f_loop(f,th)

       }

}

Sorry it is the other way around,

 if(t >= 0 && t < 0.0062)

            F_PROFILE(f,th,i) = 0;       

               else

            F_PROFILE(f,th,i) = 20; 

Hi Yoon - Song Sir,

This works!!!!!!!. but this is valid only for one time step 

Sir i want to run the pulse velocity profile for atleast 0.5hr in which atleast 180000 such time steps are generated

Sir i think with this answer i will be able to solve my problem. sir pls find the velocity profile

Thank you!!!!!!!1 

 Soory sir i had not seen the previous answer

Thank you very much sir!!!

Sir can pls suggest me the the following values:

Time Step Sizes=?????

Number of Time Steps=??????

Maximum number of iteration/Time step= ?????

Thanking you once again for a beautiful support

Hi Siddiqui, It is my pleasure to help fellow researchers.  Time step size will depend on your grid or element size of your mesh and expected maximum speed in your domain. Please follow the Courant Frederich Lewy (CFL) criteria.   As for number of time steps, it depends on how long is your region of interest, it is roughly the average speed x total time(time step x number of time step). I hope this is of help.

Ok sir i am going through

let u know for tge further queries n updates

Hi sir,

this is the error which i am getting in fluent while uploading the programme

Error: C:\\Users\\SIDDIQUI UMAIR\\Desktop\\New Text Document 3.c: line 7: parse error.

Error: C:\\Users\\SIDDIQUI UMAIR\\Desktop\\New Text Document 3.c: line 8: t: undeclared variable

Please check that there is a semicolon

double t = (CURRENT_TIME*0.01-floor(CURRENT_TIME*0.01))/0.01;

or you can attach the udf file to see where are the errors in the code

Sir this is done error rectified.

Thank you

Running the iteration with your udf

will let you know the result after 3 hours

Hi, Yoon - Song Chong Sir and every body

sir the program me is successful one small query ?

is the same user defined program me is valid for CFX if any change sir please tell me

I have attached the udf 

Hi Siddiqui, As far as I know it should work for Fluent as well as CFX

Thank you sir, ill trie

have a good day

Hello every body and Chong sir  can any body please tell me how to use UDF in CFX procedure wise

#include "udf.h"

DEFINE_PROFILE(unsteady_velocity, thread, position)

{

face_t f;

real t = CURRENT_TIME;

begin_f_loop(f, thread)

{

F_PROFILE(f, thread, position) = 20. + 5.0*sin(10.*t);

}

end_f_loop(f, thread)

}

UDF for sinusoidal pressure/velocity inlet in FLUENT? - ResearchGate. Available from: https://www.researchgate.net/post/UDF_for_sinusoidal_pressure_velocity_inlet_in_FLUENT [accessed Dec 20, 2015].

Hello ,Can someone please give me a udf for the below attached file.it is a square wave udf..but i am not sure hot maintain the frequency??

are u using CFX or fluent as a solver.

I have solved recently this problem

You can mail the details on [email protected]

 Hi..

I am using the fluent solver ..The problem is pulsating veloicty jet inlet on the hole on the wall of the aerofoil which has a diameter of 0.2mm. Its a boundary layer control problem.I have to solve for 5 seconds.The velocity has to be changed accordingly with respect to the attached file.It is a 3d problem with an aerofoil of  20mm thickness thickness ..

Dear Chong,

Firstly i would like to thank you for publishing your code online. That is very useful for users like me, who are new to UDFs.

I have got a simple stenosed artery model which i would like to analyse. I tried to implement your code (published dated 23 july 2015) but i am getting error for line 6: syntax error. Do you know why i could be getting it.

I have attached your code for quick look.

Alternatively if somebody else have figured the cause of the problem please let me know.

Thank you very much in advance

Regards

Shiju

Hi Yoon-Song Chong,

I am running blood flow simulation in abdominal aortic aneurysm, I want to use UDF for inlet parabolic velocity profile but for steady-state simulation. Do you or anyone have an example code for 3D inlet geometry. Tq

Hi Yousif Ali,

Are you performing idealized geometry or patient specific CFD simulations?  Either case, you have to make sure that you set the origin correctly.  For patient specific case, you have to align the center of the aorta inlet to the global origin so that you can apply the formula (x*x+y*y)/r*r correctly for parabolic profile assuming the blood flow is in z-direction and r is the radius of the aorta.

You can use my example code but ignore the time-dependent portion.

I hope this is of help.  

Hi Yousif Ali,

As discussed, attached is the udf that I used for Fluent CFD simulations.  It is a fully developed parabolic and transient triphasic profile to be prescribed at the inlet boundary.  The peak systolic velocity is 1m/s, period 0.93 second and the inlet diameter is 19mm in this case.  You will have to adjust these three parameters to suit your patient specific cases.  Please note that the center of the inlet must be aligned with origin of Fluent's global coordinates otherwise divergent will occur.  Good luck with your simulations. 

I am running simulations of blood flow in artery but having problem with selecting the pressure outlet and operating conditions in fluent.

I am confused about outlet pressure whether it should be zero pressure outlet or mean of systole and diastole pressure and what should be the operating condition ? (default operating pressure in fluent is 101325 Pa.)

For inlet condition I am using the velocity_inlet with parabolic profile using UDF.

Does anyone has any suggestions related to this problem ?

Hi Krunal,

As far as I know, it is the pressure difference that drives the fluid flow between inlet and outlet therefore it doesn't really matter what gauge pressure you set at outlet since you are using inlet velocity as your inlet boundary condition. Some prefer to set it to zero gauge pressure. As for me, I just set the average of systole and diastole 100mmHg or 13332 Pa. I hope this is of help.

Thank you, Yoon - Song Chong Sir

Your suggestion is very helpful.

Hello everyone!

can I get some help here? I am stuck with a problem.

i need UDF for creating pulsating ocean wave in a wave tank.

The wave height should be 17cm and wave length 22 cm. for water depth 14 cm. can anyone please provide?

Albert Yoon-Song Chong Sir, It will be really helpful for me If you kindly help me out with the following User Defined Function to run a blood flow simulation in an idealized stenosis artery. I have got the UDF from this discussion. Can you please share the mathematical function you used to develop the following User defined code and functions nature in the graph.

#include "udf.h"//file that contains definitions for define functions and fluent operations

#define PI 3.141592654

DEFINE_PROFILE(inlet_z_velocity,th,i)

{

        float x[ND_ND]; /* an array for the coordinates */

        float xx, y, r;

 face_t f; /* f is a face thread index */

        r = 0.01; /* inlet radius in m */

 begin_f_loop(f,th)

        {

  double t = (CURRENT_TIME*1.25-floor(CURRENT_TIME*1.25))/1.25; //t is the local time within each period

                F_CENTROID(x,f,th);

  xx = x[0]; /* x coordinate */

                y = x[1];   /* y coordinate */

  if(t >= 0 && t 0.1 && t 0.175 && t 0.425 && t 0.5 && t

Hi everyone,

I am trying to solve flow over cylinder but i have some problem. When i look velocity vector at in front of inlet it is developing flow but it has to be fully developed flow. I should write C code and ı am not good at C program. I saw some code and ı tried to import my case but it doesnt work. Does anyıone help me for it ?

Hi Albert Yoon-Song Chong sir and avry body. I want to simulate one wave in air water two phase flow using udf cods in 2d and 3d pipe. But i cant. I want to simulate one wave that move to end of the pipe without any changes. Do any suggestion to help me?