[Taz-CFD]

Hello fellow CFD compatriots. I am trying to model erosion within Fluent using the DNV erosion model. I am pretty new to UDF's but have managed to come up with the the following. Can someone please look at it and let me know if looks okay?

The equations I am basing it on are on page 11 on http://brage.bibsys.no/xmlui/bitstre...=1&isAllowed=y

Heres my code:

#include "udf.h"
#include "math.h"
const float PI =3.141593;
float A1 =9.37
float A2 =-42.295;
float A3 =110.864;
float A4 =-175.804;
float A5 =170.137;
float A6 =-98.398;
float A7 =31.211;
float A8 =-4.172;
float K= 2.0e-9; /* material constant /*
float pt=7800; /* density of steel/*
float at=0.0005; /* target area /*

DEFINE_DPM_EROSION(DNVV, p, t, f, normal, alpha, Vmag, mdot)
{
real erosion;
real a=alpha;
real v=Vmag;
real F_alpha;
real m=mdot;
real V;
real F;
V=pow(v,2.6);
F=A1*pow(((a*PI)/180),1)+A2*pow(((a*PI)/180),2)+A3*pow(((a*PI)/180),3)+A4*pow(((a*PI)/180),4)+ A5*pow(((a*PI)/180),5)+A6*pow(((a*PI)/180),6)+A7*pow(((a*PI)/180),7)+A8*pow(((a*PI)/180),8);
erosion=(m*K*V*F)/(pt*at)
F_STORAGE_R(f,t,SV_DPMS_EROSION)=erosion;
}

Looking forward to your replies! Here's my email if you'd prefer to email me on there: taz_78684@hotmail.co.uk

Kindest regards,
Taz

[pakk]

An important test to see if you made mistakes is to just compile it. If you get errors, you did something wrong. (If you don't get errors, it does not mean you did not make mistakes!)
The code you showed above has some mistakes that the compiler would have seen immediately, so it looks like you did not even try your code!

I can see at least two things wrong in the code.

Your comments in the code are wrong:

float K= 2.0e-9; /* material constant /*

should be

float K= 2.0e-9; /* material constant */

Also you forgot a semicolon ( ; ) after the following:

erosion=(m*K*V*F)/(pt*at)

[Taz-CFD]

Originally Posted by pakk

An important test to see if you made mistakes is to just compile it. If you get errors, you did something wrong. (If you don't get errors, it does not mean you did not make mistakes!)
The code you showed above has some mistakes that the compiler would have seen immediately, so it looks like you did not even try your code!

I can see at least two things wrong in the code.

Your comments in the code are wrong:

float K= 2.0e-9; /* material constant /*

should be

float K= 2.0e-9; /* material constant */

Also you forgot a semicolon ( ; ) after the following:

erosion=(m*K*V*F)/(pt*at)

Hi there pakk, thank you for your reply! i have made the following changes:

#include "udf.h"
#include "math.h"
#define PI =3.141593
float A1 =9.37;
float A2 =-42.295;
float A3 =110.864;
float A4 =-175.804;
float A5 =170.137;
float A6 =-98.398;
float A7 =31.211;
float A8 =-4.172;
float K= 2.0e-9; /* material Constant */
float pt=7800; /* density of steel */


DEFINE_DPM_EROSION(DNVV, p, t, f, normal, alpha, Vmag, mdot)
{
real erosion;
real a=alpha;
real v=Vmag;
real F_alpha;
real m=mdot;
real V;
real at;
at=(PI*pow(0.026,2))/4.*sin(a); /* area exposed to erosion */
V=pow(v,2.6); /* impact velocity of particle raised to constant n */
F_alpha=A1*pow(((a*PI)/180),1)+A2*pow(((a*PI)/180),2)+A3*pow(((a*PI)/180),3)+A4*pow(((a*PI)/180),4)+ A5*pow(((a*PI)/180),5)+A6*pow(((a*PI)/180),6)+A7*pow(((a*PI)/180),7)+A8*pow(((a*PI)/180),8);
erosion=(m*K*V*F)/(pt*at);
F_STORAGE_R(f,t,SV_DPMS_EROSION)=erosion;
}

But now I get an error referring to this line - at=(PI*pow(0.026,2))/4.*sin(a); /* area exposed to erosion */

It says invalid type for pointer deference: double

Thank you once again for your help,
Regards

Taz

[pakk]

Two relevant lines:

#define PI =3.141593
...
at=(PI*pow(0.026,2))/4.*sin(a);

The define line means: everywhere where is written "PI", act as if it says "=3.141593".
Let's do that:

at=(=3.141593*pow(0.026,2))/4.*sin(a);

This does not make sense, and I can imagine that Fluent has some problems with this line. I don't know how this leads to the error message you saw, but I am no compiler...

One solution: remove the "=" from your definition:

#define PI 3.141593

A different solution, better in my view: don't define your own pi, but use M_PI that is already defined by Fluent:

at=(M_PI*pow(0.026,2))/4.*sin(a);

By the way: are you sure you don't miss any brackets in this expression? I don't want to dive into your model, but I am worried it should be this:

at=(M_PI*pow(0.026,2))/(4.*sin(a));

[pakk]

While I am at it: I would simplify your code...
Why do add a variable a that is the same as alpha? Why v the same as Vmag? Why m the same as mdot?

And you use "(a*PI)/180" a lot in your code. Make that a variable!
But I think it is even wrong to use this... Angle a is already in radians, but it looks like you think a is in degrees and you still have to do the conversion...

Thirdly: you use "F" in your erosion calculation, but that is never defined. And you define F_alpha, but you never use it. I guess they refer to the same thing...

Altogether:

DEFINE_DPM_EROSION(DNVV, p, t, f, normal, a, v, m) 
{
real V = pow(v,2.6); /* impact velocity of particle raised to constant n */;
real at = (PI*pow(0.026,2))/4.*sin(a) /* area exposed to erosion */
real F_alpha=A1*pow(a,1)+A2*pow(a,2)+A3*pow(a,3)+A4*pow(a,4)+ A5*pow(a,5)+A6*pow(a,6)+A7*pow(a,7)+A8*pow(a,8);

F_STORAGE_R(f,t,SV_DPMS_EROSION)=(m*K*V*F_alpha)/(pt*at);;
}

[Taz-CFD]

Originally Posted by pakk

While I am at it: I would simplify your code...
Why do add a variable a that is the same as alpha? Why v the same as Vmag? Why m the same as mdot?

And you use "(a*PI)/180" a lot in your code. Make that a variable!
But I think it is even wrong to use this... Angle a is already in radians, but it looks like you think a is in degrees and you still have to do the conversion...

Thirdly: you use "F" in your erosion calculation, but that is never defined. And you define F_alpha, but you never use it. I guess they refer to the same thing...

Altogether:

DEFINE_DPM_EROSION(DNVV, p, t, f, normal, a, v, m) 
{
real V = pow(v,2.6); /* impact velocity of particle raised to constant n */;
real at = (PI*pow(0.026,2))/4.*sin(a) /* area exposed to erosion */
real F_alpha=A1*pow(a,1)+A2*pow(a,2)+A3*pow(a,3)+A4*pow(a,4)+ A5*pow(a,5)+A6*pow(a,6)+A7*pow(a,7)+A8*pow(a,8);

F_STORAGE_R(f,t,SV_DPMS_EROSION)=(m*K*V*F_alpha)/(pt*at);;
}

Hi there pakk

Thank you so much for your reply! I really appreciate it. I will implement these changes now and let you know how I get on. Once again thank you so much!

Kind regards
Taz

[mali28]

Just a minor correction. It should be:

F_STORAGE_R(f,t,SV_DPMS_EROSION)+=(m*K*V*F_alpha)/(pt*at);

You have to remove the density of the wall from the denominator in the above equation if you want the unit to be the same as reported in
Fluent.

[edilbertolg]

Hi, I was wondering if you could get the correct coding for DNV erosion prediciton... I am trying to use your code but I get errors

[haghasa]

One more thing to consider is the data structure in C

its a better practice to always stick to "real" format, both for double (~ for double precision) and float (single precision) datatypes

For example:
change
float A1 =9.37;
to
real A1 = 9.37;

and try to define your constant numbers in the same format too...

change
pow(0.026,2)
to
pow(0.026,2.)

pay attention to the difference between 2 and 2. (=2.0)

[Ya Hong]

Originally Posted by edilbertolg

Hi, I was wondering if you could get the correct coding for DNV erosion prediciton... I am trying to use your code but I get errors

Do you already know the correct program, can you tell me? I really need it right now.

[pakk]

What is not correct about the code that is given above? If you really need it now, then try it. If you get errors, share them.