I've been bashing my head against running a two-way coupled analysis on our local HPC. I've tried reading and applying the various command line tutorial but those get me stuck as well sadly. My predecessors here used system-coupling in an old ansys version(17.1). But now we are working with a newer version, 2019r1 and I have no documentation available to me to pull this off. I've extensively tried to follow the oscillating plate tutorial with command line options provided to me by Ansys, but I can't get it to work for me.

- Fluent.cas file
- Mech.dat file
- Coupling.sci file
- A journal.jou file for fluent
- a Shell file to submit the simulation to the cluster
- and lastly a .py file to give commands for system coupling.

My shell file does not have much information in it, just some basic IO and the following command:

Where the inputfile.py looks like this:

The only output this gives me sadly is:

sed: couldn't flush stdout: Broken pipe

My predecessor worked with an older version, but too be fair, I'm a bit over my head here and I don't know where to start to get this to work. Before I used a totally different shell file:

which used the following commands(leaving out file I/O)

The problem here sadly is that the simulation would just close with the following error:

So I was stuck there as well, and as i understand from other forum posts the above mentioned method is outdated and no longer supported.

Have all the needed files for my FSI simulation but cannot get the command line setup to execute the simulation on my HPC. Hoping that someone can point me to a good tutorial or to what I am doing wrong here.

With Kind Regards,

Rene Hiemstra

]]>On searching the net I was suggested'

parallel >> timer >> usage

but it gives me a value which is much lesser than the actual time. This actually took me more than 1.5 hours and the Total Wall Clock Time was just 29.610 sec.

Can anyone tell me how I can find the actual time taken up for the Ansys Fluent simulation?

P.S. I also tried the (solver-cpu-time) and the values are too large.

I understand that we need to enable something called Data Sampling to get the Mean Velocity.

I am doing RANS simulation and don't see Data Sampling anywhere, does it mean that we cannot obtain Mean Velocity Contour for RANS simulation? I know that we need to have a way to get Mean Velocity, I would be grateful if someone could guide me to the right direction or point out some resources.

And what is the correct method to obtain the Mean Velocity in case of an Unsteady Simulation?

]]>I am trying to model flow of mixture of air and water vapor where vapor is pulled out into a vaccum domain

I want to provide an initial condition where mass fraction of vapor and air in the domain is 0 because the domain is in vaccum condition intially.

Fluent does not allow me to do this as sum of mas fractions of vapor and air in the domain should be 1. Is there any way i could assign this vaccum condition initially?

]]>#include "udf.h"

real v_prev = 0.0;

DEFINE_CG_MOTION(osc,dt,vel,omega,time,dtime)

{

FILE *pff;

FILE *pfv;

FILE *pfd;

Thread *t;

Domain *d = Get_Domain(1);

real x_cg[3], force[3], moment[3];

real accl, dv;

real pi = 3.14159265; //pi

real rho = 999.7297; //density of fluid

real dm = 0.08; //cylinder diameter

real mr = 1.93; //mass ratio

real damp = 0.070; //total damping ratio

real fnw = 1.052892; //natural frequency in water

real ca = 1; //added mass coefficient for cylinder

real md = pi*(dm/2)*(dm/2)*rho; //displaced mass

real ma = ca*pi*(dm/2)*(dm/2)*rho; //added mass for cylinder

real mass = mr*md; //oscillating mass

real k = (mass+ma)*(fnw*2*pi)*(fnw*2*pi); //spring stiffness

real c = 2*damp*sqrt(k*(mass+ma)); //damping of the system

int i;

NV_S(vel, =, 0.0);

NV_S(omega, =, 0.0);

t = DT_THREAD(dt);

for(i=0;i<3;i++)

x_cg[i]=DT_CG(dt)[i];

Compute_Force_And_Moment(d, t, x_cg, force, moment, TRUE);

force[1] += -k*x_cg[1] - c*vel[1];

accl = force[1]/mass;

dv = accl*dtime;

v_prev+=dv;

vel[1] = v_prev;

if(NULL == (pff = fopen("force.dat","a")))

Error("Could not open file for append!\n");

fprintf(pff,"%g ",time);

fprintf(pff,"%g \n", force[1]);

fclose(pff);

if(NULL == (pfv = fopen("velocity.dat","a")))

Error("Could not open file for append!\n");

fprintf(pfv,"%g ",time);

fprintf(pfv,"%g\n", vel[1]);

fclose(pfv);

if(NULL == (pfd = fopen("displacement.dat","a")))

Error("Could not open file for append!\n");

fprintf(pfd,"%g ",time);

fprintf(pfd,"%g\n",x_cg[1]/dm);

fclose(pfd);

}

DEFINE_CG_MOTION(osc_inert,dt,vel,omega,time,dtime)

{

vel[0] = 0.0;

vel[1] = v_prev;

}

]]>I am working with centrifugal compressors with CFX, Fluent, typical Turbogrid, BladeGen, etc.

I already finished the whole machine except the volute and I don't know where to start.

In general, what is the best way to do this today?

Thanks in advance.

]]>To put it simply, CFX5 won't launch through Workbench 2020 R2 or throught the CFX5 console. This means neither CFX-Pre and the CFX-Solver will start. I've been using the license without problem since September 2020. Its due to expire on December 31st 2020.

These problems came on suddenly. 16 solutions were succesfully solved but the last 6 with identical meshes would not solve. Giving me a 'not enough stack memory". I dont know what that means. Removing any mesh interfaces solved that issue.

The new meshes had the same node count but without any interface boundaries. The next problem was that CFX-Pre could not start when started through Workbench. I could however start it through the CFX5 console and inject my mesh into it. Then export the CFX case so i could import it into workbench to solve. This no longer works. CFX-Pre will not start.

I am currently trying to solve a previous set-up file but now the CFX-Solver won't start. The entire system appears to be deteriorating around me and i need this software for my dissertation..

No amount of uninstalling and reinstalling remidies this problem.

any ideas on a solution?

windows version : 10.0.18363 Build 18363

CPU : AMD Ryzen 7 3700X 8-core @ 4.4GHz all cores

RAM : 64GB @ 2400MHz, CL17

Storage, 3TB, 45% free space

]]>I trying to simulate the air flow inside my geometry and as is clear from the pictures I have created two gaps (red circles) inside my geometry (for seperation of flow). But the problem is that CFX completly ignores them and treats these gaps as fluid domain. How is it possible that I do not have any fluid domain there but CFX creats stream lines and shows air moving inside this gap?

How should I solve this problem?

a. I have double checked to see if that region is really empty (for example if some part of the geometry was not hidden)

b. You can see my mesh quality (min orthogonal quality is 1.1 and only for very small number of meshes.

c. soultion converges up to 10^-5 for momnetum and -9 for mass

regards

]]>In ansys fluent, how can we use a a tabulated data to describe the real fluid properties instead of the expression equations in UDRGM?

I found this method in a research work which Ansys Fluent v19 was used in it, whilst I was wondering how to Implement this approach. I didn't find any interface to import a tabulated data.

Best wishes！

]]>what is this part responsible for in the solar model in fluent , and how it affects the results ?

In particular, I would like to write a UDF to get the spatial derivatives of the Reynolds Stresses - d(uu)/dx, d(uv)/dy, d(uw)/dz - that I write with the Boussinesq approximation.

]]>I have a 3D case that works fine. Now I want to simulate only half of the domain by using a symmetry plane. Thats works fine as well until I introduce the UDS. I couldn't find information about UDS and symmetry plane. Do I have to set UDS gradient =0 at symmetry plane manually? Or what could be the reason for the divergence caused by UDS with symmetry plane?

]]>I was using Fluent to simulate a wind field. After the simulation is done, I tried to display the velocity contour of a curved surface a certain distance from the wall (ground floor). I used [Surface -> Create -> Transform] to define such surface.

However, there are white spots in the contour plot.

When the transformed curved surface is displayed in mesh mode, the white spots are still there, as well as some irregular lines.

In the numerical model, I only applied structured mesh (hexahedral cells). I also checked the mesh of the ground floor and everything is OK. No white squares and irregular lines are spotted.

Does anyone know where there are white spots and irregular lines in the transformed surface? Any help is appreciated.

Thanks.

]]>- I have tried copying mesh and geometry file from previous project directory into the SYS folder of the new workbench project folder but when I run the Workbench project file then the files are automatically deleted.
- I tried to import the geometry file separately and it worked but then I cannot import the Mesh file next. (I need to re-do the meshing on the geometry)
- I tried importing the mesh directly but the imported mesh is un-editable in workbench.

I want to import the mesh and geometry into a separate workbench project and be able to edit the mesh as well. Is there a way we can do that ?

If it is the same Workbench project then we can duplicate the mesh and geometry but what if we want to do the same but in different workbench project. Thanks.

]]>I am solving a case of turbomachines with CFX. I have several input parameters (speed, mass flow, etc.) but I don't know how I should set the temperatures as input parameters (in ANSYS Workbench):

Thank you in advance.

Regards.

]]>for my problem , inlet temperature is more than inlet temperature , i am using pressure based approach , transient conditions, density is ideal gas

even i monitored velocity also , it is very very low , so i cant sure about kinetic energy , and any other factor my temperature is more than inlet temperature

Pushing 250 [m^3/h] { 250/(60*60) *1.204 ~ 0.084 kg/s }

"1.204 is the air dencity at 20 degrees C"

I have 3 Coordinate systems, one for each fan!

I just want to understand what values are required for:

Axial component

Radial Component

Theta Component

Is axial the velocity of the outer edge of the fan?

Radial should be 0

Theta is radians per second for the outer edge of the fan?

Thank you in advance :)

I am simulating VOF model with dynamic mesh. I want to use the leveI set method to increase the accuracy. However I found in Ansys help (version 19.2) this sentence "The level set method is not compatible with the dynamic mesh model". Could you help me to explain about that?

Thank you so much !

]]>I am simulating a swirling flow inside a cyclonic filter.

I have tried two turbulence models :

1/ Standard K_Epsilon

2/ K-epsilon -RNG Swirling flow ( this is the most appropriate for swirling flows)

Methods are :

3/ Simple for velocity pressure coupling ----For both turbulence models

4/Upwind second order for turbulence , mass , momentum ---- For both turbulence models

Calculation with the standard K epsilon model , has converged after 280 itérations

But Calculation with the K epsilon RNG swirling flow model has not converged , the residus curve continued to oscillate ( instable) even aftre 6000 iterations ( a time of 1:30 h) . I have stopped calculation , because i have concluded that solution will not converge , i will just waste my time.

For more details , please see figures attached.

Since RNG Swirling flow is better than k epsilon to model cyclone inside filter , i prefer using it rather than standar k epsilon model.

I ask you please :

A/ Can you find an explanation about why the residus curve oscillates continually and solution has not converged even after more than 6000 iterations ?

B/ Advice me to change something ? in the order to accelerate convergence of the model ?

Remark : I have tried to reduce the value of residus from 0.001 to 0.01 but , also used Quick Scheme instead of simple but the problem is always here

Thanks tou very much

Kind regards

I want to simulate acoustophoresis phenomena. My system is composed by a fluid domain surrounded by a mechanical structure and a piezoelectric transducer. The fluid domain has an inlet where the fluid (water) and particles are inserted. The piezo vibrates and pressure waves are generated in the channel.

My goal is to have acoustic plane waves in the channel where the fluid flows (channel), but I have problem with the outlet I think. From the picture below, a perturbation origins from the outlet and travels into the channel. I have mass flow inlet and pressure outlet (I used the compressible liquid method).

Do you have any advices?

I have simulated steady state high temperature laminar flames with the chemical reaction. I have included the reactions and all the required Arrhenius parameters in the reaction panel. I followed the procedure as in the fluent manual. First the cold flow is created and disabled all the cold flow equations. Then volumetric reactions are activated and solved all the species. Later all the equations (i.e., species, energy and flow) are activated. While solving I reduced the under relaxed factor (URF) of energy to 1e-05 in order to increase the stability of the flow. And I got all the distributions of species included. Now, if I try to increase the URF, I find all the species distributions to get diffused.

Whether reducing the energy URF value to 1e-05 is correct? or what should I do now?

Please kindly give me few suggestions and your thoughts on it.

]]>I am trying to simulate transcritical methane for heat transfer in rocket cooling systems. I imported methane from NIST and should I have to create a lookup table? In my case fluid changes its phase within the simulation so do i have to do anything extra to include this phase change.

]]>I'm using a udf to specify the properties of one degree of freedom body (rotation around z-axis), I compiled it using the built-in compiler in Fluent 2020 R2. I specified the cg location in the dynamic mesh zone, but it seems like the location doesn't get updated with the solution. The motion file has the same value for the cg point with different angles.

Am I missing something? Why the cg location doesn't update? I think it's affecting the solution, that's why I need to fix it.

This is the UDF I used:

#include "udf.h"

DEFINE_SDOF_PROPERTIES(flap, prop, dt, time, dtime)

{

prop[SDOF_MASS] = 21.5;

prop[SDOF_IXX] = 5;

prop[SDOF_IYY] = 5;

prop[SDOF_IZZ] = 5.5;

prop[SDOF_ZERO_TRANS_X] = TRUE; /* boolean, suppress translation in x-direction */

prop[SDOF_ZERO_TRANS_Y] = TRUE; /* boolean, suppress translation in y-direction */

prop[SDOF_ZERO_TRANS_Z] = TRUE; /* boolean, suppress translation in z-direction */

prop[SDOF_ZERO_ROT_X] = TRUE; /* boolean, suppress rotation around x-axis */

prop[SDOF_ZERO_ROT_Y] = TRUE; /* boolean, suppress rotation around y-axis */

prop[SDOF_ZERO_ROT_Z] = FALSE; /* boolean, suppress rotation around z-axis */

printf("\nflap: updated 6DOF properties");

}

and the motion file I got:

I had a blue screen problem on my Windows 10. After that I cannot continue my simulations on Fluent.

I am always receiving a message --> Error: token too large Error Object: (), when I try to initialize a case or run it.

Could someone help me to solve it ?

Thanks in advance

]]>/file/read-case /home/example.cas

/file/read-data /home/example.dat

/solve/set/data-sampling yes 6 yes yes yes

/solve/set/time-step 0.005439874

/solve/dual-time-iterate 1440 50

/file/write-case /home/results.cas

/file/write-data /home/results.dat

exit

OK

]]>I'm trying to simulate a single solid spherical particle moving in a fluid flow in a pipe. This particle moves due to the drag force and some external force that is imposed through an UDF. The flow is laminar and I'm using a simple 2D steady simulation, but this can change if needed.

I've used the DPM model to do this. However, when I increase the particle diameter, the fluid flow does not change. I've tried turning ON the "Interaction with Continuous Phase", but there was no change. Is there a way to couple the fluid flow and particle trajectory calculations, such that the particle diameter has an impact in the flow?

I've also tried the MPM model, which takes into account the volume of the particle and also uses the lagrangian method. However it seems that there is no option to impose an external force onto the particle through an UDF. Is there a way to do this?

If neither of these are possible, which model / strategy would be suited in this case?

]]>I am simulating a steady state laminar flow inside a reactor operating at low pressure and high temperature. If I use low value of momentum URF (0.2 in my case) my solution gets converged up to 10-14 and net fluxes are almost 0 and point monitors also gets steady. But if I increase the URF value from 0.2 to 0.3 and re run my solution without initializing it again gives me convergence only up to 10-1 and my fluxes are also not close to zero.

My question is can I trust the solutions obtained from low value of URFs?

Regards.

]]>I am simulating melting behaviour of glass substrate with laser as heatsource.can anyone please explain How can I simulate melting and solidification of glass in Ansys fluent.

]]>I am currently simulating the effect of bubble in liquid immersion cooling. My cooling tank consists of 16 CPUs and 16 heat sinks and cold plate (surrounding the tank) and it is filled with viscous fluid. Air is injected from the bottom of the tank.

I was able to demonstrate the circular convection of fluid and upward flow of Air. When air flow rate is increased, it is expected to help the heat transfer from heat sink to the fluid and thus it cools the CPUs.

However, no matter how much air I increased, air volume fraction near heat sinks reminds around 1%. I even increased it from 1.0 g/s to 10.0 but with no difference in fraction. Which means no increase in cooling efficiency.

Reading the manual and trying out different parameters haven't solve this problem. Please kindly point me to the right direction, possibly a section in the manual or a tutorial video.

Air volume fraction is measured at 20 different points in and around heatsinks, inlet and outlet.

Thank you very much for your kind attention and help in advance.

Nat

]]>I have to understand the flow dynamic through the gas cylinder with inlet and outlet and it include the piston too, outlet fluid is affected by the vertical oscillation movement of the piston.

b. How can we create that movement using moving mesh, do I need to write udf? For UDF, can we use matlab or just C++ code? How should I begin with it?

c. I am not sure, how can create the movement or start tacking this problem? Can anyone suggest any good example to begin with, later on I can look for the complexity?

d. If the problem have motion , I think 2-Way FSI must be adopted with the transient analysis? please correct me if I am wrong.

thanks.

]]>1- The domain is a closed one (think of a closed container) . There is no inlet-outlet. Believe me that CFD problems without mass exchange (inlet-outlet) are tough, especially incompressible ones!

In this problem The driving force of dynamics is only "gravity". Assume that we use a big value for gravity to accelerate the process and overcome the time scale problem.

2- Assume that the Number of particles is extremely large. Let's say 100 billion. So, DDPM is literally useless. Keep in mind that the particles are tiny. Despite their large number, the initial volume fraction is assumed to be 0.4, so that DDPM assumptions are not violated.

3- If we start with a uniform suspension (tiny solid particles in an "incompressible" fluid) with volume fraction 0.4, we expect that the particles start to settle and after sufficient time, they aggregate at the bottom of the container. They should build up a pile of particles whose volume fraction can reach up to packing limit. at the end we should have two separated sub domains! I checked Euler-Euler approach in which the particles are treated using granular constitutive relation, but No success. DDPM is out of use due to large number of particles.

Any Idea?????

]]>My name is Neel Shah and I am trying to run a spray simulation on Ansys Fluent.

I have got a Ansys student version Ansys 2020 R1.

I have gone through many webinar of the same and I found that everyone is using workflow tool in meshing section. However, when I open Ansys meshing part I am not able to see any workflow option. Could you please guide me through this?

Waiting for your prompt reply.

Thank you

Neel Shah

Email ID:- shahn3@my.erau.edu

]]>How to resolve this problem

]]>I am doing a steady-state two D catalytic combustion of methane with fluent. I found the settings of URF would change interation results. To be specific, when URF of species is 0.7, then steady state combustion could not be sustained. However when this value increases to 0.8, then steady flame could be maintained.

As far as I am concerned, URF would only influence speed of convergence instead of iteration results. I am wondering why would this happen? That is URF changes iteration results

Best regards

weiqiang

]]>What model should I use?

Any particular settings I should be paying close attention to? Boundary conditions?

So far, I've been getting these errors:

"Update-Dynamic-Mesh failed. Negative cell volume detected." I had to reduce the step size to 1e-30 to avoid this error but this got 0% through the calculation in 5 hours.

"Minimum Orthogonal Quality is less than 0.01. Consider improving the mesh quality before proceeding with your simulation" However, I am on the educational exe. and can only use <512,000 cells.

The setup also closes after I click "Calculate" and sometimes the case gets thrown.

]]>recently I've been using fluent to simuation some combustion application with detailed chemical mechanisms. I found inelt condition would influence convergence performance of my case.

For example, when inlet velocity of fuel air mixture is 0.3, then convergence would be bad and there are unsmooths and flunctuations in profiles report. However, when I increase inlet velocity to 0.5, then resutls are very smooth. The other settigns are just exactly the same. I am wondering why would this happen? why would just inlet velocity influence convergence performance?

Best regards

weiqiang

]]>I am trying to simulate a simple flow through a smooth pipe. Later on I plan to add porous structure on its interior, but only after I get a sattisfying results from this simple model first.

I used a quarter model, pressure inlet of 50000 Pa, pressure outlet of -50000 Pa at an operating pressure of 150000 Pa. As I understood, this would equal to an inlet pressure of roughly 2 bars and outlet pressure of 1 bar. This is the solution of pressure I got:

As you can see from the solution, absolute pressure at the outlet (left side) is what I expect it to be - 1 bar. But on the inlet (right side) it is just a little bit more, even though I expect it to be 2 bars. Why is this the case? I tried calculating the pressure myself for inlet and outlet:

I add the operating and the total pressures together to get the absolute pressure, but I dont get the same results for the absolute pressure as ANSYS. Why is this the case?

Thanks in advance!

Stefan

]]>e.g. if (0 < t < 5, dt = 1) and if (5 < t < 10s, dt = 2) and if (t > 10 s, dt = 3)

When I tried to insert more than 1 if statements, it syntax error.

I tried this,

if (t <= 5[s], 1[s], if(and(t <= 10 [s],t > 5[s]), 2 [s], if(and(t > 10[s]), 3 [s])))

But when I run it, the flow times were mixed up. I think it goes like 1s, 5s, 7s, 9s, ...

Thanks.

]]>