segmentation fault error due to zero flow at porous jump boundary

  • 212 Views
  • Last Post 19 February 2019
  • Topic Is Solved
zypresse91 posted this 08 December 2018

My problem is that I keep getting this error of Segmentation fault due to the zero flow at porous jump boundary. I tried to create cell zones just above the porous jump boundaries and set their Y-velocities as '0' to ensure that there's no slip in-situ boundaries. It didn't work out. Also for the values of porous jump boundary condition, how do I assign them?(I know that thickness is '0').

Any comment would be much appreciated. Thanks in advance.

Cheers,

Selvi

Cheers, Selvi

Order By: Standard | Newest | Votes
abenhadj posted this 09 December 2018

Add some screenshots so that we understand geo and model.  For proud jump parameters follow the guidelines documented for porous cell zone condtions. 

Best regards,

Amine

  • Liked by
  • zypresse91
zypresse91 posted this 09 December 2018

Dear Amine, thank you for sharing your insight.It is much appreciated.My system is composed of 3 bodies, and the middle body will be the host of a reaction. All those three systems have the porous zone option activated. I want the water liquid which flows through the undermost body and arriving the porous jump boundary, passing through this boundary, reacting with the chemicals I designated in the middle body, creating those other products, and finally once again those products for this time passing through the second porous jump boundary and reaching to the uppermost body.

Cheers, Selvi

zypresse91 posted this 09 December 2018

I hope you could imagine the system. Now I created 2 cell zones just above these porous jump boundaries with a width of 1 grid cell which was 8 m for my case, and also set the Y velocities of those interior cell zones as 0. Logically, since they are just above the porous jump boundaries they are actually in-situ of the reaction area(the middle body). So, I did activate the porous zone and reaction options for my cell zones, and set '0' value for the Y velocities in the fixed values tab(a researcher proposed this method).

Cheers, Selvi

abenhadj posted this 09 December 2018

Why yvelines is set to zero? You will have no flow normal to the jumps and so massflow is zero. 

Best regards,

Amine

  • Liked by
  • zypresse91
abenhadj posted this 09 December 2018

One grid cell the X direction?

Best regards,

Amine

zypresse91 posted this 09 December 2018

Sir, I understand your point, I didn't propose this method, a PhD researcher tried it out and he said the system is solved in his case. So, I just wanted to give it a try. Still, even if I don't set the Y-velocities(tangential velocities) as '0', I do still keep getting the error of segmentation fault, detected zero flow at the porous jump. as for cell zones I created, it was the method itself, creating cell zone just above the porous jump to trick the solver as if there's no slip at the porous jump so it solves the system.I hope I made it a bit more clear.If you think its not the solution, what do you suggest to get rid of the "zero flow" error Sir?

Cheers,

Selvi

Cheers, Selvi

abenhadj posted this 09 December 2018

What are you trying to setup? I still do not understand the model.

Best regards,

Amine

zypresse91 posted this 09 December 2018

Ok.Let me explain.It's a 2D earth-system model which represents the fluid flow through a system which composed of 3 bodies.In the images I shared to you,you see the bodies Surface Body1,Surface Body2,GHSZ(the region which hosts a chemical reaction).Now, all those 3 bodies are set as porous medium.You see the boundaries which I assigned as porous jumps.The system presents the following: Water liquid will be flowing from undermost body(Surface Body2) reaches to POROUS JUMP-2 passes through it(since it acts like a membrane,thickness of it is 0), reacts the chemicals I designated in the middle body.Once it reacts,products will be produced.Then,those products will be passing through POROUS JUMP-1 and reaching out the Surface Body1.This is it.But when I run it, I receive this error: fatal error, segmentation fault, detected zero flow at the porous jump boundaries(regardless of setting Y(tangential)-velocities as '0').If you could lead me on that matter I would be much appreciated.You may contact me if you like: turkdogans@itu.edu.tr

Cheers,

Selvi 

Cheers, Selvi

abenhadj posted this 09 December 2018

Why is the thickness zero? With zero thickness there is no pressure loss.

Yvel is normal and not tangential to the jumps.

So you have inlet at the upperbottom? Can you disable reactions and start just by looking into the flow.

Best regards,

Amine

  • Liked by
  • zypresse91
zypresse91 posted this 09 December 2018

Ok Sir, I read a couple different sources and manuals of ANSYS which says since it behaves like a membrane it's thickness must be set to "0". If it shouldn't be zero, How do I decide the value?I have a mesh system of 5000mx5000m and 40000 elements(grid) in it.Actually, reaction is the primary cause of the fatal error,at first our system didn't have this reaction part,and it worked out well.Reaction is causing a problem at these boundaries but I cannot find why Sir.Also I read that discrete phase boundary should also be activated if you're dealing with porous jump.I did also activated it,still no solution.Do you think the reason that I keep getting error is the values?I need to find out how to set these values of FACE PERMEABILITY(default:1e+10),thickness,pressure jump coefficient,and thermal contact resistance(default:0.1).Where do I find these values or how do I calculate them for my system of 2D planar?I'm also looking for.Thank you very much.

Cheers,

Selvi

Cheers, Selvi

abenhadj posted this 09 December 2018

Who did till you all that regarding thickness and discrete phase? Which references?

Regarding parameters: check my first reply . If you do not have data stick to default and high permeability membrane or just do not use porous jump.

PS: If you are research fellow (not undergraduate student) from an university in Europe please contact your channel partner there to follow up.

Best regards,

Amine

  • Liked by
  • zypresse91
abenhadj posted this 09 December 2018

You are welcome anytime. But having a width of one cell is wrong along the x direction. If you think that the error is due to reaction add some more details on that.

Best regards,

Amine

  • Liked by
  • zypresse91
zypresse91 posted this 09 December 2018

Sir, don't worry about the cell zones now bc me and my professor just figured out why its not working. As you told me, the default values of the porous jump boundary condition has to be assigned regarding to the constructed system.For our system, the size of one grid is 50mx50m so thickness could be at most 50m(isn't it right?), now I have to find the corresponding values of pressure-jump coefficient(C2) and face permeability(default:1e+10).And then I will try it out again.

Thank you so much for your comments.It means the world to me.

Cheers,

Selvi

Cheers, Selvi

abenhadj posted this 09 December 2018

If the thickness of the membrane is 50 meter then you need to give as input in the porous jump panel. The same applies by the way for the normal porous zone conditions. The C2 term is relevant for turbulent flows (or higher velocity= higher inertia). Actually setting all porous parameters requires either measurement of pressure loss, or correlations or input from the manufacturer. 

Best regards,

Amine

  • Liked by
  • zypresse91
zypresse91 posted this 09 December 2018

Sir, now everything has cleared in my mind. I left the value of face permeability as default, i.e. 1e+10, thickness of the porous medium as 50 m and C2 value as "0". As for C2 value, I read through ANSYS manual that it could be assigned as "0" for laminar flow simulations just like you said bc it's actually for turbulent flows. Unfortunately, when I run the program it still doesn't respond, no solution. I keep getting the error segmentation fault "zero flow" at porous jumps.Please see the screenshot I shared.The video you shared earlier actually explains very well how one could derive the coefficients of a porous region. Could we apply the same procedure you did?(curve fitting in excel by referencing the values of VOLUME FRACTION of SOLID MATERIAL and DIMENSIONLESS PERMEABILITY BETA of GLASS WOOL) They are a bit strange for my case since our porous jump is actually a transition zone between marine sediments, imagine 3 layers(bodies) and these porous jumps behave like membrane which allow water liquid flow and passes through them.So, for that kind of a membrane can we talk about some properties to be useful to extract the coefficients by curve-fitting them in excel? If so, which values(I mean you derived coefficients from experimental pressure and velocity data and for laminar flow one can derive coefficients by volume fraction of solid material and dimensionless permeability beta of gw) and how ?

I hope I didn't confuse you.Sir I cannot express my graditute for your time and consideration. You literally helped me out to realize what I'm dealing with, what is it that I'm missing and at which points I'm doing things completely wrong. I thank you Sir, for your generosity to share your knowledge.

Cheers,

Selvi

Cheers, Selvi

abenhadj posted this 09 December 2018

Just use interior zone and not porous jump if they just interface or transition between different porous layers. 

Porous prameters are user's responsabilty and are poros media dependent (intrinsinc).

 

Best regards,

Amine

zypresse91 posted this 09 December 2018

Sir, I understand you. At first, we didn’t define them as porous jumps, actually we didn’t define them(porous jumps, i.e. boundaries btw the marine sediment layers) at all. But then this error arrived : Error: Species are not solved in neighbor cell threads (14 and 16)\n of interior zone 6. \n Error Object: #f Then I searched the cause of this error and saw a comment on cfd forum that suggests to define transition layer as porous jump which should conceptually behaves like a membrane and let fluid to flow through itself. But here it’s, just can’t find what’s wrong with it. Is it the values ? Not sure yet but other than values , everything seems in its place. I’m keep searching, asking and sending messages to experts like you, me and my professor will have another session hopefully she will also give some insight to the problem, so can’t wait to find some answers and run the program properly. I thank you again, and have to say that I am much appreciated for your time to read my messages, care to answer them and give some insight to me. Cheers, Selvi

Cheers, Selvi

abenhadj posted this 10 December 2018

I am not sure about anything now as all is getting confusing now. Just describe in precise manner each part of the model. Post screenshot fom ANSYS Fluent and tell us more about the reactions. Why are you not solving the species in all domain?

Best regards,

Amine

zypresse91 posted this 10 December 2018

Sorry for the late respond. Connection was lost. Because it’s the purpose of this study. This project will be reflecting a real geological and geophysical earth model simulation for fluid dynamics and interactions of it through 3-layered marine sediment. It’s a real place where indeed a chemical reaction takes place just in the middle body, creating products and reaching out the uppermost body. I know it’s a bit complicated, but earth system itself is complicated in the first place. Thus, we have to simulate its design as it is. Thanks for the discussion. Cheers, Selvi

Cheers, Selvi

abenhadj posted this 11 December 2018

Please get in touch (you or your prof) with the channel partner from Turkey.

Best regards,

Amine

  • Liked by
  • zypresse91
zypresse91 posted this 11 December 2018

Sir, we had a session w/ my professor and I will get back at you soon. Cheers, Selvi

Cheers, Selvi

zypresse91 posted this 19 February 2019

let me wrap it up:

segmentation fault error that I kept receive while I was trying to solve my problem was due to failing assignment of contact regions in mesh in the process of  transfering the data to the setup section of ANSYS Fluent. I don't know how and why at sometimes these contact regions are defined well, sometimes not at all. So once I watched these tutorials, I noticed the failure:

Following instructions from these tutorials(especially the first one) I managed to solve the error. I didn't need to define any other boundary layers between those packed beds including porous jump boundary. I don't have great knowledge about the implementations of ANSYS Fluent(keep learning) but once I tried "ADD FROZEN" method in the first link I shared, I just got rid of the error. 

So, problem is solved. Thanks for the discussion.

Cheers, Selvi

Close