Define Material Dimension (particularly the length) in CFX

yusofharun · January 24, 2015, 02:45

Greetings to all, bowing to the Super-Mod, Mod, Senior

One of the fluid I'm trying to simulate is a fiber. While the diameter could be set in Particle Mean Diameter box, how can I specify the length of the fiber? Better yet, fibre with multiple distribution of length. I have the length distribution data (e.g. 5mm 20%, >5mm 80%).

Thanks in advance

ghorrocks · January 26, 2015, 18:29

I have never been bowed to before.

CFX does not have a fibre model, let alone one with varying diameter over the length.

What are you trying to model? What does the fibre do in the flow?

yusofharun · January 27, 2015, 23:36

Upon entering the dojo, one must bow to the master

Thank you for the attention

I'm trying to simulate the process of oil extraction using extruder. I want to check the mass distribution of the process components (oil, fibre, etc.) in relation to the extruder's geometry and movement.

In actual condition, the oil is residing within the fibre and the compression force induced by the extruder's screw flight movement will squeeze out the oil as the fibre began to re-arrange and eventually become compacted. To simplify the simulation, I just assume that the process fluid is a heterogeneous mixture; with the volume fraction of each component specified.

I encountered a bump in the road; the fibre strand has the shape of a long thin cylinder (with the assumption of constant diameter) and they are of various length (e.g. 5mm 20%, >5mm 80%). This particular component needs to be represented properly in the simulation because it makes up the majority of the process fluid.

ghorrocks · January 28, 2015, 01:21

What are you trying do this model for? Are you trying to optimise the oil extraction process? Or reduce energy consumption?

The reason I ask is this can be modelled on many different levels. You could model this as a fluid in a porous medium, you could use either Eularian or Langrangian particles, or you could model the fibres directly with a rigid body or even a FSI model.

But I cannot say which form of model is most appropriate until I understand what you are trying to achieve with the model. Obviously you want to use the simplest possible model which captures the important physics in your application.

yusofharun · January 28, 2015, 02:32

I want to optimise the oil extraction
(in the context of separating the oil from the fibre, basically by checking the individual distribution of the fluid components (oil, fibre, etc.) in relation to the extruder's geometry and movement.)

Here are some descriptions of the process:

The flow domain (extruder) consists of a perforated barrel and two screw flights. The available area of for fluid flow (cavity) would be the volume of the barrel minus the volume of the screw flights.

The process fluid (oil, fiber, etc.) would enter from one side of the barrel and eventually exits at the other end. The fluid is expected to occupy the extruder's cavity and will be brought forward to the end of the extruder as the screw flights turn.

Along the way, majority of the oil will be drained through hundreds of small perforations on the barrel (3mm in diameter, thus not allowing fibre and other solids to pass through). As for fibre and other solids, they are expected to be transported to the extruder's end (remnants of oil would be there as well).

ghorrocks · January 28, 2015, 07:09

What design things do you want to modify to optimise the oil extraction?

Regardless, here are some comments. They may or may not be relevant as I do not understand what you are trying to do yet:
* The simplest model of this would appear to be a porous flow model where the fibres are assumed to be the porous material. But this ignores a lot of detailed physics so may not be useful - but it might get the oil flow.
* You might be able to use a fluidised bed type of Eularian multiphase model. You will need to modify it to generate the oil from the fibres but this could be done with source terms. The fluidised bed model includes maximum packing fraction on the solid phase which will be useful.
* Modelling this with lagrangian particles will be tricky. The CFX lagrangian model does not work when particles pack together.
* If you are interested in what happens to single fibres then a rigid body or FSI model may be of interest
* If you want to do a lagrangian particle but include packing then I would consider coupling CFX to a discrete element model software like EDEM. EDEM has more sophisticated particle models to handle shapes like fibres and packed particles.

yusofharun · January 29, 2015, 10:53

Among the variables that can be manipulated in the optimisation study is the screw flight rotational speed, clearance between the two screw flights, clearance between the flights and the barrel, screw flight's pitch + root diameter, etc...as for the process fluid, not much can be manipulated except maybe the feed rate.

Thanks for the suggestions. I really appreciate your help.
I've been looking into EDEM and it looks very promising. Unfortunately, my institution does not subscribe to this software

(A weeping emoticon would be more appropriate here).I need to consider other options.

Thanks again~

ghorrocks · January 29, 2015, 18:09

You say the oil is released as the fibre is compressed. Is the compression due to pressure in the fluid? Or is it mechanically squeezed by the screw squashing it against something? Or is it shear stresses in the fluid?

yusofharun · March 11, 2015, 13:19

Yes, the oil is released as the fibre is compressed. This is due to
(a) the compaction of the process mixture due to the movement of screw flight
(this happened during the initial stage of the process)
(b) mechanical squeeze - as the compacted process mixture move towards the end of the cage, they will be squeezed against the cage end. The cage end is movable yet retained by a hydraulic module to counter the force exerted by the upcoming material. After a period of time, the hydraulic module will retract the cage end to discard the exhausted cake
(c) gravity
(d) viscosity

ghorrocks · March 11, 2015, 18:30

It has been over a month since your previous posts on this thread. Have you done some simulations since then? Have you made any progress? Is anything working or not working?

yusofharun · March 16, 2015, 00:35

Still looking around for a feasible solution since most of my attempts reach a dead end =(...need to work harder...

ghorrocks · March 16, 2015, 02:14

It is a very complex model you are attempting with several physical processes in it which will require development of a custom model in CFX. This model was always going to take a lot of work to get useful. I suspect you should consider simplifying it a bit. For instance, is the oil extraction a function of something (pressure, shear strain rate)? If so you could model the stuff as a fluid and infer the oil extraction from the press/SSTR.

I should also mention that one of the most common mistakes made by new CFD people is to jump straight into the "model everything" simulation. In my experience they are so complex and the interactions are not known enough - resulting in a model which is never accurate and never works. A far better approach is to start trivially easy - maybe a simple single phase fluid in your geometry and get that working well - then add the complexity one step at a time, checking that each new layer of complexity is working before you add the next one.

And my final point about this approach is that 95% of the time as you work through this, you find something important on the simple model which explains something you never knew before. You then spend the rest of the project optimising stuff on the simple model and never end up needing the complex model with everything in it anyway. This has happened to me too many times to count.

yusofharun · March 16, 2015, 03:05

You're absolutely right, there's no shortcut to master CFD. When I first began, my superiors warned me that this area has a lot of pitfalls especially with the advancement in commercial CFD packages. Those fancy buttons and drop-down lists...

From your message, I'm glad to know that I'm on a right track. I did kick-off my project with a very simple setup, similar to what you have described. As of now, I'm entering the so-called complex phase; attempting to refine the simulation.

I started this journey alone and along the unpaved, rocky, dusty road, I met people like you who are more than willing to lend a hand in helping me to get through. I'm very grateful...

“If I have seen further it is by standing on ye sholders of Giants”
Sir Isaac Newton (1643-1727, English physicist and mathematician), in his letter to Robert Hooke on 15 February 1676.

January 24, 2015, 02:45	Define Material Dimension (particularly the length) in CFX	#1
yusofharun New Member Selangor Join Date: Sep 2013 Posts: 10 Rep Power: 13	Greetings to all, bowing to the Super-Mod, Mod, Senior One of the fluid I'm trying to simulate is a fiber. While the diameter could be set in Particle Mean Diameter box, how can I specify the length of the fiber? Better yet, fibre with multiple distribution of length. I have the length distribution data (e.g. 5mm 20%, >5mm 80%). Thanks in advance

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Hydrogen storage by metal hydride	longbk	FLUENT	12	August 1, 2023 21:13
[stressAnalysis] how i can define the material type?	e-nikookar	OpenFOAM CC Toolkits for Fluid-Structure Interaction	3	January 2, 2015 01:01
Can CFX use a material from Fluent's library?	JuPa	CFX	2	June 30, 2014 09:46
how to define a material of gas	kiwishall	CFX	4	May 8, 2013 02:03
Adding material to CFX material library	shiva3689	CFX	1	March 8, 2013 10:16

January 26, 2015, 18:29		#2
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,872 Rep Power: 144	I have never been bowed to before. CFX does not have a fibre model, let alone one with varying diameter over the length. What are you trying to model? What does the fibre do in the flow?

January 27, 2015, 23:36		#3
yusofharun New Member Selangor Join Date: Sep 2013 Posts: 10 Rep Power: 13	Upon entering the dojo, one must bow to the master Thank you for the attention I'm trying to simulate the process of oil extraction using extruder. I want to check the mass distribution of the process components (oil, fibre, etc.) in relation to the extruder's geometry and movement. In actual condition, the oil is residing within the fibre and the compression force induced by the extruder's screw flight movement will squeeze out the oil as the fibre began to re-arrange and eventually become compacted. To simplify the simulation, I just assume that the process fluid is a heterogeneous mixture; with the volume fraction of each component specified. I encountered a bump in the road; the fibre strand has the shape of a long thin cylinder (with the assumption of constant diameter) and they are of various length (e.g. 5mm 20%, >5mm 80%). This particular component needs to be represented properly in the simulation because it makes up the majority of the process fluid.

January 28, 2015, 01:21		#4
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,872 Rep Power: 144	What are you trying do this model for? Are you trying to optimise the oil extraction process? Or reduce energy consumption? The reason I ask is this can be modelled on many different levels. You could model this as a fluid in a porous medium, you could use either Eularian or Langrangian particles, or you could model the fibres directly with a rigid body or even a FSI model. But I cannot say which form of model is most appropriate until I understand what you are trying to achieve with the model. Obviously you want to use the simplest possible model which captures the important physics in your application.

January 28, 2015, 02:32		#5
yusofharun New Member Selangor Join Date: Sep 2013 Posts: 10 Rep Power: 13	I want to optimise the oil extraction (in the context of separating the oil from the fibre, basically by checking the individual distribution of the fluid components (oil, fibre, etc.) in relation to the extruder's geometry and movement.) Here are some descriptions of the process: The flow domain (extruder) consists of a perforated barrel and two screw flights. The available area of for fluid flow (cavity) would be the volume of the barrel minus the volume of the screw flights. The process fluid (oil, fiber, etc.) would enter from one side of the barrel and eventually exits at the other end. The fluid is expected to occupy the extruder's cavity and will be brought forward to the end of the extruder as the screw flights turn. Along the way, majority of the oil will be drained through hundreds of small perforations on the barrel (3mm in diameter, thus not allowing fibre and other solids to pass through). As for fibre and other solids, they are expected to be transported to the extruder's end (remnants of oil would be there as well).

January 28, 2015, 07:09		#6
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,872 Rep Power: 144	What design things do you want to modify to optimise the oil extraction? Regardless, here are some comments. They may or may not be relevant as I do not understand what you are trying to do yet: * The simplest model of this would appear to be a porous flow model where the fibres are assumed to be the porous material. But this ignores a lot of detailed physics so may not be useful - but it might get the oil flow. * You might be able to use a fluidised bed type of Eularian multiphase model. You will need to modify it to generate the oil from the fibres but this could be done with source terms. The fluidised bed model includes maximum packing fraction on the solid phase which will be useful. * Modelling this with lagrangian particles will be tricky. The CFX lagrangian model does not work when particles pack together. * If you are interested in what happens to single fibres then a rigid body or FSI model may be of interest * If you want to do a lagrangian particle but include packing then I would consider coupling CFX to a discrete element model software like EDEM. EDEM has more sophisticated particle models to handle shapes like fibres and packed particles.

January 29, 2015, 10:53		#7
yusofharun New Member Selangor Join Date: Sep 2013 Posts: 10 Rep Power: 13	Among the variables that can be manipulated in the optimisation study is the screw flight rotational speed, clearance between the two screw flights, clearance between the flights and the barrel, screw flight's pitch + root diameter, etc...as for the process fluid, not much can be manipulated except maybe the feed rate. Thanks for the suggestions. I really appreciate your help. I've been looking into EDEM and it looks very promising. Unfortunately, my institution does not subscribe to this software (A weeping emoticon would be more appropriate here).I need to consider other options. Thanks again~

January 29, 2015, 18:09		#8
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,872 Rep Power: 144	You say the oil is released as the fibre is compressed. Is the compression due to pressure in the fluid? Or is it mechanically squeezed by the screw squashing it against something? Or is it shear stresses in the fluid?

March 11, 2015, 13:19		#9
yusofharun New Member Selangor Join Date: Sep 2013 Posts: 10 Rep Power: 13	Yes, the oil is released as the fibre is compressed. This is due to (a) the compaction of the process mixture due to the movement of screw flight (this happened during the initial stage of the process) (b) mechanical squeeze - as the compacted process mixture move towards the end of the cage, they will be squeezed against the cage end. The cage end is movable yet retained by a hydraulic module to counter the force exerted by the upcoming material. After a period of time, the hydraulic module will retract the cage end to discard the exhausted cake (c) gravity (d) viscosity

March 11, 2015, 18:30		#10
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,872 Rep Power: 144	It has been over a month since your previous posts on this thread. Have you done some simulations since then? Have you made any progress? Is anything working or not working?

March 16, 2015, 00:35		#11
yusofharun New Member Selangor Join Date: Sep 2013 Posts: 10 Rep Power: 13	Still looking around for a feasible solution since most of my attempts reach a dead end =(...need to work harder...

March 16, 2015, 02:14		#12
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,872 Rep Power: 144	It is a very complex model you are attempting with several physical processes in it which will require development of a custom model in CFX. This model was always going to take a lot of work to get useful. I suspect you should consider simplifying it a bit. For instance, is the oil extraction a function of something (pressure, shear strain rate)? If so you could model the stuff as a fluid and infer the oil extraction from the press/SSTR. I should also mention that one of the most common mistakes made by new CFD people is to jump straight into the "model everything" simulation. In my experience they are so complex and the interactions are not known enough - resulting in a model which is never accurate and never works. A far better approach is to start trivially easy - maybe a simple single phase fluid in your geometry and get that working well - then add the complexity one step at a time, checking that each new layer of complexity is working before you add the next one. And my final point about this approach is that 95% of the time as you work through this, you find something important on the simple model which explains something you never knew before. You then spend the rest of the project optimising stuff on the simple model and never end up needing the complex model with everything in it anyway. This has happened to me too many times to count.

March 16, 2015, 03:05		#13
yusofharun New Member Selangor Join Date: Sep 2013 Posts: 10 Rep Power: 13	You're absolutely right, there's no shortcut to master CFD. When I first began, my superiors warned me that this area has a lot of pitfalls especially with the advancement in commercial CFD packages. Those fancy buttons and drop-down lists... From your message, I'm glad to know that I'm on a right track. I did kick-off my project with a very simple setup, similar to what you have described. As of now, I'm entering the so-called complex phase; attempting to refine the simulation. I started this journey alone and along the unpaved, rocky, dusty road, I met people like you who are more than willing to lend a hand in helping me to get through. I'm very grateful... “If I have seen further it is by standing on ye sholders of Giants” Sir Isaac Newton (1643-1727, English physicist and mathematician), in his letter to Robert Hooke on 15 February 1676.