Discussion Forum

Hi

FEM will give you nice flow "theoretical" views, but still you need to define the "physics of the erosion" and to validate/calibrate your model agains measurements, hence precise cases where you know the geoemtry, the input/output flow parameters, the environment parameters, the proportion of "sand" etc.

It's not an obvious task, certainly challenging, but I do not see what to give you as arguments "why COMSOL would be much better than any other tool", apart that it is a modern tool with, a lot of expansion (users developped physics) possibilities, compared to most other tools

In anycase use the latest v4.1 version, and not the early 4.0

--
Good luck
Ivar

Dear Ivar,

Thanks for your reply. i appreciate it.

I am going to estimate and find the erosion rate in oil and gas pipe due to sand transport then compare the result with experimental data which has carried out before.

Do you think its better to use another tool and different software for example (ANSYS)?? do you think that I can do in comsol since i am a beginner in this program.

thanks for your help

Hi

I do not have enough experence in your filed to really say which one is "best". From my point of view COMSOL is as good if not better than the others, it's certainly more modern and has a unique "multiphysics" approach, that you will not, in my view find elsewhere (yet;).

Today I do not know of anything Ansys can do that COMSOl cannot do but then I do not know everything ;)

Then it's also a question of usage, previous knowledge, exchange with other clients etc. many things to consider

--
Good luck
Ivar

Hi,

Thanks again.

I heard and rad a lot about the ability of comsol in multiphysics but the question is, is it possible to estimate the erosion rates due to sand transport in oil and gas pipelines?? how would be difficult for me me as a beginner?? does it take long time ??

Thanks again.

akar

Hi

there are two different issues a) to learn how to COMSOL (or any advanced FEM tool), if you have a good engineering base or are a physicist, I would say 3-6 months to use the tool efficiently + 1-2 years to become an expert, but it depends also on how many "physics" you use, how skilled you are in math and physics + PDE/ODE (for Comsol).

I usually call ANSYS or NASTRAN or the other FEM tools "clickable Engineering tools": most operations are "pre-cooked" and you select what is available.
With COMSOL you have more "clickable differential equations (PDE/ODE) for physics" and you select Physics chapters, then you might taylor them, you need to think a bit more "physics".

Then for your case, you have to know the physics of the erosion, this is not at all clear for me, so here I cannot tell, but I doubt that it's trivial. I would myself count several months if not a year to feel comfortable ! but again your case is far from my domain

--
Good luck
Ivar

Hi Alan,

I think you should be able to use Comsol for simulating the sand erosion in the pipelines. And probably Comsol is better than ANSYS because you can simulate 'corrosion' as well in addition to the erosion and if you really push the boundaries, you can try simulating 'erosion-corrosion' in case of carbon steel pipelines.

If you are a beginner, I suggest you start working on understanding the physics behind the phenomena of the above scenarios and start putting together the differential equations to replicate them. If you can create the required equations for simulation, the next part should be easier.

To give you some hints for erosion only, what you need to consider is:

1. DNV-RP-O501 as a starter
2. Think how would you replicate/approximate the multiphase flow for the fluids which would be containing the suspended sand particles
3. Think what parts the individual particles would take whilst taking the drag forces and gravitational pull on the particles would effect. And how to simulate the non-elastic collision of the sand particles with the pipewalls.
4. What geometries would you want to simulate for erosion, is it only pipe depressions and bends or it includes any valves etc. ???

Depending on your circumstances it might take you between 1 month and 1 year to get to the bottom of a basic erosion model (working 22 days a month for 8 hours a day).

But one thing I know is that if you get the physics of modelling right, the end product of the simulation would be much better than at least ANSYS.


Good luck,


Sohail

Hi Alan,

I think you should be able to use Comsol for simulating the sand erosion in the pipelines. And probably Comsol is better than ANSYS because you can simulate 'corrosion' as well in addition to the erosion and if you really push the boundaries, you can try simulating 'erosion-corrosion' in case of carbon steel pipelines.

If you are a beginner, I suggest you start working on understanding the physics behind the phenomena of the above scenarios and start putting together the differential equations to replicate them. If you can create the required equations for simulation, the next part should be easier.

To give you some hints for erosion only, what you need to consider is:

1. DNV-RP-O501 as a starter
2. Think how would you replicate/approximate the multiphase flow for the fluids which would be containing the suspended sand particles
3. Think what parts the individual particles would take whilst taking the drag forces and gravitational pull on the particles would effect. And how to simulate the non-elastic collision of the sand particles with the pipewalls.
4. What geometries would you want to simulate for erosion, is it only pipe depressions and bends or it includes any valves etc. ???

Depending on your circumstances it might take you between 1 month and 1 year to get to the bottom of a basic erosion model (working 22 days a month for 8 hours a day).

But one thing I know is that if you get the physics of modelling right, the end product of the simulation would be much better than at least ANSYS.


Good luck,


Sohail

Hi again,

Apologies for a typo in the point 3 above, its paths and not parts that the sand particles are supposed to take.

Cheers, Sohail

Hi Alan,

I'm working on something similar where I'd have to model high velocity abrasive particles travelling over a surface and in the process, 'eroding' it little by little. I'm having some trouble modeling the scenario. So if you have found a way to model the erosion of surface due to particle collision, would you like to share it with me over here?

Thanks.

In my opinion, the best information about erosive wear in piping systems was developed by DNV (Det Norske Veritas) and the title is "Recommended¨Practice RP O501 Erosive Wear in Piping Systems.

www.dnv.com/searchresult/index.asp?query=erosive+wear