Material erosion with particle size variation

[torbern.klason]

Hello, 

I have a case where we have a particle size distribution of particles, with the density of 1550 kg/m3, see the table below.

Size (µm)	Number of particles	Type
1000	8	Hard/Sand
650	9	Hard/Sand
200	953	Hard/Sand
100	3749	Hard/Sand
50	13334	Hard/Sand
15	76317	Hard/Sand
5	202150	Hard/Sand

Is there any way I can simulate this (i.e. can I put in the whole particle distribution 1000 µm --> 5µm and the correct numer of particles)? 

I notice in the "Trace" menu that it seems like I can only do I particle size per calculation (is this correctly understood)?

But how do I specify for example 953 particles of a size of 200µm?

As I view the result (in this case the 200 µm, where the streamlines/traces are colored by velocity), how can I see the effect of erosion? 

If i make an iso-surface contour plot (where the blue surface is 1e-5 g/mm2/s) then I have no idea how see the size of the iso-surface. How can I get the area of the iso-surface? Else I cannot calculate how long time it takes for the erosion in this area to "eat its way through the 0.5 mm wall"? Is there away to get forward the area?

In my example that I used for calculation is the Autodesk standard EGR-valve with an inlet velocity of 7.5 m/s. 

The geometry can be found here.... https://knowledge.autodesk.com/support/cfd/learn-explore/caas/CloudHelp/cloudhelp/2019/ENU/SimCFD-Tu...

I hope anyone can share some information on what I shall do since I am working with Heat-exchanger (much more complex geometry) where there seem to be erosion in the tubes and it has the same size distribution table as above and the material need to last 40.000 hours.... I thought it would be easier to use a simpler geometry first and then try on the more complex one

Thank you

Torbern 

[KubliJ]

Hi @torbern.klason 

I am going to have to make an assumption from your description, and that assumption is that the number of particles are in units of rate. Such as 8 particles per second. If not, you are going to want you data in terms of particles per second.

CFD isn't going to let you enter in a table of different particle sizes, so we are going to have to do this based on one particle size at a time and combine the results. Also something to keep in mind is that the results are not going to be combined within CFD, this will need to be done externally.

Each trace is going to be considered one particle per unit of time (or basically one particle per second). So, assuming your particle count in the table is the number of particles per second....

1. You only need one flow analysis.
2. Add 8 traces evenly distributed (or biased however you see fit).
3. With the traces, enter in the density and radius of the largest particles.
4. Use the flag write_bounce_data to get CFD to output the erosion results for this particle size to a CSV xyz coordinate file.
5. Modify the traces for the next data set. Which is 9 particles.
6. And repeat. At some point it is going to get too hard to add the number of traces required to represent every particle. Your smallest particle has over 200 thousand particles in the flow. For that, you can use the flag Default_trace_multiplier
   1. This is going to multiple the number of traces by a scalar value (a multiplier) to virtually increase the number of particles.
   2. For example, you are simulating the smallest particle. You have 10 traces, so you would want to put in a multiplier of 20000 to get your 200,000 particles per second.
7. You can then sum up all the XYZ data and look at the combined erosion rates. You will have to use the coordinates to work backwards in probably the CAD model to best see where the max erosion is.

These steps are not exact unfortunately.... I came up with them in discussing this issue with another coworker. So it may not work as well as we hope. Or there might be a better way, or we missed some additional steps needed.

For visualizing the results per particle size. Surface results are going to be better, not traces or ISO surfaces. You will likely need to hide the fluids and use Surface Blanking to peer into the model.

Let me know how it works out.

Thanks,

James

[torbern.klason]

Hello James

Thank you for your help! It is much appreciated. 

I just have a couple of questions:

You wrote:"Use the flag write_bounce_data to get CFD to output the erosion results for this particle size to a CSV xyz coordinate file." uhm, I activated the flag write_bounce_data, rerun the traces but I can't find where it writes the csv-file. Where is it saved?

For visualizing the results per particle size. Surface results are going to be better, not traces or ISO surfaces. You will likely need to hide the fluids and use Surface Blanking to peer into the model.

• Where in the viewer do I find surface blanking?
  • I suppose that it is placed on the VIEW-tab but I just can't find it 🤔. I guess it is right in front of my noose but can you send a screen dump where I find it 😁

Then I ran into a problem....I made one flow simulation (just as you described) and then I decided to make a "erosion case" for each particle size (i.e. 1000, 650, 200 ... 5 µm) and run TRACE for each case (and I made a summary image for each case). It worked out fin until I decided to  back to a previous case (I jump from 200 µm to 650 µm). Then I get this error Assertion Failed and the program crashes. It also crashes each time I restart the program.... I assume I must have done something wrong.... 🤔🙄.... the crash is reported as error report 318789273 & 318789466.

Thanks for the help

Torbern🙂

[KubliJ]

Hi @torbern.klason 

I would imagine the write_bounce_data flag would have written in the Scenario folder. It may be in the Solver folder of the scenario. I've personally never used it. So not 100% sure. But most outputs go into one of those two folders. If you browse to your model folder, you will find subfolders that have the same Design and Scenario names. In there you will find file, or in the solver folder for each design/solver folder.

The surface blanking is on the view tab. It is something you turn on and off. And it changes the behavior of the middle mouse button. Instead of hiding the bodies, it will hide the surface under the mouse. It can be found here:

For the assertion error. I'm currently in the process of documenting a problem with version 2021 and mass particle traces that causes model file corruption. The model cannot be reopen and has to be recovered from the CFZ file.

Are you on version 2021?

Thanks,

James