Moldflow Insight Forum
Welcome to Autodesk’s Moldflow Insight Forums. Share your knowledge, ask questions, and explore popular Moldflow Insight topics.
cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 

Tips and tricks to improve the pressure predictions.

16 REPLIES 16
Reply
Message 1 of 17
PascalGosset
3735 Views, 16 Replies

Tips and tricks to improve the pressure predictions.

For several weeks, I have been comparing the pressures predictions of Moldflow to the real ones.

To do this, I take into account all the information I can :

checked parts thicknesses, hot or cold runners, press nozzle, machine settings, injection speed profile ….

And then, I superimpose the machine plot pressure / time over the calculated one.

 

Doing this, I regularly found underestimations around 50%.

Mainly, because the lost of pressure in the runners is much more important in reality than in the simulations.

 

I join to this post a picture with 4 examples.

The pressure underestimation is important for 3 of them.

Until now, I got a pressure plot close to the real one only with the 4th part.

 

So my question is now : do you have tips and tricks to improve the filling pressure predictions ?

 

I submitted this problem to the Moldflow support.

On my side, I tested some changes to correct those errors.

I hope I’ll be able to give you solutions in few days.

 

Last week, I posted a question about this problem.

I only got 1 reply. That surprised me.

So, I prepared a the little survey and I’d like you to fill it.

Normally, it’ll take you less than 1 mn.

Please, follow this link :

http://www.surveymonkey.com/s/2SNHYTD

 


Thanks for your help.

Signature: "Maybe Moldflow does not work properly, but the real world neither" my son...6 years old 😉

16 REPLIES 16
Message 2 of 17
tim.vanast
in reply to: PascalGosset

Hello Pascal,

 

You have a good topic here that has a number of possibilities for the error.  It looks like you've looked at most of them. 

 

One thing I've found for better pressure accuracy is to model the runners as 3D instead of beam.

 

Another thing to consider, as you mentioned, is the material data.  At a previous job, I had Moldflow do testing on many materials.  Their test report has a validation section at the end and many times the validation showed that the pressures did not match very well.  And the report show the validation at a very low pressure compared to actual injection molding.  I don't have a solution to this problem since I know that they use the state-of-the-art technology to do the testing.  But this could be the source of much of the under prediction.  You could ask your material supplier to see the information (if they have it readily available, and are willing to share it).  So this variation would be different from material to material.

 

Many times I've found that even at 50% difference, that it doesn't change the simulations that I run or change the recommendations that I make.  Maybe that's why we collectively don't have such big issues with the pressure inaccuracies.

 

Best regards,


Tim VanAst

Message 3 of 17
PascalGosset
in reply to: tim.vanast

Thanks for your reply Tim.

 

I plan to try 3D modelling for the runners later.

 

In fact, I found a trick to improve the pressures predictions with beam runners.

Take a look at the picture joined to this message.

It shows the results for the 3 parts I took as examples before.

 

For the moment, I don't tell you how I did.

I just can say you it is really simple.

I'm waiting answers from Moldflow about the mails I sent them.

And, I'm testing this solution on other parts, with others materials.

 

 

Signature: "Maybe Moldflow does not work properly, but the real world neither" my son...6 years old 😉

Message 4 of 17
tim.vanast
in reply to: PascalGosset

Wow, the results are shockingly close.  Great work.  Share it when you can.

 

Tim

Message 5 of 17
nishit78
in reply to: tim.vanast

Pascal,

 

Very impressive work. I too spend some amount of time trying to compare moldflow with reality, when possible. However, that kind of real data is not always available to me. And, while i consider myself to be a well-organized person, i am not as organized as you appear to be with this validation stuff.

 

My short shot comparisons are extremely spot on.

As i said in other post, my pressure predictions are usually below reality but i rarely model the hot runner due to the lack of design data and the preliminary nature of the work. For a cold runner mold i go as far back as the sprue when design data is available.

 

I hope to see more results for you and looking forward to shared tips/tricks for various other aspects of moldflow.

 

Nish.

Message 6 of 17
PascalGosset
in reply to: nishit78

I think, I understood from where the problem comes.

 

It happens with all kinds of meshings, with all materials, even doing the best you can to set good entries.

 

I'm talking with Moldflow about this enormous error.

 

I'll give you an explaination after.

 

I just received the results got with another rheologic simulator. Visibly, it makes the same error.

Signature: "Maybe Moldflow does not work properly, but the real world neither" my son...6 years old 😉

Message 7 of 17
PascalGosset
in reply to: PascalGosset

Please, take 1 minute to fill the 10 questions of my Survey.


It’s only to know how much important is this problem for you.


Follow this link : http://www.surveymonkey.com/s/2SNHYTD

 

If you cannot access this site, let me know.


Thanks.

Signature: "Maybe Moldflow does not work properly, but the real world neither" my son...6 years old 😉

Message 8 of 17
efoltz
in reply to: PascalGosset

Wow, that is really impressive. I agree this is a major issue in Moldflow. Please share when you can.

 

Thanks

Erik

Message 9 of 17
piepCAE
in reply to: PascalGosset

Hello,

 

The pressure prediction underestimation is in fact a very important issue that needs improvement, I'm looking forward to know about your tricks. I personally don't have the time or availability to do any extensive testing and comparison, so I'll have to rely on your work! Congratulations for you iniciative.

 

Regards,

Carlos

Message 10 of 17
sadashiv
in reply to: PascalGosset

1) very nice work

 

I use the 3 D runner system with Machine nozzle and the extension viscosity if it is measured.

 

with Beam element it is useful to add machine nozzle and some correction can also be made in Viscosity model to get accurate result.

 

Best regard.

Message 11 of 17
PascalGosset
in reply to: sadashiv

Sorry everybody…
No news from Moldflow for the moment…
Waiting them, I propose you to take a look at my survey results.
See the picture joined.

 

Reading it, I notice a lot of you take into account the runners.
Personally, I have doubts about them since a long time.
And now, I think they can lead to underestimated pressures bigger than 100 %.
(even with runners perfectly design, all meshes, all materials… )

 

For the moment, I advise you to make your simulations without the feeding.
To consider the pressure lost in the hot runners system, ask your supplier.
But, check he does not gives you values got with simulations.

 

Signature: "Maybe Moldflow does not work properly, but the real world neither" my son...6 years old 😉

Message 12 of 17
csegard
in reply to: PascalGosset

Hello M. Gosset,

 

just to present myself for others, i am working for the moldflow reseller in France and deal with the case for M. Gosset.

 

I am really surprised to see that, for you, there is no answer from Autodesk.

From my side and from Autodesk side, we give some advices on materials datas to check and on modelization to have better results on Pressure.

 

Autodesk and I will be really pleased to discuss with you directly during a meeting in your office.

 

Regards

 

Cédric

 

 

Message 13 of 17
PascalGosset
in reply to: csegard

Hello Cédric,

 

It’s true that you worked on my case.
But you did not look in the sense I was expecting for.
That is why, few days ago, I explained you clearly what I made by my side.
I told you my tracks, my clues and my conclusions.
For the moment, it seems we do not agree.
I think I have to work more on this subject.
But I hope, by your side, you are going to take my hypotheses more carefully.

 

Best regards.

Signature: "Maybe Moldflow does not work properly, but the real world neither" my son...6 years old 😉

Message 14 of 17

Sounds interesting.

looking fwd to your input

 

whereas moldflow does not include pressure losses at the screw and barrel, i wonder what were your 'tricks' to have a significant closer correlation on your models.

 

feeding system - meshing quality perhaps?

 

 

Message 15 of 17
nwvanvliet
in reply to: PascalGosset

Are the incorrect pressure plots done with a single beam element per segment?  

 

Then the correct one was done using multiple beam elements per segment, especially at the gating position?  

Message 16 of 17
PascalGosset
in reply to: nwvanvliet

I'm working on this subject since few months.

But I have not finished.

 

About my trick, I can tell you what it is. But, I don't know if it works always.

What I know is it worked for 3 parts I chose  by chance.

 

Take a look at the 1st curves in the picture 1.

The distance between them means there is an important under-estimation of the material compressibility in the runners (all : cold and hot, beam, 3D).

 

To correct this trouble, I only DOUBLE the length of ALL runners (using the tool Scale).

See the results picture 2.

 

In fact the problem comes from the compressibility factors included in the materials data.

The fields c1, c2 D3 are not filled for 80% of the database (see picture 3).

They are for materials certified "Gold".

That's why the pressure prediction is good in the 4th example of the picture 1.

 

Signature: "Maybe Moldflow does not work properly, but the real world neither" my son...6 years old 😉

Message 17 of 17
nwvanvliet
in reply to: PascalGosset

I was under the impression that Moldflow's solver always assumes D3 is zero, regardless of what was input.

 

Here's a snippet from their wiki about C1 and C2 values...

http://wikihelp.autodesk.com/Moldflow/enu/2013/Help/2.0Insight/4441-Material4441/4995-Juncture4995

Can't find what you're looking for? Ask the community or share your knowledge.

Post to forums  

Autodesk Design & Make Report