Co-relation of Moldflow results & actual molding machine results.

[Mandar]

Good Morning all,
Can somebody please share views on, how to co-relate moldflow results with injection molding results like
-Injection Pressure & time
-Cavity pressure
-Cooling time & temperature
-Injection time,Holding pressure, Holding time,& packing pressure & packing time
-Weld lines
-Sink marks

Thanx & regards,
Mandar

[Suhas]

Hello Mandar,

This has always been an issue. Most people learn from their experiences and get better and better in modelling. There are also several companies, each with their own expertise - MoldFlow, SigmaSoft, Moldex3D and so on. None are perfect.

People also have developed their own ways and have kept it a secret. There are a few people who know how to do this well but they won't tell you their secrets.

Sorry!
Suhas

[Mandar]

GM Suhas sir,
Thank you so much for your view.

Have a good day

Regards,
Mandar.

[Valentin Leung]

I am not really against sharing my "secrets", if we have to call these this way, but I think that in order to have any feedback on this touchy topic, I need to give in first. So, I'll feed in the thread and see what you guys think. At the end, it's like CAD, it's not because we know how the tools work that it makes you a better designer.

We use SimpoeMold (the one behind the plastics package for Solidworks). So far, I found it useful only for the filling phase in order to find weld lines, air trap etc.. There are a lot of things I don't see with the software while I can see them at the machine. There are a lot of things I don't understand or am not sure if I do correctly.

1. Fill profile and screw inertia... (for this section, consider that there is no PACK/HOLD, 0 speed, 0 bar, 0 time, I just fill the FILL page of the machine/software )
VP switch... That's probably our biggest headache. We don't know what to do with it on the software side to represent what happen at the machine. Inside Simpoe, I would set my filling profile as I do with the machine's UI. Aim for speed, so for example, I set a starting point, a position and a speed. I have a drop down box where I can set the exact % value at which we should perform VP. That's it... . I often set it at 100% of volume filled. We tried many times in the past to set it at 98%/95%/90%... etc... but every time, it ends up saying that the computed flow reach a value that is 1/1000 of the original flow, so there is short shot happening. So, we just ended up using 100%.
Now comes the issue, on the machine, anybody who performed a viscosity curve knows that the screw will never stop at the VP point when we are at parameters on the right side of the viscosity curve, the screw will push a little bit further because of its inertia. On my machine I can see that the reported cushion value is closer to VP and different to the latest position reached. For example, VP is 15mm, cushion is 10, but I can see that the actual position of the screw is 5, then it would bounced back to 10.
Because of this, when my designer try to set up an early simulation and ask me for VP switch... I don't know what to tell them, from the trial reports, they see that from the total screw stroke, they can calculate a VP % switch... and then, don't understand when I tell them that it's meaningless in their software.

Also, Fill profile is different, if I set speed 100mm/s, the software's screw is going to reach that speed within 5mm of stroke... while on my machine, I need more, like 20-30, some product I run have such a short stroke that I don't even reach my set speed within the time the part is filled.

At the end, I was asked by some Sigmasoft engineers: "does it really matter to know exactly what happens at the start? the most important thing is how you interpret the results."

So, in the setup of the FILL side... does it really matters? Am I doing something wrong?
At the end, I think the Sigmasoft engineers are right... my worries are end of fills, weld lines, sinks etc... air traps (that's the worse thing). I need to be able to see where they are.

2. single cavity/multi cavities.
That's my second issue, we usually do pilot molds. What speed should we run them? Do I plan to run at a speed that is equivalent to the production mold I am targetting to do? So, for example, if I plan a production mold, 16 cavities on a 150mm/s machine, should I just do 150/16, that's the speed of the pilot mold?
This also translates completely in the software world, I never have hot runner CAD and my engineers don't spend their time meshing 16 cavities, they'll do simulation on one single cavity and that's all. What do you guys think?
I had an issue on a part lately (to illustrate that my design engineers and the process engineers don't talk to each other), a part has been designed, production mold done. At some point, we saw an air trap in the middle of the part, the designer was surprised, saying that his simulation did not show that, after we dug into the problem, he was using filling speed from the production tool (6 cavities) to simulate a single cavity fill. When we dropped the speed to the equivalent speed to fill a single cavity, things were much closer to reality. So, for now, I set some upper limits of speed for their simulation (machine max/nb of cavities they plan for mold making).
The problem is that for high cavitation... I can't do it, I fall into the extreme left side of the viscosity curve and I am screwed... having to use a VERY big pressure to overcome viscosity... I think I should dig that viscosity curve and try to find an equivalent shear and not just a plain speed portion. Like I need to drop the speed enough so that shear on a single cavity is similar to the level seen on a multi cavity.

Thanks for your time.

[Valentin Leung]

That's two points I wanted to share first, let me know your thoughts and comments. It's a bit of a mix between actual machine and pilot mold testing, but that's really what we are after, flow simulation early in the development stage..

No engineer is perfect, we all learn on mistakes and success.

Sorry I had that at the end of my post, but this forum limits the messages to 5000 characters...