Drafting Mass, Inertia, CG automatically

[ingallspw]

Does anyone know of a way to automatically calculate the Mass, Inertia, and a Center of Gravity and add a call-out onto an assembly print that automatically updates each time the part is changed?

On a side note...

Is there a true way to draft the center of gravity? I've only seen it on one other print and I doubt the method I saw is a part of any ANSI spec (or equivalent).

Thanks

Keegan

 

[JohnRBaker]

Have you changed you default...

File -> Options -> Save Options...

...so that the 'Generate Weight Data' is toggled ON for all your parts? If this is not toggled ON when you last saved your part files, none of the Weight Data will have been saved and therefore it will need to be recalculated whenever it's requested.

John R. Baker, P.E.
Product 'Evangelist'
Product Design Solutions
Siemens PLM Software Inc.
Industry Sector
Cypress, CA

http://www.siemens.com/plm

http://www.plmworld.org/museum/

To an Engineer, the glass is twice as big as it needs to be.

 

[ingallspw]

That's good to know but on load it only takes about a minute extra (and I am happy to know there is a way around even that little bit of time).

But... that's not what causes the hang up.


It's upon APPLY or OK. Here are the steps to get me to that hang:



Analysis > Measure Bodies

I select the bodies I want to measure. (Associative is already checked)

After a minute, the information window pops up with all the calculations.

I select OK on the Measure Bodies dialog box.

Now I wait 20 minutes.

After 20 minutes, it is in my Part Navigator. As long as if I never "edit parameters", that is the only time I see the delay.


Keegan

 

[JohnRBaker]

When you say a 'fairly large assembly', how large is large? Are we talking hundreds or thousands of Components?

John R. Baker, P.E.
Product 'Evangelist'
Product Design Solutions
Siemens PLM Software Inc.
Industry Sector
Cypress, CA

http://www.siemens.com/plm

http://www.plmworld.org/museum/

To an Engineer, the glass is twice as big as it needs to be.

 

[hudson888]

The system really ought to look at dealing with weight data for assemblies somehow more efficiently. The saved weight data works only for components with solids in them, assemblies are assigned no mass and that to me seems to be a shortcoming for any of us who display mass on parts lists for example.

Best Regards

Hudson

www.jamb.com.au

Nil Desperandum illegitimi non carborundum

 

[ingallspw]

John,

In the model that is hanging at the moment, there are 67 parts.

34 of the parts are simple extrusions (such as solid rivets and a sheet metal plate in the center).

The other 33 solid parts are actually two part files, arrayed, with lots of complex surfaces.

________________________________________________________


As for Hudson's suggestion, I know for us end users its always easy to say it should work better but in this case I'd agree. What I'd suggest is that the mass properties feature allow you to determine what is going to be calculated.

eg: In this instance I do not need, Surface area, Volume, Radius of Gyration, Products of Inertia, and Radii of Gyration.

So to speed things up, give us the option of calculating these results. If we don't need them, don't take the time to process them.

And if I am understanding Hudson, he just needs the weight. If I get the mass properties of the individual component, NX calcualtes these results Speedy Gonzales like. Now take it over to an assembly and understandably it takes a while. But if all he wants is the weight (nothing else) then it's just a simple equation for the solver to create.

I know, dang pesky end users and their suggestions!

For anyone reading this, that is new to NX , its a great CAD system and this website is full of reasons why! So don't take this as all complaints... just trying to learn as I go and hopefully it will get even better!

Thanks,

Keegan

 

[cowski]

to speed things up, give us the option of calculating these results

This initially sounds like a good idea, but due to the nature of the calculations - won't give much of a speedup. The time consuming part is the volume calculation: NX divides your model into small cubes and adds up the volumes; it splits the model into smaller and smaller cubes until your specified tolerance is met. The more 'organic' your part is the longer this process takes; if you model large boxes then it is lightning fast. Weight is then calculated as volume * density. The other parameters are simple 'one time' (no iterations required) calculations that depend on weight and/or volume.

I'm not sure how the surface area is calculated, if it is an iteration (similar to the volume) then you might see an improvement if there is an option to turn this off. But NX may use the same cubes to estimate surface area that it used for the volume calculation, so surface area might 'come along for the ride' with the volume calculation. The rest of the calculations you basically get for free once the volume is calculated.

 

[JohnRBaker]

If this is only happening with one assembly, I would suspect that there is a problem with the file (or one of more of the component parts) itself rather than with the code which is adding the Mass Property (Measurement) Expressions to the assembly model.

I talked to our developer responsible for this code and he states that you should not be surprised that it actually takes a bit longer to create the Measurement Expressions, after you hit the OK button, than it took to perform the initial mass property calculation when you first started to select objects. The reason for that is because, despite what might be logical, we first RERUN the calculations before saving the results (I know, this seems like a waste of time because we already displayed the results temporarily and technically they should still be good 2 seconds later when you hit OK, but that is just the way the code works, however this is a very small part of the time it takes to complete the task so it's not really a problem).

The extra time that it DOES take has mostly to do with the fact that we need to create what we call an 'object record' in the Assembly file which the Measurement Expressions are actually assigned to since for them to be associative and automatically update if changes are made, they have to be part of a smart object or feature (even if you can't see or 'touch' this feature, it still needs to be created).

Note that we don't create this feature when the initial calculation is done for a couple of reasons. A) you might change your mind and add or remove selected objects and since we don't wait for you to do anything before displaying an initial result, we wanted to keep this working as fast as possible. B), you may decided to NOT save the results permanently (you've toggled OFF the 'Associate' option) so why waste any time, after displaying the temporary results, creating a feature record which we are just going to have to throw away anyway. Therefore, we wait until you hit OK and then check to see if the Associate option was toggled ON and if it was, we then recompute the information, create the Object Record, assign the Measurement Expressions to it. add them to the list of Expressions in the Assembly file, thus finishing the task. Now because we are actually creating something permanent, we have to do some additional checking to make sure that we have all of the data needed to complete the task (which might require some additional time if for example the components had only been partially loaded or there were some Interpart links that needed to be checked) so it does take some additional time, above and beyond the actual mass property calculations, which as you already seen, is very fast indeed.

Now what is the typical time you should expect?

I did some testing with a couple of average sized assemblies. I had a couple in the 150 component range and for both of them, with the worst case scenario, it took about 2 seconds (these tests were done on my 3 year old Dell M65 laptop) for the initial computation of the mass properties and about 15 seconds, after I hit OK, to actually process and save the results as Measurement Expressions in the Assembly. I then tested this using an assembly with about 370 components, where it took perhaps 3 seconds to compute the data and 27 seconds to actually save it and complete the operation.

So, if your assembly only had 67 components and it's taking you 20 minutes to complete the task, I suspect that we have problems somewhere else.

As to what you can do, try doing a Part Cleanup on all of the components and re-save them and see if that helps. If it does not, then you may need to send this assembly file to GTAC so that our developers can look at it as this should NOT be taking this long no matter how complex your parts may be, something else is slowing this process down, but we would have to see the parts to determine what the real cause is as I suspect that what you're seeing is only the symptom of something that may not be related to Mass Property Calculation at all.

Anyway, I hope this helps.

John R. Baker, P.E.
Product 'Evangelist'
Product Design Solutions
Siemens PLM Software Inc.
Industry Sector
Cypress, CA

http://www.siemens.com/plm

http://www.plmworld.org/museum/

To an Engineer, the glass is twice as big as it needs to be.

 

[ingallspw]

Wow, Thanks Jon! I appreciate all the work and effort you put into this!

(For all of you reading this I challenge you to get this kind of customer interaction from PTC/Pro-E even calling them directly!)

I'll have to call GTAC. I just tried creating a new assembly using very similar parts, on a different PC, and I still have the same results. The components that I am running the process on are very "organic" as Cowski called it so maybe that has something to do with it.

And to reply directly to Cowski, I still would like to see an option to Not calculate volume if it is not necessary to calculate the rest of the values but your explanation was awesome and explains a lot! Thanks!!

And I'll take one last stab... any ideas on how to get the Inertia to calculate automatically as a value? Or do I have to wait for NX7 to do that?!

THANKS TO ALL!

 

[JohnRBaker]

Inertia would have to be derived (computed) using the values provided by the system when you perform the Measurement.

And there are no changes in what is computed nor in how the Measurement function works in NX 7.0 and NX 7.5.

John R. Baker, P.E.
Product 'Evangelist'
Product Design Solutions
Siemens PLM Software Inc.
Industry Sector
Cypress, CA

http://www.siemens.com/plm

http://www.plmworld.org/museum/

To an Engineer, the glass is twice as big as it needs to be.

 

[ingallspw]

As an FYI to everyone...

I mentioned it taking 20 minutes to calculate the Mass properties. Potrero also reported this issue. Recently I loaded up NX7. It also has the same issue.


However, this is the response I just received from GTAC:

“It has been determined that the performance problem is due to eleven (11) calls of the Parasolid function MASSPR due to the associativities of the Mass Properties feature”.

The fix will be available in NX605 which is ready for the customer around March 2010.

 

[hudson888]

One supposes that it may be in NX-7.01 whenever that comes out as well.

Best Regards

Hudson

www.jamb.com.au

Nil Desperandum illegitimi non carborundum