topic Re: modelcheck metrics in 3D Part & Assembly Design

modelcheck metrics

MikeLockwood — Thu, 03 May 2018 00:55:52 GMT

I've been at Edwards a little over 3 months now. Lots going on - major initiatives to "raise the bar" in many areas, including CAD quality and becoming more model-based.We're actively establishing / updating standards in various areas, and attempting to automate checking

Re: modelcheck metrics

JoeFrancz — Tue, 29 Mar 2016 20:51:16 GMT

Hi Mike,

I would be interested in the cheat sheets and diagram you put together. I am trying to get back up to speed with MC.

Thanks

Joe

Re: modelcheck metrics

BenLoosli — Wed, 30 Mar 2016 13:25:05 GMT

Mike,

I have a lot of tutorials from Pro/E 2001 timeframe on ModelCheck. We have it implemented here, but the users don't get too concerened with the resukts, from what I have seen. We do have an established limit on errors that will prevent check-in, but I haven't heard of any complaints on those.

I would like to get a copy of your cheat sheets, too.

Thanks,

Ben

Re: modelcheck metrics

JoeFrancz — Wed, 30 Mar 2016 13:52:23 GMT

Ben,

I take it you are using the "gate keeper" to limit the number of errors allowed to be checked in. My question can you prioritize what errors can go thru and what errors will be stopped.

For instance I have in my model check to make sure if there are no suppressed objects. Can you set up gate keeper to not allow any models/assemblies with suppressed items from getting checked in?

Joe

Re: modelcheck metrics

BenLoosli — Wed, 30 Mar 2016 14:05:26 GMT

I did not setup ModelCheck here, but I did in another company. From what I can remember and tell, the only 'block' is if the number if errors/warnings exceeds the threshold in the gatekeeper. There is no way to have it prevent a check-in of 1 suppressed item, that being the only failure, if your limit is set to 2 failures.

The key is setting the items to error/warn and then setting your gate keeper to an acceptable level.

I do agree that there should almost be a third code for those items that you absolutely do not want in a released model. Lacking more granularity to the failure levels, maybe a defined set of conditions that Windchill can read that must return a 'pass' before check-in will proceed.

Re: modelcheck metrics

JoshH — Fri, 10 Jun 2016 13:52:33 GMT

Hi Mike Lockwood,

I'm at a meeting with a bunch of like-minded individuals at the moment discussing ModelCheck stuff.

Have you gotten anywhere with your reports?

Re: modelcheck metrics

dschenken — Sat, 11 Jun 2016 04:23:32 GMT

Why no suppressed objects? I use them from time to time to future-proof models. It allows adding features that are sometimes useful in the next assembly/ies via flexible features without disrupting all current assemblies and without creating a confusing array of potential parts in Windchill.

Re: modelcheck metrics

RussellJ.H.Hsu — Fri, 01 Jul 2016 02:51:40 GMT

Hi Mike,

Metrics Tool has been re-built on ThinkWorx. A prototype is available currently. If you are interested in this new app, I can provide you a trial build.

Any feedback is appreciated.

Thanks,

Russell

Re: modelcheck metrics

TomU — Fri, 01 Jul 2016 11:59:25 GMT

Russell Hsu‌, you might want to repost this over in ModelCHECK Administration‌

Re: modelcheck metrics

CADQueen — Fri, 01 Jul 2016 13:53:06 GMT

Russell,

If available I'd love to see the prototype. This could be the topic of my next book!

Rosemary

Re: modelcheck metrics

MikeLockwood — Fri, 01 Jul 2016 20:16:43 GMT

Hi Russell, we would love to be able to take a look at this. What is involved in using ThingWorx?

Re: modelcheck metrics

RussellJ.H.Hsu — Mon, 04 Jul 2016 03:18:38 GMT

A video and brief configuration doc have been uploaded onto group ModelCHECK Administration. Let me know if you need further information.

Re: modelcheck metrics

MikeLockwood — Tue, 05 Jul 2016 15:56:21 GMT

Very nice!! Major question: I checked some months ago into setting up MC metrics. Seems that it requires use of a very old Apache install - which is not allowed here. What is required (in 2016) to get MC metrics collected, such that they can be summarized and presented in ThingWorx?

Re: modelcheck metrics

RussellJ.H.Hsu — Thu, 07 Jul 2016 02:02:41 GMT

The platform support depends on ThingWorx. I know Apache Tomcat 7.x works for ThingWorx. Not sure the the latest Apache is supported.

So far, the user needs to upload MC metrics manually. The Metrics tool checks whether the data is already in database or not. Metrics data will be appended to the database if it is not there.

Re: modelcheck metrics

MikeLockwood — Thu, 07 Jul 2016 14:36:50 GMT

What I found a few months back on this was that Java SDK is required to be 1.4, Tomcat needs to be 5.0 and Apache needs to be 2.0.

I would LOVE to find out how to use metrics to output automatically to a network drive on exit with modern versions.

Re: modelcheck metrics

MikeLockwood — Thu, 07 Jul 2016 14:38:10 GMT

Attached are some rough notes I made on this in March this year after pretty extensive discussion w/tech support.

Re: modelcheck metrics

RussellJ.H.Hsu — Fri, 08 Jul 2016 02:19:34 GMT

The requirements for ThingWorx Platform 7.1 are:

Oracle JDK 1.8.0 (64-bit)
Tomcat 8.0.33

Please refer to system requirement for ThingWorx 7.1.

Re: modelcheck metrics

gspencer — Wed, 31 Aug 2016 16:15:56 GMT

All folks on this thread -

I have setup a fairly extensive MC configuration for my company, and went through the PAIN of poor documentation, poor PTC support, lots of trial and error, to get it working.

The MC Metrics is really a POS, but it does offer a few useful items. The AccessDB it uses/creates is actually pretty useful. What I do is export the data out to Excel and work with it there to create pivot tables, etc.

NOTE: Keep in mind that the AccessDB is a compilation of all MC runs by engineers. For instance, it a user has a model that has 1 error, and he runs MC 50 times, the AccessDB will count all 50 errors. The reverse is true if you have a clean model (lots of OK's in the AccessDB) What this means to me is that the data gathered needs to be filtered down to individual runs (by last date) to get data that means more than a trend. Something else to keep in mind is that there isn't a good way to tell which models actually got checked in, and if the errors/warning were fixed before check-in.

This is where Creo Distributed Batch comes into play - I have been playing with the tool for about a week now, and the best caveat of the tool is the ability to run MC on "checked-in" models! You can batch check against a workspace, or a context (could be brutal if many objects exist).

BTW - I COMPLETELY agree with Mike Lockwood that ALL data should be allowed to check-in, to keep the data secure. However, if using Gatekeeper, much easier said than done. Mike's way of controlling the quality through the Change process is an interesting one.

If anyone on this thread would like to have an open teleconference to discuss MC more, I'd love to see that.

Perhaps we can even get Rosemary on the call

Re: modelcheck metrics

gspencer — Wed, 31 Aug 2016 17:56:08 GMT

Hi Mike,

Can I get those cheat sheets from you that you mentioned?

Also, at some point, would you be willing to share a bit about how you've incorporated MC quality checks into the EC process?

Thanks,

Gary Spencer

Re: modelcheck metrics

rburgess-2 — Thu, 13 Jul 2017 11:40:30 GMT

Hello all,

I'm a little late to the party but, I too am trying to set up ModelCheck at my company. (I've used MC for years OOTB but not configured.) Any help is/will be greatly appreciated. My first task for MC is to clean up 20+ years of legacy garbage. Has anyone attempted a nightmare such as this? (parameters, layers, drawing format/config......) I would also like to add check parameters to the ".dsumm" reports for metrics etc. Has anyone had any luck in this area??