Session Information:

How to Identify and Resolve “Points of Failure” in Business Processes

This session will establish critical process and application capabilities required to deliver the Core Model-based Definition Infrastructure require to enable the Model-based Enterprise. We will additionally walk through the critical process and application capabilities required to establish the Model-based Enterprise through Mfg. Engineering, Quality, Tooling, Manufacturing, Procurement, and your Mfg. Outsource Supply Base.

Key Takeaways:

1. Strategic Roadmapping is Critical as the Journey will be complex and staged
2. Substantial and Measurable ROI has been proven by A&D an Automotive Industries
3. Digital Enterprise is not attainable absent a full conversion to 3D
4. The Model-based Design Environment will drive additional ROI from your CAD & PLM Assets

Session Transcript:

Hello everyone, Welcome back to. I BPM Live, were Processed Meats, Tech Knowledge, People, and Culture. So, our next segment is really excited, and we're going to talk about technology. We're going to talk about strategy, we're gonna talk about creating a model based system for the answer price. And that we have one of the best when it comes to that. Ken ... is the senior partner and manager director of DSA. Hello can we're really glad to have you here. Yeah, DSA is a PLM and Digital Enterprise and Consulting Practice and can has more than 35 years of experience in engineering systems and process design. And the development of global digital enterprise strategies. So, real honor to have can hear.

We went from ..., Israel, now, from close to Cleveland, Ohio, and can, it's awesome to have you here this morning, and, and that we're very honored to have you share your expertise with us today around the world and just a few minutes. Ha! Very nice. Thank HSA.

Alright, good morning everyone. Thank you for being here today.

I am going to talk about deriving a measurable return on investment from implementation of a model based enterprise.

See, here, I'm having a little trouble, moment.

Oops!

OK, I apologize for having a little trouble getting the presentation to ever go. All right, first, we're going to talk about the Pillars of Success for Industry four, dato. It's critical to understand that, before you can move through digital, enterprise, and industry forward auto. There's a substantial amount of foundational work that you have to do.

What I'm going to be talking about today is primarily moving from two D to three-d. so that you can better enable not only the engineering organization, but better enable automation, digital automation, through the six Pillars of Success.

In industry for it auto.

So, the first pillar, increased uptime, reduction in defects, new product introduction, improved productivity, human interface, and energy consumption. Now, if you look at reduction in defects, improved productivity, and human interface, you have to have very high quality engineering data.

Specifically today we're talking about, um, a technical data package that has all of the information in it required to enable a high first pass yield of the downstream dependent supply chain broth efforts.

So I'm going to talk today about how to get there.

So we start with the first foundational capability you have to have standards. You have to have modeling standards, PMI standards, ballooning standards. And when you're creating a bill of characteristics, you need to create a three D machine, readable bill of characteristics, so that you can enable all the potential downstream automation.

Getting the standards developed as the first TAF, getting them, and making sure the organization understands them clearly is the next hard task.

And a lot of organizations throw this into self paced training, and that has not proven to work very well. Also, just real quick, this is just say here, your video feed got frozen.

So what I would suggest you do is just turn off your camera angle with audio without your video feed, OK.

Thank you.

Thank you.

That's cool.

All right.

So institutionalizing the standard through the standards, through extensive training, and then modifying your engineering change process, at least, initially, in a manual way, so that there is a checklist of things that you'll look for when you're evaluating. The release of a three-d. model is critical to make sure that your organization is adhering to the new standards.

After that, you'll need data definitions. What a data definition is for a specific part family, or an airfoil versus a gear, for example.

What is all the information required to enable that high first pass yield of the downstream dependent supply chain processes? That's what a data definition is.

That's the all the information. It's a specification for your technical data packet.

So when you look across your part family, if you have 30 part families and you're part family tree, you're going to have somewhere between 8 and 15 different data definitions. So, for example, if we looked at their foil, you'd have a data definition for casting.

And then you'd have a data definition for complex machining.

So just for an airfoil, you would exploit two different data definitions as you move that part, through those two critical stages of its manufacturing process.

Once you've got your data definitions and your standards in place, then, you can build some initial technical data packages.

And, you can do a walk through of metrology manufacturing, engineering, quality, tolling, procurement, manufacturing, planning, and particularly with your outsource manufacturers, your critical, outsource manufacturing base. Once that set of walkthroughs is executed, it will allow you to do two things that will allow you to tune the standards and tune your data definitions and execute some early pilots.

So, for example, try to auto Gen, a CMM code.

Have some manufacturing engineers take technical data packages that drive manufacturing work instructions, and you'll see that it takes a lot less time for them to do that from a TDP than it did prior to creating that data definition and building network product for them.

OK?

So, establishing the flight procedures. There is a very high level picture here of the process. What is the core infrastructure that you have to build?

So you have to have marbling standards.

You're going to have to work spoilt PLM extensively.

Um, and you're going to have to add some technology, very likely to PLM.

You're gonna want your need to exploit middleware for storing CAD clones. You're going to have to update your part management and part object model. Um, it's relationship management roles.

You're going to have to modify your engineering release process to automatically Trevor trigger particular validations of the CAD data, and perhaps even print, do produce ability, validation of the CAD data.

In an automated, automated way, and then automatically generate the CAD derivatives, J K, step, three-d. PDF, et cetera, and generate the technical data package at particular, life cycle states or states of maturity through the park management process.

So, it's critical that you understand what the end to end infrastructure looks like.

We're going to talk a little more detail about this.

So, if we take CAD data Validation, and there really are two critical components, once you have standards and data definitions derived, there's two critical, or I should say, three critical capabilities you need to have. You need to have data good data validation, good produce ability validation. You have to be able to generate technical data packages, and all three of those things are going to require substantial modifications to your PLM application environment.

When you're doing CAD data validation checks, it's important to remember a couple of things on. there are some easy stuff to do, and that is to implement generic checks.

So, is a component missing? Do you have a duplicate assembly? Is there a blocked whole? Do you have a void in the model?

Those are generic CAD data validation checks that will work across all part families.

Then, you have to put in, as you move through multiple phases of increasing the sophistication of your CAD data validation process.

You have to move to parts specific checks. So, if you have a complex machine part that requires a tolerance of plus or thousands, you'll have to configure a tolerance check for that part family at that.

At that tolerance, if you have a a small, non complex machine part that is plus or thousands, you'll have to make a copy of that same tolerance check, and save it out, so that you could run it against that particular part, family, that it's related to.

So, in order to do that, you have to make some pretty substantial PLM extensions. You have to define a park, family, tree, update your classification schema to enable people to classify each part into a park family.

The engineering change process is going to read that value, and be able to determine what set of checks you run on that particular part family, and what data definition you use to produce technical data package.

You'll have to deploy a dispatcher.

You'll often displace cloning tools Um, and you'll need to, eventually update the ACO process to enable ... of the validation reports that you get, such that if something is a tolerance check, for example, is critical, and your tolerance check fails that, or your, your raw validation process fails. That tolerance check, then you need to be able to automatically stop the change process, and notify the change owner, for example.

So as you move through this, the requirements for exploitation of PLM get more and more extensive.

Very similarly, if we look at technical data publishing, you have to specify the data definitions across a family tree, Specify the carousel views to be placed in each technical data, package by part family they will vary. The specific views, the downstream supply chain, requires, for a given part family can vary. Every CAD system blogs about 100 different views. So you just have to pick the views most relevant to each part family.

Then you have to drive a template for each data definition, that's that's the form, basically, and then you have to derive recipes to move data from PLM ERP.

And perhaps other systems, into that technical data package, now, The interesting thing about those two things, a lot of people go out and buy a third party application for technical data publishing, when they have a sophisticated PLM environment, You don't have to buy a third party technical data package. You have all the tools right in your PLM application to do this. So think of a template, as nothing more than a form.

In the average PLM system, you can create a complex form and validate that form in 3 to 4 days.

And then the recipes, as they're called, in those 33rd party tech technologies, are nothing more than data migration scripts, which you could write in your own PLM.

So by building it in PLM, you don't incur all the capital costs for servers. And for the software, you use the configuration, cos it's about the same. And then, of course, you're not maintaining another or another application environment through helpdesk support and regular updates on server maintenance, et cetera, et cetera, et cetera.

So as you think about technical data publishing, think about doing it right in PLM and run a pilot.

So that really talks about the critical infrastructure required to enable the model based design environment. Model based design environment is critical to enabling the model based enterprise and the model based enterprise. We will do things like automatically generate CMM code.

Um, do intelligent first article inspection, um, increase the, or reduce the amount of time substantially to drug manufacturer, work instructions, et cetera, et cetera.

So, that's where we look across the entire set of available ROI for model based design environment and model based enterprise. There are about 100 points, plus or minus a value proposition for any given company.

Over the last three years we've been working with a number of organizations in automotive and aerospace who are well down the road on this environment. And we have a very sophisticated group of companies that come together on a quarterly basis and assemble return on investment metrics.

So we have a substantial evidentiary library on ROI for model based design and per model based enterprise and all of those companies are also reporting the critical capabilities that they had to put in process and application capabilities to achieve the ROI. So any of you that would like to have that kind of detail about how to move forward, please feel free to reach out.

I'd like to talk now, before we wrap it up, a little bit, about our, oh, hi. So I'm going to provide some case studies here.

So case study one, jet engine manufacturers, six new engine programs, a year average of 2000 parts per program. The average number of drawing sheets, per part, with five, the average time to validate each sheet was 3.3 hours fully burn rate at $120 an hour. So the annual validation time, should now, the dollar sign, there, was 40,000 hours, plus or minus, at an average annual cost of four point eight million dollars.

After we put in automated validation.

And it was both on generic checks and part family specific checks across 32 different part families.

We were, we achieved a real, measurable reduction, cost savings of 1.7 plus million dollars, which was a productivity gain of 7.2 man years. And so, the company was able to take all seven of those people out of product validation, and deploy them on new projects to do real engineering work. To do systems, engineering, work requirements, work, and CAD design work.

So, you can, even though they kept those employees, they did achieve a one point seven million dollar annual reduction in the cost to validate CAD, real measureable money in the bank, for that particular case, for that particular process requirement.

Next example, on average annual engineering quality notifications to habit from suppliers to get the model to match the drawing that the automotive OEM was selling.

Um, so, there was an average total hours to remediate, and there was four hours to 35 minutes.

The, they had, they did 374 of these on average per year. I'm sorry, back up?

And, it cost them 3200 hours and lost engineering CAD designer productivity, that's 1.5 man, years, at a cost of a half a million dollars a year.

So, for that keeps changing, going to take time off the average annual ... six months, after we put model based design, environment, and was 11.

So they substantially reduce that waste.

So, that was an increase in productivity.

Next, exit two examples.

Um, this has produced the ability.

So we have 12,000 parts per year. 13% are built in-house, and that's 1860 work instructions per year, the average time to derive and review and release the work instruction.

Before technical data package was available was 22 hours.

The average amount of time it took to create, review, and validate manufacturing or construction after TDP was 9.5 hours.

At an average annual savings of two point four billion dollars.

In this particular case, this organization laid off eight manufacturing engineers and put that money in the bank.

So that way, in this particular case, high-tech manufacturer, they actually were able to achieve a staff reduction and still reduce the cycle time to release of a manufacturing or construction, and by the way, the yield went up significantly. So we did not get, um, I think they reduced the number of returns to manufacturing engineering of a work construction. That was incorrect by more than 90%.

Automotive OEMs are subjecting parts to produce ability review there, plus or parts subjected annually.

In this one case, this was the one on product line.

By the way, the average total hours executed groups produced ability review was about 2.5.

The total annual manufacturing engineering work hours was 15,000 hours at an average annual cost of one point seven million dollars.

Once again, this was a situation where we were actually able to make a staff reduction of manufacturing engineers and achieve, in this case, about a $900,000 real savings money in the bank.

So, as you then look across product design, manufacturing, trolling engineering, you might want to grab a screenshot of this because these are some of the other areas where you can achieve ROI. And once again, if you want evidentiary ROI, you want to know what real clients have gotten for. Each one of these things. Reach out if you want to know what's on, how to calculate it, reach out.

If you wanna know what critical process, an application capability is required, we'll reach out, Quality procurement, outsource manufacturing, Probably, want to get a picture of this, as well.

Now, the big takeaway from this, for us, from this slide, is if you have a very large, outsource manufacturing base, you're going to want to know what their model based enterprise maturity level is. What is their capability maturity level?

So, when we survey outsourced manufacturers for our clients, we're able to determine all what clients are immediately ready to consume a three-d. technical data package, and provide a return on investment back to our client.

If you don't do this, if you just send the TTP's out to the outsource manufacturers and don't put a process in place for getting the roi back to them, you're going to leave a lot of money on the table.

So, you have to understand their maturity for all those critical suppliers. It's, you know, it's 80, 20, 20% are critical. Have a maturation plan for them.

And then understand from internal data and from evidentiary data.

What ROI you should expect for every, for every point of ROI there is and look at the history. You have with that outsource manufacturer, if they're billing you 26 hours to create a work instruction.

And, you know, it takes, are you 11 hours.

You should, you should be ready to negotiate what they're going to charge you from now on, on average, to drive a manufacturing work instruction so that you can see a line itemized reduction in your cost per unit, Engineered and manufactured by your Outsource Manufacturing team.

So finally, the last thing, because this is so complex.

This is a project that is, as large as implementing PLM, for example, putting in model based design and model based enterprise. is that large, an initiative. You are going to be the executed stages. You are going to need to execute a strategic road mapping exercise.

This is a good picture to snapshot, because it, it tells you what all of the critical work products are that you have to put in place. And also, don't forget, then that your model based design and your model based Enterprise roadmaps are part of a much larger set of chapters in the overall industry, forward, auto strategic roadmap.

So, call to action, derive a comprehensive strategic plan roadmap, and get busy with the foundational work and do what? You can do that while you're doing the roadmap, by the way.

Don't skip on standards, training, assure institutionalization. And then, make sure you design and build this thing for real ROI noteworthy Royals and know how to get it.

That's all for my presentation.

If your screen capture this and want to reach out, you know, that's a good time to do that, and I think I am supposed to, wow, stop sharing my screen.

Yes, can, because, because your, your video feed is not all is not gold transmitting very cleanly. I'm gonna ask you to, in this case, share your presentation and steal, and go back to that page, where you had the strategic roadmap, if you will, at least, where you have the, the big picture on your presentation, and the very good.

That one right there. That one right there. And that, and that will, there are questions that are coming in from the audience here. I'll turn my camera on. You may want to try to turn your camera on just to see if your video feeds maybe may have gotten approved at this point. Give it a shot.

And the, and if that, if that's not there, don't worry about it's not it's not critical. That would be on the interface of the goto webinar. It's not coming on.

Yeah, so, so, let's, Let's leave it off so that it doesn't, it doesn't impact your audience very clean. So, I would just ask you the questions, and I think that the reason I asked you to put that strategy map back there, is because it provides the context for a number of questions that I have seen coming up. So, for those of you know who are listening in and watching this, keep providing your questions OK. I'll get to them as as you provide, as you provide the questions and insights.

So the first question is that it was as a matter of fact, there was a comment. That was that was funny, because there was a page where you show that? The, the Initial part of the journey. And there was an incredibly complex diagram of the steps They use that to be taken, and then you said we're gonna get into the details of that, which was which was awesome, because I was like That is the high level folks because this stuff is hard, So I want you to talk to us a little bit about this.

And we appreciate the the, the, the the Expertise that you have because just simplify you must master complexity And and it's This is not necessarily a simple thing to, to do. It's a, it's a complex process, and it's a mastering that complexity that you can come up with simplifications. But there's no, there's no obvious, no trivial solutions for some of this. This, this challenges. So the question, the first question, is on, on that, on the, on the, on on on, how do you even start this journey?

In an organization that may not have, you know, thought deeply about all these Interconnections? So how do you work with an organization that that's just getting started?

Yes, so, the first and most critical task is executive, education and executive development.

So, what we would typically do is, we would come in to a client site, and we would set with the team, the engineering community and some of the operational folks.

And we would draw, we would provide some initial project plans, some outlines of roadmap, and we would work together to derive a an executive level deck to educate the executive team on what the real ROI is available for this. How should they get started? What resources will they initially need, et cetera, et cetera?

So, we offer a two week Bohm readiness preparation activity that we do for clients on that call.

So we'll work with them to put all that, inform the initial information together so that they can take it to the R, to the executive team and have that built in to their strategic digital enterprise roadmap. So, that's the first thing to do.

Once the executive gets on, executive staff gets on board.

You can start ferreting out all the rest of this detail for the entire journey and while that's going on, you start the derivation of standards and data definitions and that's how you get started.

And the next question that we had here had to do with and you gave several awesome examples on how you apply and how you derive the Roi from from those applications. But, But, if you look across all your, your clients. The, the, the opportunities that you see in the industry, in the industries that often go beyond what even your clients can see themselves, because they have limited understanding of the capabilities.

Where do you see the best applications for model based enterprises across different industries? Where is there, is there is there are certain industries or certain applications that that stand out?

Well, certainly, if you're looking at the heavy equipment, aerospace, automotive, engineering, procurement, and construction, where they're building. So, you know, for example, if you're building powerplants, you have bar turbans. You have bag houses, you know, a number of things that are standard products that you develop. It's a very good application for those things. In the high-tech industry, there is good ROI, but it's a little less than it is for a hard mechanical manufacturing provider.

Then, of course, if you look at an organization that, for example, does, their product is primarily mold injection.

So, for example, and an organization like audience, that makes automotive interiors, we still don't we are still not able to derive really good ROI from an organization that is largely a mold injection firm. But if you're an automotive OEMs, tier one, tier two, tier three supplier of mechanical parts, you can make a lot of money with model based design, a model based enterprise. If you're an aerospace, you are already marching down this path. Northrop Grumman, for example, is allocated more than $100 million to move into model based design and model based enterprise. Lockheed is almost back to that amount of money.

Pratt Whitney is already spent $15 million perhaps on, on this program. So, aerospace is really far ahead.

But large mechanical manufacturing firms or raw electromechanical manufacturing firms can achieve a substantial roi.

So that's that's very, very helpful. Another question that came up is a little bit on the and on the perspective, on the model based design environment industry. A little bit of a history on it and how he has evolved to where it is today. Any thoughts that you can share a bit on the, on the history of this discipline?

Yes.

So, a number of years ago, the United States Air Force, actually, what was the first DOD organization to really start looking at the opportunities available for them to drive their supply base to three date.

And so, there were some consortium's that were put together.

Very shortly after that the Air Force went out to the National Institutes of Standards and began to develop a group at Nest.

And there are, by the way, a couple of several committees at best one on standards.

one that's deriving, modifying a SME standards. There's a Committee for Mil Standard, 3000 A and B, which is PMI Standards, and then at D M S C.

Today, there's now an order, a group of guys, that are taking a look at how to create a machine readable bill of characteristics.

So, as the Air Force started to evolve, they started to notify their, met their manufacturing base, their suppliers, that at some point in the future that they were going to make as part of their contract, a requirement, that a baby provided a full three-dimensional technical data package.

So the aerospace industry started taking an early look at it. And as they started running pilots, they started to realize how very complex the program was and how very long it was going to take them to fully transition their entire corporation.

So they started in earnest.

The aerospace industry, about three years ago, really started going fast, and just on the heels of that, General Motors, Chrysler, Ford, BMW, et cetera, et cetera, also started large projects and started pushing their supply base, Tier 1, 2, and three, to drive to model based enterprise as well. So, those are the industries that are currently leading the charge, and making progress.

And then, of course, over the last three years, we've seen a number of organizations come in to play, to provide third party software applications.

So, produce ability, validation, CAD data validation, derivative validation, translation, technologies, and non technical data publishing tools.

We've also seen organizations produce things like intelligent three, I'm sorry, Intelligent App, AI capabilities. So, you can take the data right off the CM, and post it right back onto the model. So, you can see your issues color coded and annotated and be able to move through the FBI process much more effectively and much more efficiently.

Very good, 1 1 other question that we have here, is around the concept of digital twins, and the question is whether, you know, How how this That's kind of this three-d. model approach that you have here may marry or support, you know, this concept of digital twins. That's something you are familiar with.

Absolutely, yes. So while some people are suggesting that a two dimensional drawing of a part.

And then the physical part are digital twins they're not a three-dimensional model that has a machine readable bills of characteristics, intelligent ballooning. And, and those kinds of things, that is what a Digital Twin is.

So let's, let's, for example, take a look at a smart factory with intelligent robots and machines that all talk to one another, as there, as a product, is moving through a manufacturing process. And that's where we're headed, by the way.

Organizations like Rockwell have committed Rockwell committed last year to spend more than $3.5 billion to drive smart factories. So if you're if you have a smart factory and your machines and your robots are going to talk to one another and make autonomous decisions, they're not going to be able to read a two dimensional drawing.

It's just text, so you'll have to, as you start to move through all of those pillars of success, for industry forward, auto, you'll have to understand that creation of the Digital Twin means the physical part, and, and intelligent digital representation of the part, and that's critical and intelligent, digital representation of the part.

That's, That's, That's incredible, just the vision that you're sharing there, and that's a full automation approach with a digital tween. It's quite impressive. The, the, the last question that I have here is, related to, is related to Advanced Technologies, and you may have partially answered this already. But just to confirm, how, do you see, the play of AI, RPA, machine learning, you, know, this new X, but not new, But this exponential technologists, there are becoming more dominant, currently playing with, with this concept that you're talking about here.

Well, sulfur RPI, I've kind of describe that a little bit in the last question. Let's take something else that is emerging that will be part of a smart factor.

Let's take three D, printing.

So when you look at three-d. printing today, a, um, you have to you have, of course, you have to have a three-dimensional model or you can't do three-d. printing with a two-dimensional model. And in particular, when you are designing three-d. parts, you need to design three-d.

parts using a lattice structure, which we could never do before. We can never build, apart with a lot of structure. Because I can't take a raw piece of metal and make, make that raw piece of iron, have a latticed structure. So, we can only do that with three-d. printing.

So, today, um, you have to apply generative design techniques, lot of structure and then you change how you do real time simulation of those models so that you can create a three-dimensional model And once again, a machine readable bill of characteristics is also required so that you can feed that model directly into the three-d.

printer. The three-d.

printer will generate its own code automatically and then you'll be able to derive that lattice structured part.

So, for those of you who might not understand what I mean by lattice structure the simplest example is if you think about a woodpeckers big and his ability to power on an oak tree and not break it, the reason that is, is because the interior ever speak it's very much like a honeycomb. It's a lattice structure, and it makes it much stronger, and much lighter.

A fascinating, Ken. Fascinating. Thank you so much for sharing your expertise with all of us today. Incredible journey, complexity value, the drive for innovation and value creation all very clearly represented. Thank you so much for sharing your expertise and insights with us.

Thank you so much, Josie, and thank everyone for attending Y'all. Have a great day.

Thank you.

Ladies and gentlemen, this completes our session. and then, and they, in fact, this will be the last segment that I'll be able to facilitate today. And you'll be in the very capable hands of Brian for Russell and this beautiful British voice that will guide you through the rest of the sessions. So, thank you for joining us. Make sure that you provide feedback on this session as you exit this, this this segment and log back on for the top of the hour and we will carry on the discussion on how business process management marries technology. People, process, culture. Thank you very much.