Episode 2: Model-Based Definition in Practice: Dennys Gomes on Digital Engineering at Vestas

In this episode of the Share PLM Podcast, we are joined by Dennys Gomes, Senior Director and Head of Engineering Excellence at Vestas. 

Dennys is a mechanical engineer specializing in automotive engineering with over 23 years of experience across R&D, engineering leadership, and industrial transformation. Having held roles ranging from designer and specialist to team leader, manager, and senior director, he brings a strong end-to-end perspective on how engineering organizations operate and scale. Currently a Senior Director and Head of Engineering Excellence at Vestas, Dennys focuses on engineering excellence, PLM strategy, and cross-functional alignment, with deep expertise in Model-Based Definition (MBD), Model-Based Enterprise (MBE), and PDM/PLM systems. 

Here’s what you can look forward to in this insightful episode:
⚉ Discovering the Opportunity for Model-Based Definition
⚉ Explaining Model-Based Definition in Simple Terms
⚉ Addressing Legal and Contractual Considerations 
⚉ Moving from Document-Centric to Data-Centric Thinking
⚉ The Efficiency Gains of Model-Based Engineering 
⚉ Enabling Machine-to-Machine Communication 
⚉ Starting the MBD Journey at Vestas
⚉ Collaborating with Suppliers


CONNECT WITH DENNYS: 
⚉ LinkedIn: https://www.linkedin.com/in/dennys-gomes-1b49b6/ 

CONNECT WITH SHARE PLM:
⚉ Website: https://shareplm.com/ 


Join us every month to listen to fascinating interviews, where we cover a wide array of topics, from actionable tips, to personal experiences, to strategies that you can implement into your PLM strategy.

If you have an interesting story to share and want to join the conversation, contact us and let's chat. We can't wait to hear from you!

Transcript

[00:00:11] BEATRIZ GONZALEZ:
Hello everyone, and welcome to Share PLM podcast. I'm Beatriz González co-founder and CEO of Share PLM, and as always, I'm here with my cohost, Jos Voskuil. Hello, Jos. How are you doing? And who is with us today?

[00:00:25] JOS VOSKUIL:
I'm doing great. I'm talking a lot about AI, but today we're not going to talk about AI. Today we have a very interesting guest whom I met the first time at the PLM Roadmap in the PT Europe, in Paris, and his name is Dennys Gomes.

And he’s passionate and clear keynote about the introduction of model-based definition at Vestas hit me right in the heart because I'm also an advocate of model-based enterprise, but he's doing it. So Dennys, a warm welcome and we are very eager to, to learn from you about your personal journey, your model based definition journey at Vestas. So let's start first with you. Who is Dennis Gomez?

[00:01:02] DENNYS GOMES:
Alright, thank you very much for having me here. My name is Dennys Gomes, as you just mentioned, and I actually from Brazil. I'm half Brazilian, half Portuguese, and I actually live my whole life in Brazil where I studied mechanical engineering. And then I actually moved to Europe, more especially in the Sweden to study automotive engineering.

And there I actually work for Volvo trucks and Volvo cars. And while in Volvo I work quite a lot with mechanical engineering. I'm a mechanical engineer by trade, and I got a bit curious about the role of digital 3D models into the production basis and how actually we produce things.

And while I was working at Volvo trucks, I learned about 2006, 2007 that we could actually didn't have to be dependent on to the joins, we could move from other things. I cut a little bit from that point because I actually decided to leave the fossil fuel industry, which is basically the automotive industry to move to Denmark, where I actually start working for wind turbines OEM first I work in a company called Suzlon, is an Indian company who had the research center here in Denmark, and after that I actually moved to Vestas, which is the biggest OEM for wind turbines.

And while working at Vestas, I again came back. I was working as mechanical engineer first, working with mechanical and electrical components. Then I learned a bit more about the blades. And once I got to know the entire type of work and components we had in Vestas is actually, I started applying that knowledge together with what I learned at Volvo or PLM.

And a mix between mechanical engineer and PLM has been quite good, and that actually starts moving me towards looking what could I do in order to make everyone's life better, faster, more efficient? And my career in PLM has been always towards digitalization and always towards finding better ways of doing things.

And then of course, I stumble again in a model-based definition, when I found out that we were producing drawings and materials for our sub-suppliers that had contained much more information than they needed, and then I decided to actually turn and move towards that direction.

[00:03:23] JOS VOSKUIL:
Okay. I think that you mentioned a magic word, model-based definition. I think it was not completely new for you because the automotive industry and the aerospace, they are, I would say, front runners on the model-based definition because they see the benefits, the efficiency in it. But how was your start with model-based definition in, in a company like Vestas where you're talking to do it?

[00:03:45] DENNYS GOMES:
No, in Vestas. I actually took in 2016. A, uh, idea of looking strategically what we needed to do to improve efficiency. And then I remember all the, the work, actually I study in 2007 and the trial I've given to Volvo and I say, okay, I think this is something I know the automotive and aerospace are much more major.

They have much more volumes and uh, they have much more as what they call, like this maturity. The wind is a pretty new, it doesn't have a lot of maturity, but then it came to me, that could be a leapfrog. Event that would bring the wind industry much closer to the rest of the, those major industries.

Actually, if you start moving on that digitalization of the documentation that we have, as I mentioned, I start realizing some of the patterns that we are producing drawings with too much information that we're not being used by our sub-suppliers or our car machines and so on. So I realized why we need to have all this. Then the question being asked, we start realizing that we could do something to improve our efficiency.

[00:04:50] JOS VOSKUIL:
I'm just thinking, Dennys, we know what is model-based definition, but how would you explain model-based definition to people that are not familiar with the concept?

[00:04:59] DENNYS GOMES:
Okay. I think the first thing that comes into people's head when we talk about model-based definition is that, it's a model, right? And it's defining something. And what I try to explain to them is that model-based definition is about changing a paradigm from paper to the drawings that has been with us for around six millennia.

I mean, the first time that as anyone has actually seen documented a 2D drawing was from a Sumelian tablet that was actually pointing out dimensions for the temple. We actually had that specific technology for all this time. And now with the digital age, we're able actually to translate all knowledge that has been put on a paper that actually people have to study years and years to be able to understand into something that is much more intuitive. Something that can be literally understood by a child when they actually are moving, mend, rotating.

And nowadays with the current technology that everyone has a screen, everyone knows how to use a screen and rotate things, they actually can operate this. So that's actually the intention was to simplify. So model-based definition for me is move out from to the paper to a simplified way that anyone could access and verify and understand a specific object.

[00:06:16] JOS VOSKUIL:
Okay. Does it have to do with the generations? You say the, the old Sumerians they were, I would say paperous based or clay based. Do you see also model-based definition as something for more modern people that are grow up with the a digital 3D environment?

[00:06:32] DENNYS GOMES:
I think not Jos, actually, we experimented this a little bit when talking to our sub-suppliers, is that the fact that you're creating something in 3D, it makes it very, very intuitive for anyone. And we have experienced that it takes around one, two hours to have anyone learning how to manipulate this either in a tablet with a screen touch, or in a normal computer with a mouse. It goes really quick and the level of appreciation that we get from people who are looking to this.

It's quite high because you know, if someone that is not an engineer that has been formally trained, understanding how views are placed, how the rule of dimensioning is in place, once they pick up a model and they start moving it and start inquiring what they see, it's much easier for them because they do not have a lot of interpretation rules to be used on their own ads.

So literally anyone any age can actually move around and test how this, I had actually my kids playing a little bit around with some of the models they have had to see if they understood and it was, yeah, I, I know what you're talking about. It's interesting.

[00:07:36] BEATRIZ GONZALEZ:
And because we are talking about two different things, so the viewers of the models said that it might be easy and intuitive, and also we have the engineers that are creating the model itself, and they come also with a 3D designer background from the university.

But what about the other departments that needs to also to contribute with information to the model, but they are not so familiar. How do you manage to train them? I would say, or that they are involved on this?

[00:08:05] DENNYS GOMES:
Bea, what we have actually experienced was the following. There's quite a lot of people who use this documentation, right? And some people have the knowledge from the 2D drawings because they been actually seeing them for their whole careers. But we have a nice example. It was someone from our procurement team. Who has been actually manipulating these paper drawings all the time and they actually have been asking, no, we cannot move to model-based definition, or we don't know how to use the models and so on.

And once I actually, together with my team, we actually bring them the model and we tell them exactly how to use this. Again, it takes half an hour, one hour to get anyone that has no experience to be able to, to learn how to, these small tools will be use, right. How to measure things, how to do this. And my experiences for everyone that I've trained from our supply chain, from our procurement team, even the people who have been helping me on the legal team.

Because sometimes for to be able to look at patents and see what we can and cannot send to our sub-suppliers, they have to walk around. Right. And they actually have been telling me that this became much easier for them to do their work using a 3D model of annotations. Then with the paper, because they, as I mentioned, it's all discussion about the loss in translation that could happen with people who do not have the training or the experience basically disappears.

[00:09:26] JOS VOSKUIL:
You are mentioning, uh, many dimensions of the model based definition even the legal part, which sometimes I realize and see at companies is also important because historically the drawing was the authoritative source of truth, and the model was just, I would say, a small help if you need more understanding. Did you have challenges with the legal change of model-based definition with US suppliers?

[00:09:49] DENNYS GOMES:
No, to tell the truth, it was a surprise for me Jos, because when I was working at Volvo, I was in the time when they actually made the massive switch between a 2D to 3D document to become the legal binding document.

And it was a quite painful process, right, I think to rewrite contracts and go through this. But I have to confess that the legal department at Vestas had actually extremely fantastic insight that instead of actually enforcing the change from 2D to 3D, they were smart enough to just change for data. 

So the data is what is important and data become agnostic. If you have a model, if you have just dimensions, if you have Word documents and the suppliers, they actually did not complain about it. They just understood, okay, data can be consumed and used in many different ways and can be actually delivered in different ways as well. So that actually made our life much easier to make this master switch between a 2D and a 3D model.

[00:10:46] JOS VOSKUIL:
It's a clever approach. My first question would be, is the word data not generic enough or too much generic?

[00:10:52] DENNYS GOMES:
What I hear from my legal team is that it was a generic enough for the purpose we had. And of course, and that actually helped us quite a lot in terms of being able to handle legacy data.

Because one of the things when you have implementation of model based enterprise and model based definition, one of the questions is what we do with the legacy data. We had investors maybe 30 years of 2D drawings, right? We cannot convert everything one from one day to the other one. It'll take time and moment.

And that's the legal team has been clever enough to include that. And just make sure that we explain when we are giving the suppliers the information that from this point onwards we are gonna do the, the 3D models with annotations, but everything before applies as well because it's all data that has been provided to you at a certain point of time.

[00:11:41] JOS VOSKUIL:
Okay. That could be a good advice for listeners that are also considering moving to model-based definition and sometimes struggle with the the legal part.

[00:11:50] BEATRIZ GONZALEZ:
Yeah. And also we have a lot of listeners that they are struggling on changing the mindset of the company from document centric to data centric approach. How would be your advice of how to convince or share the benefits of the data oriented approach to the organization so the mindset change?

[00:12:10] DENNYS GOMES:
Yeah, Bea, I think the first thing is that there's no way back, right? Especially with the advances we are having with databases and now AI, unless you have structured data in, in some sort of database, some sort of way to convince this to, to anyone, you actually gonna struggle and gonna lose competitiveness.

So I think the fact is not how to convince is just what is the tempo and the pace of doing this transformation between document base to database pair. And again, it doesn't mean that documents will stop existing, it's just as documents become a subset of a structured data being presented in a database or in a model or anything that actually comes.

And sometimes even the model is a subset of a database because everything exists in our case, in the PLM. And the PLM is actually transformed, at least translated into a model with annotations, but everything exists on the database first, and then is actually expressed in different ways.
 
[00:13:12] JOS VOSKUIL:
Okay. You're mentioning competitiveness, Dennys, and I think also you gave the examples during the the conference, but for the listeners, why would you do model-based definition? What are the major impacts of this methodology and moving to a data-driven environment?

[00:13:27] DENNYS GOMES:
Jos, what we have learned first from the studies that we have made with the standards and also what we have learned from aerospace and the automotive, is that you have a gain, and the gain is primarily effectiveness and efficiency on your engineering organization, meaning that you're gonna consume less hours to produce your work. It's not because there's less information, it's because your model contains a lot of information already implicit that you do not need to express.

The second is that once you have a model with a information or a database, as I mentioned, you can actually enable something called machine to machine interaction.

So traditionally, everyone who has been needing to produce something, they had to create a drawing, and this drawing has to be sent to a sub supplier. The supplier has to interpret and then program a cam machine or a machining machine and so on. What you do when you do moving towards model-based enterprises, actually try to bypass as much as you can this translation period. So you do as little as possible manual work and that actually increases the efficiency of the engineering organization and as I mentioned, have less issues on understanding the data that you're sending.
So it's much less phone calls between, oh, what? What about this dimension here? What does it mean? Are you sure this tolerance is this? And this actually help us quite a lot on this.

And for our supply chain, the moment you have something that can be used as a reference. For the final product, it helps quite a lot. So I, I give an example, a model which is with annotations is basically a virtual twin on the reality. And when you get a physical component made by our sub-suppliers, we ask them to either do a measurement of all the important information or a laser scan to tell the truth. A laser scan is actually the best because the laser scan creates a solid and that solid can be, uh, compared to what we have sent them as our source of the truth for the digital twin. And then the model-based defined model becomes a pass gauge. So if it fits approved, if it doesn't fit, go back and actually do something different. So we use it for our PPAP.

So our PPAP and our first article inspection in production has been significantly improved due to that. So our suppliers are happier, because they don't have to have measurements for everything. And our PPAP is happier because we have this past gauge that can increase our quality because we're not losing any opportunity to say, okay, a component is not in the, in the right quality, or we miss to measure something. So it actually came to a lot of improvements in terms of rework. We have much less rework.

[00:16:54] JOS VOSKUIL:
And how to get there. Uh, I mean, you, you started there, you said 2018 with your project in Vestas?

[00:17:01] DENNYS GOMES:
Yeah. Let's put like this in Vestas, we start flirting a little bit with model-based definition in 2018. Then it took us a couple of years to convince everyone that there will be a good idea and no, I cannot say that it was something easy, but we needed to capture the value.

And to capture the value we had to study quite a lot. We have to look outside who has been trying to do this. We have to see what is the implementation of the existing standards on this. So we look at the military standard in US, we look at AS and I standards, we have looked at ISO standards, and we start probing a little bit who has done, who has succeed, and then we manage to get a bit of information from the US government. They had a benchmark and then we used a benchmark to assess our company and then we could actually from that point, trace what we needed to do to be able to get there.

So it's nice that there is a couple of benchmarks that we can use as a reference, and I think we are becoming a benchmark for the Nordics as well. I think there is a lot of companies have been contacting us to be able to get some information on what we have done. But the moment you actually lay out what you want to do, basically, you just have to convince by showing me the value. Every time you show a value, every time you show a gain, you get a new person who is with you and then you start increasing.

It's a journey that takes some time. You need to do some pilots, and if you start actually using the power of having a model and pointing out that you can actually have much more faster and higher quality tolerance analysis, cost verification, if you can actually use the tool that you actually, the model that you create to point out that you can actually do a lot of more virtual simulations. The whole discussion with the digital twin, you actually enable quite a lot on the mechanical design parts. You slowly apply little by little and you get a lot of things that can actually move forward you on the journey.

[00:18:56] JOS VOSKUIL:
Okay, so what I hear, it is really a journey. It's, uh, studying and it's, I would say being very consistent, looking to references. I've seen also companies that say, let's buy a system and use the MBD tool from this, uh, vendor. Uh, did you face this type of approach? Also, people jumping too fast to a tool, to a solution instead of doing the, the groundwork?

[00:19:17] DENNYS GOMES:
No, Jos, I think we actually got some resistance here on the journey because what happens that is not that the systems, I mean CAD systems and PLM systems are supporting model based definition, at least since 2010, right?

So there is quite a lot that could be done in the past. But intention is you need to change the mindset of a organization, a supply chain, and a value chain. From this 6,000 years or 120 years, if you look for paper, the industrial paper drawings, and that takes some time, right? You need to be able to show the value and slowly pick a pace.

But what is actually super interesting is that the moment you actually get to convince and get to show the value and start capturing the value, it's kind of an exponential curve. So right now we are in a situation that we are getting a bit of a hockey stick, and the amount of reviewers that I have to do reviews in the models is actually, I'm running short of them. So I have actually, I'm training right now reviewers because there's so many items, so many components, so many parts of the wind turbine is actually going to model based definition. There is not enough people with the right skills to review a 3D model, which is a bit different from reviewing to the drawing.

[00:20:29] JOS VOSKUIL:
Okay. Do you work a lot with, uh, suppliers? Because I think that's where I always hear the friction comes. Yeah.

[00:20:35] DENNYS GOMES:
Yeah. So the suppliers have been from the beginning with us because we knew, I suspected the suppliers will be a problem, especially when they saw a letter from, uh, Rolls Royce turbines stating to their suppliers in 2019, either you in or you're not our suppliers anymore. And then I say, okay, that's interesting. 

So what has happened is that we started, when we looking at the opportunities to simplify our documentation, it's not what I think is what should be simplified. Primarily we are picking up our huge A zero drawings with many pages and coming to the suppliers and asking them, frankly, what are you using out of the information we have. And then they were actually highlighting that maybe 60% were not being used. And then we ask them the following, if we actually only produce a documentation, who contains the remaining 40% that you actually are using, and we give you a 3D model, will it be better for you? And they say yes.

And then of course we work, we prepare some prototypes. We start sending to them and ask them the following question, are you able to produce this component that you have for us with the same quality or better with the documentation I sent you? And then there is always a discussion back and forth between us and the supplier to fine tune what it is.

And then of course, we have done that with 82 suppliers, I think, different suppliers from different areas, different technologies, and you end up finding what is your sweet spot. And I think right now in Vestas, we are covering composites, we're covering weldments, we are covering shit metal, we are covering extreme big castings, we are covering blades and we are covering as well, machining. 

There's some, still some technology that we need to go to. We have not talked yet about plastics. We're not talked yet about other parts in terms of a big assemblies and so on. But it's the journey we are going through. But you can see the moment you start dialoguing, your suppliers will get to buy in.

And of course there's some suppliers who will say, no way, I don't want that. We have a dirty shop. A lot of people who have their crazy hands, they cannot use a laptop, they don't know how to use a mouse and so on. And then what happens that in Vestas, and my advice for everyone is basically to send someone there. And actually show them how it works, point out how easy everything can become with their production. I mean, can carry them because your suppliers basically are the ones who are providing you with the quality products you're actually putting together. And I think this is the journey that we have.

[00:23:08] BEATRIZ GONZALEZ:
And coming back to the journey, I'm interested also in the, because you started at first, you trying to study and identify the value and couple of years to have everyone on board. Then you were looking at the standards and then you were simplifying the amount of data it was needed, and the next step would be like looking at your system landscape and how was that step indeed like, because probably you have a lot of system, you were saying like legacy data, legacy system, so it's also going through a simplification of the systems or it's not necessary?

[00:23:43] DENNYS GOMES:
The implementation of model-based enterprise is very agnostic of systems. So you need to have systems, of course, you need to have a CAD system. You need to have a PLM system. You need to have an ERP system.

[00:23:55] BEATRIZ GONZALEZ:
But what happen if you have more than one of those?

[00:23:58] DENNYS GOMES:
That work as well. So because you know it's supposed to be, as I mentioned to you, especially the CAD systems, they have actually, all of them are in a situation. They can produce the information and the information has to be the final outcome that you get is basically a new to format, which is covered by ISO. And that one contains the information. If you want to include more information, you have to go to other forms called like QIF and so on.

But the point is, any system can generate the output. How the systems interact is what you need to do as a company to find out what is your ecosystem, how you actually put two CAD tools together to generate similar output. How we can generate, put two PLMs or three PLMs together to generate something for the value chain through the ERP.

How you connect everyone together is not something that is impossible, but you know, you don't necessarily have to simplify, but of course it much easier. If you do this. In Vestas, we actually have simplified, we have a quite strong streamline with one sub supplier, but I have seen while I was, uh, talking to other companies, plus doing the investigations that is not necessarily mandatory to have one system that rules it up. You can actually use different APIs to connect. I can see the trend for the last three and four years. More and more, the PLM suppliers actually open up to get APIs shared so they can actually use the multi-CAD.

[00:25:29] JOS VOSKUIL:
Okay. I think then this time is really flying and I'm excited to hear such a positive MBD story. I talk a lot with companies. They see the problems, you see the opportunities, and that's why I'm, I'm curious about your experience and, and that's my final question. Always like experience is what you get when you don't get what you expect. And maybe you can share one of your best experience in your career for the audience.

[00:25:54] DENNYS GOMES:
I think regarding the model based definition, I think one of the biggest roadblocks we got is that we have, I still some suppliers who wants to have a drawing, right? And what we have been doing is basically I managed to find a very good compromise. So I told them, I'm not creating a drawing, but I can help you to print the 3D.

So, as you might have heard, the model based definition is you create some sort of views in a 3D that everyone can actually toggle to. Everyone can rotate, but you can toggle to different perspectives that you actually put on a model. In my team, we actually had this great idea to actually, to enable a way to print those into a sort of 2D drawing that the suppliers could use.
And that means that I am providing them a step file. I'm providing them a annotated 3D model and I'm providing them a printout of a document which represents every single combined state or every single, let's put like this view of the 3D model. And I think that's a compromise we took.

And they all companies who actually do not do that, they just say you have to learn how to use the viewer. There are some companies who actually are still doing the 2D drawings in parallel. It depends on how is your dialogue with a subsupplier. Jos, right.

[00:27:14] JOS VOSKUIL:
As long as there is no workaround. I think that's the main thing. Yeah.

[00:27:19] DENNYS GOMES:
That's what we try to do. Because you know, the workarounds just create more problem for you in the future because at some point then you're gonna have to phase that out.

So the intention we had at least, was to integrate as much as we could and do as much as we could in a one line of logic that could easily, uh, with the nature of our sub-suppliers, they will be just dropping the ends of the peak.
My vision is that in the short time future, I would not need to print PDFs anymore because everyone will know how to use the models, but I'm not actually excluding that the suppliers can print it by themselves in the future.

[00:27:54] BEATRIZ GONZALEZ:
Mm-hmm.

[00:27:55] JOS VOSKUIL:
Okay.

[00:27:56] BEATRIZ GONZALEZ:
What would be then the key message you would like our listeners to remember from this episode?

[00:28:02] DENNYS GOMES:
I think Bea, the first thing I would say is if when you are doing and going to a model based enterprise, you're not doing it alone. You need to do with your procurement team. You need to do with your supply chain team. And most important, you have to have your suppliers in mind. You cannot push the towards them. The resistance you're gonna receive are gonna be impressive.

The second thing is that it's not about making it a 2D into a 3D. It's about recombining your information and data. In a way that you can actually have machines, understanding machines, and you can make it so intuitive that anyone can go through it.

And the third and most important one is remember that we are hardwired to look from a technical perspective to a 2D drawing for 6,000 years, right? So it's a change management. You need to be able to have everyone on board. You need to ask them to, what is a perks that you're gonna foresee in the future? Work together, solve the problem slowly. Do not rush into anything, and use a change management team to support you. And I thank you guys, in Spain are pretty good on change management. So a team like yours, it'll be really useful for anyone to get this through.

[00:29:28] BEATRIZ GONZALEZ:
Thank you very much, Dennys, and thank you very much to our listeners for tuning in.

[00:29:23] DENNYS GOMES:
Thank you for having me.