A Renaissance of Innovation through Generative Engineering and Integrated Validation

Summary

Jennifer Piper: Welcome back to another episode of the Next Generation Design podcast. I’m your host, Jennifer Piper – and on today’s episode, we have a bit of a roundtable. We have Boris Lauber – Product Manager for Additive Manufacturing Simulation – and we also have Tod Parrella, Product Manager of NX Design. So, we have a chat about generative design. What is it? How does it look in action? And how is it going to change the landscape of the industry? As it turns out, design is a lot more than colors and cool smart technology features when it comes to the future of vehicles. Let’s just go ahead and jump into the topic and learn more about generative design. So, Boris, could you tell us a bit about what exactly generative design means?

Boris Lauber: I think ‘generative design’ is a term that is used in many applications. Actually, I think Siemens is viewing that topic of generative design even a bit more broadly. So, typically, we talk about generative engineering – so, having and using design and generative methods to generate new design ideas, but also using that kind of technology to build up system architectures in a broader way. So, that means it’s really helping engineers to build up their designs and systems in a kind of computational way. So, typically, engineers are supported by computational methods and also artificial intelligence to get good design ideas and to improve the product, the functional behavior, and the functions behind. So, typically, if you think about designing a product, in former days, it was maybe just designing the part from a kind of design way, so just CAD design and building up the geometry.

Boris Lauber: And more and more functions are involved quite early. So, what we really try to focus here on is bringing the functional aspects into the design. So, really thinking about requirements and functions upfront and then using simulation techniques before the design. So we’re really doing a front-loading of different simulation techniques, in order to find new designs that are really optimized for certain functions. We have much more design freedom from a manufacturing side, and therefore, we have much more freedom in designing parts like designing for a certain load path, or for certain functionality that could never have been manufactured before. And therefore, this kind of generative methods play a more and more important role in that domain.

Jennifer Piper: Sure! And it sounds like it’s a very innovative process as well – and like you said, allows you to be able to design parts that probably never before would have been able to be designed.

Boris Lauber: I would also mention that it’s not only a new design paradigm from a design perspective, but it’s also changing the workflows in companies. Because typically, what happened before, was you designed a part and then you used simulation techniques to validate the parts. And using this front-loading will really shift the CAE technology and all that we do in simulation before the design. So, we really let the computer design new components based on the functional requirements and then we get these design ideas and transfer them to the designers to really transfer these ideas to real CAD designs that then can be used in the further engineering process. Typically, methodologies like, for example, a topology optimization tool that is really finding new design topology is used in that domain. But it’s not only the topology optimization – I think that technology is already existing for about 15 to 20 years – but meanwhile, shifting that upfront to the design engineers and bringing that methodology into the design environment, this is really helping to bring these processes into real product development.

Jennifer Piper: Very good. Yeah! Tod, did you have anything that you wanted to add to the topic?

Tod Parrella: I guess I just wanted to stress that Siemens takes a multidisciplinary approach to this process. And we also are taking a very broad and all-encompassing approach to leverage generative engineering. When we say generative design, generative engineering, we’re really talking about advancements in computing algorithms to do complex problem solving for us. And we become the orchestrators. As the designer and the engineer, we become the orchestrators to help guide the algorithms towards solving a complex problem. And, as Boris mentioned, a lot of what we see in generative engineering today is around topology optimization – or optimization of a form. But I think it’s important to note that Siemens views this as, generative algorithms can be used for a lot of different applications. And so, you see this in our solutions for optimizations of fluid flow and thermal fluid flow engineering challenges, structural challenges, as we mentioned. We’re also seeing the use of generative engineering algorithms in our electrical design applications as well. So, we do see this, holistically. We see it as able to offer great value add and benefit to our customers and we do look to apply this wherever we can in the product that will help solve complex and multi-disciplinary engineering challenges.

Jennifer Piper: Right. Okay. So, what would you say are the benefits or the risks, I guess, in using a generative engineering approach? And I guess, Boris, I’ll start with you.

Boris Lauber: Well, of course, the benefits are quite obvious. And there are so many benefits of using this kind of technology. What happens here is that you really can formulate and concentrate on the real problem definition, on the requirements, and then letting computer technology just take all this input and iterate over several design ideas, and judge them, and validate them automatically. So, that means you’re not going into only one direction – maybe that’s some individual decision – but you can really get a large number of different design ideas that fit, for example, for certain manufacturing techniques, or for certain structural requirements; and then, you can pick the one that fits best into your process and into your engineering process. I think that’s mainly supporting engineers to rethink their designs and really to get completely new design ideas. And all the new requirements that come into the game are really considered and you don’t forget anything.

Boris Lauber: On the other hand, that’s of course, also, a risk you have in your setup because if you use computer technology to find new designs based on some requirements, as soon as you miss important requirements, of course, they will not be considered in the design idea and that’s, of course, a risk you have. So, you carefully have to see what are the real requirements you have. And maybe it’s sometimes even better to take more of them to get robust designs and to get robust engineering processes at that point.

Jennifer Piper: Yeah, I mean, it definitely sounds like it could be a balancing act, for sure.

Tod Parrella: I guess I would just add to that, that a lot of people have asked us when we’ve been talking about topics of artificial intelligence or generative engineering algorithms and what the role of the engineer is, and I think the role of the engineer remains critically important to help guide the decision making in this process. It’s very easy, as Boris alluded to, that you could go down an entirely wrong path with the wrong information. And so, having human interaction to guide this process and evaluate the analytical data is critically important to the success of using generative engineering.

Jennifer Piper: Tod brings up a really good point. Often, when we think of technology advancing and moving forward, we think that, ultimately, the people behind the technology will become obsolete. Soon, a robot or a computer will be able to do everything without us. But this isn’t necessarily the case. The engineers, the people behind the ins and outs of the technology are an essential part of the generative design process.

Tod Parrella: We’ve worked with a number of companies, including some of our other Siemens divisions that rely on our technology for their engineering design work as well. One of our internal Siemens successes within our Siemens oil and gas industry, where they were looking at new ways which they could simplify a complex system in the oil and gas burning process – where they have a very complex system that burns residual fuel, the sludge of fuel to extract further useful elements from it. It was a complex system, it was a complex manufacturing and assembly process and they really wanted to leverage the power of what generative engineering could provide them, to try to simplify this system.

Tod Parrella: They were extremely successful using the Siemens suite of tools across the entire portfolio of what we offer in the generative engineering space. They could really run through hundreds of iterations of optimizations of the parameters of the system, and really get down to a much more elegant design. In fact, actually, as Boris mentioned earlier, additive manufacturing and the role that happens to be, overlap there where, when you increase the design freedom of possibilities, it’s helpful to have manufacturing freedom on what’s possible to actually make – and in this case, Siemens oil and gas actually leveraged additionally additive manufacturing to be able to take the output from generative engineering data and produce it as is, without having to essentially re-engineer the system for traditional manufacturing processes. So, they really got double benefits.

Tod Parrella: On the back end of that success, they had a system that had significantly reduced product and assembly complexity, they had reduced manufacturing costs on the system, and they were able to minimally maintain but also improve burner performance through optimization of the cooling of the system – it’s a very high-temperature system, incredibly high temperature; it’s actually higher than the melting temperature of the metals that make up the part. So, cooling was very important there and they were really able to optimize that cooling as well. So overall, a big success! And done entirely within our portfolio, which for them was a huge benefit on the integration of what they were able to do in their production CAD tool.

Boris Lauber: I think what I could also mention is quite an interesting example for the usage of generative design methodologies and generative engineering for the design of injection molding tools. We have a partnering service bureau who is using our tools for designing molding dies – and for injection molding, you can imagine we already have a kind of multidisciplinary requirement for these kinds of parts. So, on the one hand, we want to have lightweight parts, we have high-positioning velocities that need a certain accuracy. So, lightweighting and stiffness is a tough requirement here at that point.

Boris Lauber: On the other hand, if you do plastics injection molding, you, of course, want to optimize the throughput of parts. So, the parts have to be cooled down quite fast to be ejected from the mold. What helps here is the usage of conformal cooling at the area where you inject the plastics. This combination is a quite challenging system because you have, on the one hand, thermal and fluid flow requirements; on the other hand, it’s also a structural domain you have to consider here. They were using our technology, on the one hand, really, first of all, to optimize the channels for the cooling – so, having a real optimized cool flow for the certain part they wanted to produce at the end. In a sequential procedure, then going for the structural requirements – and having the manufacturing methodologies, like powder-based fusion additive manufacturing, these parts could also be produced, everything could be set up completely in the NX workflow. So, from the first design idea on the flow channel to the final part for the manufacturing, including all the support structures that are needed for the manufacturing type – that is something where all the interfaces in between are just reduced. So, we can really have a nearly fully associative way to come in from start to the end and doing changes in that design process. And that is really something that helps a lot using this kind of generative methods in that design process.

Jennifer Piper: Across the industry, I guess, what are the different ways that this technology is being leveraged?

Tod Parrella: I guess, in general, you see this as a tool used to optimize the performance of a system. That’s at least one area, and maybe I’d have to ask Boris to speak about opening up the design space of possibilities. But, from the optimization of the performance of a system, because the computing algorithms can really evaluate hundreds, if not thousands of alternatives of a design, where somebody really couldn’t do that using purely human engineers behind the wheel, it’s really been opening up the performance envelope, for example, in aerospace structures, where they can really eke out new lightweight, higher-performing parts, whether it’s in jet engine technology or actual aerostructures design.

Boris Lauber: Yeah, I see it exactly the same way as you, Tod. And I think it’s really a large variety of applications you can use this technology for. On the one hand, it can be really a single part design where you’re really going into a local performance and really optimizing heavily; on the other hand, it can be integration into a large assembly or sub-assembly to investigate this kind of influences. We saw, for example, also applications were working generatively on components in a whole assembly, where we tried to use the simulation techniques to transfer, for example, load path into a certain small component. So, we took a whole system to get the right boundary conditions – to overcome the problem that we mentioned at the beginning of the risk we have, what boundary conditions would we use to get the right responses at the end? And if we can use simulation to transfer the load in the larger system to the small party we’d like to investigate, that is really something that offers you a large variety of applications here.

Boris Lauber: Maybe another topic is also, especially if we talk about fluid flow performance, that is something that’s also a quite interesting application to use, a design finder topology optimization for fluid flow applications to really minimize, for example, for pressure drops in a system or to optimize for certain flow paths. We also have customers who are using this kind of technology to optimize their systems from an airflow perspective.

Jennifer Piper: Okay, thank you, Boris! It sounds like a lot of companies are using the technology in a lot of different ways, which is fantastic! So, why would a company be driven to consider Siemens NX over the other solutions that are currently on the market today?

Tod Parrella: As you know, Jen, our value proposition to all of our customers is our integrated set of tools and workflows within a single environment. We bring together the most comprehensive end to end capabilities than anybody else can offer, that we believe. I also see the NX as a CAD system is uniquely positioned with various specific and innovative capabilities that really helps this process. In this new world, we live in, we work across a multitude of different geometric design mediums – traditional CAD geometry, we work with polygon mesh geometry from CAE systems, and others. And you really need a tool that can handle all aspects of the design without having to go through multiple separate tools in the process. Again, our value proposition is we offer sort of the best of all worlds in a single production CAD environment.

Tod Parrella: So, our ability to easily take these engineering concepts that are generated using generative engineering methods, and transfer them and incorporate them into a usable design using tools like NX Convergent Modeling technology for working with mesh data, and NX Realize Shapes subdivision modeling for working with generating complex shapes and forms, really is unprecedented in the market, and it helps our customers incorporate these designs very easily into real-world, industrialized component and assembly designs. We see NX as the most productive modeler in the market. It’s one of our major pillars of investment and innovation for NX. And we see the benefit of being built on top of the most robust and ubiquitous modeling kernel in the industry – being Parasolid – not only in our ability to have robustness within our own system, but also data reuse and sharing across other Parasolid-based systems.

Boris Lauber: If we think about the offering we have as Siemens, it’s on the one hand, of course, the powerful CAD and design capabilities; on the other hand, that’s also connected well to our Simcenter and Simcenter’s 3D offering. Of course, if we think about simulating and doing simulations upfront to the design, you need to be able to capture the different requirements and to validate for those. So, our simulation technology that we provide really allows you to have a large number of different solving technologies coming from standard structural analysis to CFD – so fluid flow, thermal analysis, even going to acoustics, motion, fatigue life. So, there’s a lot of capabilities we have to capture the requirements and to get the right responses – and all that is integrated in a way that we can associatively change geometry and then automatically update meshes. And the connection between what Tod mentioned, for example, this hybrid convergent modeling to the simulation, is really handled in a seamless way. So, if you do any changes in the CAD design, you can update that on the simulation, and evaluate these designs. And that’s, of course, extremely important if you think about closing the loop for optimization and defining multi-objective targets. This is also equipped by engines that can really automate this process and capture responses and automatically changing parameters on the design to really come up with a very flexible way of designing any kind of optimization setup you would like to do here.

Jennifer Piper: This kind of design provides a suite of tools at the hands of customers. The possibilities of innovation are vast with generative design.

Tod Parrella: Often, design today is starting from something that’s already been created, already been developed. If you’re an aero engineer, you’re starting probably from the last airframe model that you built off of and you’re trying to optimize it or increase its performance. There’s a limit to working in that way. If you’re always starting from something very well known, you tend to already be focused down on what you can and can’t do. When you open up the exploration of the design possibilities, as we mentioned, you take this different approach as being sort of an orchestrator to allow the generative algorithms to do some of this investigation for you. We see that as an opportunity to really innovate in new ways, to explore design spaces that maybe you never had time or resources to explore before, to leverage, as we mentioned, new manufacturing methodologies and possibilities to say, “Well, maybe that design was left on the drawing table because we simply couldn’t manufacture it using traditional methods. But maybe we can reconsider an entirely new design, taking into consideration that we’re going to use additive manufacturing to produce this part.” Really, we see the possibilities of innovation as endless. In fact, I’d say it’s a renaissance of being able to innovate in new ways that maybe was a challenge up until the advent of generative engineering.

Jennifer Piper: So, some of our listeners may have heard a lot of buzz about how Siemens is really kind of changing the game. And I know that a new tagline was just announced a couple of months back now – where today meets tomorrow. Can you comment on how that theme ties in with generative engineering?

Boris Lauber: Especially if we talk about generative engineering, I think we have a lot of technologies in place today. I’d say many pieces in the game are existing maybe even for longer times, but bringing all together in an environment that is really combining the different tools you have and linking them associatively, that is really something that brings up a completely new way of working and a new way of engineering. And I would also mention, maybe, it’s not only the tool itself to do the work, but it’s also capturing results in a comprehensive way in a data backbone. Typically, we talk a lot about NX and Simcenter and Teamcenter is, of course, a kind of data backbone that will really help to judge all the design variants and all the engineering processes of a team. Because we mentioned at the beginning that the human is also, of course, an important part in that process. And, of course, all the data that you generate somehow has to be provided in a way that all the engineers and the teams can really work on that in a collaborative way. So I would say, from my perspective, is somehow taking maybe existing tools and new developments and bringing them on a new next-generation platform.

Jennifer Piper: Okay, and how about the future of technology? We have to keep ahead of the curve – which we talk a lot about – but what are some of the possibilities for future function and perhaps the capability of generative engineering?

Tod Parrella: If you look at generative engineering, we classify generative engineering as a form of artificial intelligence where, again, we’re letting computing algorithms that humans have written – letting computer algorithms trying to solve complex problems – but in order to do that, they need to leverage other concepts like machine learning. Whereas Boris alluded to the wealth of data that computers can generate in terms of useful results, but actually not only generate that as output for humans to interact with but also have the algorithm try to learn and adapt to its findings. Doing that, in cooperation with human interaction, I really see that as the next step of innovation in generative engineering, where the algorithms begin to be able to learn from their results, learn from their past designs, successes, and failures. Once we have that fully in place, that will be the next level or the next game-changer in engineering and design. The computer becomes much more powerful and can provide us with further useful opportunities and further guidance on how to improve the designs of the future.

Boris Lauber: Tod summarized it quite well – and having this kind of artificial intelligence is really a game-changer in computer power and all the data that is collected to use artificial intelligence. Because, of course, you always need this kind of decision-making tools, you need some basis and some data you can rely on. And nowadays, we really start to have this kind of toolsets and the power to store all this information. And maybe we do not yet know what we can take out, but there will be new algorithms coming into the future taking new ideas and really bringing up this kind of stuff into the engineering world. So, that’s exactly where it will go in the future, from my perspective.

Jennifer Piper: I guess the last million-dollar question that I have for today’s discussion is, if you had one suggestion to companies who aren’t yet leveraging this technology in their process, what would it be?

Boris Lauber: My suggestion would really be, if you’re not yet into that technology, you don’t have to cover everything at once. But you should start as soon as possible to get this kind of technology into your processes because I would say it’s not only buying a tool and installing it, like a Microsoft Office or something like that; it’s a really complex engineering environment. And starting with that, you also will have to incorporate all the different engineering departments. Maybe you will also have to rearrange your company structure in order to bring all the technology in place. And, I remember from my first days when I was working in topology optimization and development, we always talked to customers that they should front-load the simulation, but the processes and the companies were not at all set to do that. So, it was really discussing a lot in that area as well. So, now people are doing that, and I think that’s quite important to start with that stuff and really to get it in place in the engineering processes.

Tod Parrella: I would add to that, just many companies look at new technology and they may fail in the application of it because they look at it as a silver bullet solution or they don’t really think upfront about how or where would be the best area to apply this new technology to what they do. Failure to do that can result in a failure to apply something that you could have really benefited from, let’s just put it that way. When I see it happening out there, what I would suggest to others is to look at areas in your process or areas in your design, which are big challenges for you today, that take a lot of time and effort and resources and money to solve and look at if you could apply a generative engineering approach to those particularly challenging areas. And to not look at it as a wholesale holistic solution to everything that you do in design, but to really identify where you think your biggest challenges are, and where you think maybe you could apply this for the biggest benefit. That’s basically what I would say – think hard, do your homework upfront into where it could be applied and to consider it, as Boris mentioned, maybe at a small scale, using a proof of concept, proof of prototype, and then take it from there. Take small bites. We have a saying in our world, in development: “Let’s not boil the ocean, but let’s take small bits of success and build on that success towards you further success.”

Jennifer Piper: Thanks to Tod Parrella and Boris Lauber for giving us some insight into the workings of generative design.

Jennifer Piper: Siemens Digital Industries Software is driving transformation to enable a digital enterprise where engineering, manufacturing, and electronics design meet tomorrow. Our Xcelerator Portfolio helps companies of all sizes create and leverage digital twins which provide organizations with new insights, opportunities, and levels of automation to drive innovation. For more information on Siemens Digital Industries Software products and services, visit www.sw.siemens.com or you can also follow us on LinkedIn Twitter, Facebook, and Instagram. Siemens Digital Industries Software – where today meets tomorrow!