Harnessing the Benefits of Cloud-based Simulation

Adriano: Yeah of course. So in SIMULIA, we make software solutions that allow our customers to model and evaluate real-world conditions virtually on, you know, a computer workstation or on high-performance computers. Specifically, I work within the strategy team at SIMULIA and my focus is on how best to integrate the benefits of cloud computing into our solutions for our customers.

Now, how I became involved in the field of simulation, and it goes back to, really, a young age when I've always been kind of fascinated by computer technology and the kind of massive scope the virtual world provided us, or provided me, as a young child, to create new things. 

You know, I've always had an aptitude with computers and when I was about eight or nine, I received my first computer, which is an old Atari 65 XE, for the retro middle-aged dads out there. And I used to write simple programs on it. It was never a fantastic experience, it could take an hour to load any games, for example. But it never tempered my fascination. You know, for me, it was always an opportunity to create things we do in the real world. And the limit was, you know, ultimately my imagination.

And that led on to essentially, through secondary education, to my choice on going to university to study engineering. However, when it came to simulation itself, I only really discovered it after I completed my degree. You know, doing a bit of research, I came to learn about computational fluid dynamics, which is a field of simulation that allows us to model the motion and behavior of fluids and gases.

This really piqued my interest, so I started looking around for PhDs and luckily an opportunity arose at Glasgow University. So here I am, my PhD covered applying high-lift devices to helicopter rotors. And from there, my career spans multiple industries applying CFD, mainly in aerodynamic applications, but also through to developing methods and processes.

And today, I'm now at Dassault Systèmes. I've landed at Dassault Systèmes’ SIMULIA, and now I rather focus on the application of simulation directly to real-world problems, I look to investigate the different ways we can make simulation easier by leveraging cloud computing.

Arvind: Well, let's dive into our topic.

Adriano: So, the path to creating a new product is very challenging. And, as much as anyone thinking about concept, you want to create how it looks, how it works, how it will handle, how it will feel. There's still a long way to go from an initial product definition to actually having something physical that, you know, a customer will want to buy. 

And for me, the biggest challenge throughout this process is how a company is best able to manage the risks for developing a product whilst delivering on the user values they want their product to offer.

And the first big challenge is you ultimately have different groups with different specializations and working on different components. That's all meant to come together to a single product. And the challenge for these different groups is, you know, they're trying to develop from an initial concept where there are multiple competing objectives they need to keep on top of.

They need to keep control of their costs and stay within budget. They need to make sure they have confidence in the tools they're using to generate the data. And so they're making the right choices throughout that process. You know, it's about having the right information at the right time and to make the right decisions, not getting trapped in a silo, not losing awareness of, you know, what they're doing and how it's impacting other parts of the product.

And there's this kind of misnomer that the best solution for their component may not actually be the best choice for the overall product. So there's a lot of challenges there that need to meet. Then if you think about when you have to integrate these components into a single product, there are further issues.

How do you ensure that what each team's doing is going to fit within the overall product strategy? And how do you ensure that the final product is going to deliver on the expected customer value? Is it going to meet any kind of safety, robustness profiles? And how can you get it to market faster than the competition without necessarily sacrificing quality?

So, you know, risk is prevalent throughout the entire process. And for me, the companies who manage that risk the best whilst delivering on the key values they want their customers to access are most likely to be successful. And this is where simulation, and 3DEXPERIENCE, plays a key role in helping these businesses be successful.

Arvind: A couple of takeaways that I found to be quite interesting is the competing requirements that different teams face. And then each of these teams produce their masterpiece and it has to come together and the final product has to meet its own set of requirements. I think that's a challenge by itself.

I think it will be fascinating to see how simulation plays a key role in helping companies meet these objectives. Now, with that, we can lean into simulation now.

Adriano: For me, it really comes down to what simulation means because it can mean a lot of things these days. You do hear the word used a lot. For SIMULIA – our kind of software solutions, are designed to provide physics-based simulations, and these have been continuously validated against real-world problems for decades and they’re heavily ingrained already into the product development processes of many of the top OEMs around the globe.

The reason for that is they have confidence in our solutions as a tool for effectively managing risk through their design programs, whilst being able to deliver on actual solutions to problems. Some examples are, they're able to support the detection of problems very early on in design proposals. That helps kind of prevent wasted efforts and it limits the real risk of detecting problems very late on and incurring some fairly heavy costs.

We enable customers to propose new design solutions by making it easy to simulate multiple variants of a single product without needing to go out and actually build and test, you know, real-world versions or prototypes every single time.

We also allow customers to get a deeper understanding of how their product works. And that in itself can then help guide the future design direction the company takes for a particular product. And in the end, you know, as I mentioned for the previous question, it's all about providing the right data for a business to make the informed decisions to make a successful product. And that's really what simulation can add to the overall product development process. 

And the beauty of simulation is we have different solutions for different types of problems, that enables customers to do that. 

There are quite a few fields in simulation. The most dominant one and most prominent one is structures, or finite element analysis, and that sees widespread use across all industries, whether it's a company looking at virtual crash testing of road vehicles or simulating a product packaging process, structures allows designers and engineers to answer some of these big questions when it comes to designing a kind of new product.

And I have a couple of examples. If you imagine a company looking at designing a new mobile phone design, one of the big questions, the impacts that they need to answer, that impacts us as consumers is, you know, if the phones drop, is it going to survive the impact with the floor without getting damaged or cracking the screen?

There's also a big focus on sustainability these days. And if you look at manufacturers and the packaging goods sector, you look at when they're designing new plastic bottles, how sustainable can they make them whilst making sure they stay robust? So if they end up having this plastic bottle design and they want to try and thin the walls of the bottles to use less material, is it going to retain structural integrity? You know, when it's in transport, when it's filled, when a customer drops it and obviously customers have certain product expectations, they don't expect bottles to bust and to have to clean up all the goo all over the floor.

But there's obviously many other opportunities where really understanding the structural response is really important to making sure the product you're creating is going to be successful and a robust product for the market.

Next, we have Computational Fluid Dynamics, which is my particular background. And this is one that's maybe more visible to the average user. I'm sure everyone kind of follows Formula 1, and they've maybe seen the kind of fancy videos of the lines kind of flowing over the car, or the streamlines with nice colors on the streamlines.

You know, that's some of the type of visualization we do with CFD to understand what's happening around the car. But CFD isn't just limited to aerodynamic analysis alone. There's many things and many problems you can solve with CFD, such as modeling the blood flow through arteries, trying to design effective cooling for electronic circuit boards, understanding the noise signature of air con units in industrial buildings. Even trying to assess the visibility of a car as it drives through the rain. All these things that have direct impact on a user, on the environment. 

There are three other domains that we can support. There's electromagnetics. It's where we aim to understand electromagnetic properties of devices. You know, the flow of currents across circuit boards or the performance of antennas.

There is motion or multibody dynamics, where we look at the complex interaction of multiple bodies under the influence of mechanical forces. And then there's vibroacoustics, where we look to simulate noise and vibrations of objects across the audible frequency range.

And ultimately, introducing any of these physics-based simulation tools into a product development process is going to be a really kind of strong step to actually de-risking that process. But, although it's kind of important to understand what you can use, it's also important to understand when to use it and my general guidance is, we should be simulating more and we should be simulating earlier.

And to understand that we have to kind of first take a quick look at the overall product development life cycle. You can really break it down into three main stages. There's a concept phase where you have teams working together to try out early design concepts, explore the design space, and look to really refine on the key performance indicators they're going to target for their particular product.

There's then the core development phase, and this is really a period of rigorous product refinement where they really try to narrow down on the specific performance they're going to get out of the product.

And then once that's done, there's a validation phase. And here you do tend to have a really high-quality physical prototype. And obviously there will be a high-quality virtual twin of that. And here you're aiming to ensure that the product performance of what would be a manufactured product is going to match the specification that you chose or designed at the very beginning of the entire life cycle. 

And through that process, traditionally, there would be a lot of physical prototypes required to really have a look at the product itself and assess whether it's doing what it's meant to be doing. You know, physical prototypes, they do generate a lot of cost to the business. They're not cheap to make and they take time to build. There's also an environmental impact of having so much potential waste material to consider. And if you take all these factors together, you end up really being limited by how many physical prototypes you can actually build in a given period.

So without simulation, although you would require more physical prototypes to get a proper understanding of your design space, there's actually less opportunity to do so because of the fact that you're having to wait more, there's cost impacts. 

As I've kind of said before, the less data you have available, you may take more time to make the right decisions and this can lead to greater risk of delays to get your product to market. And the real risk of delays and bad information when you're creating a product is it can bring unexpected engineering costs very late on. And at the really worst case, it can mean you're not meeting regulatory requirements, which is a huge cost to a large business. 

So the key message for me is, the earlier you integrate physics-based simulation into your design process, the sooner its benefits really can be delivered. It allows designers, engineers to refine target designs by being able to look at many different design concepts early on, on a virtual model. And it also helps reduce the number of physical prototypes required to validate a design and potentially, significantly shortening the development cycle themselves. 

Arvind: You know, so far we have talked about simulation and doing simulation. Nowadays, you can do it over the cloud.

Adriano: The big challenge that's always been faced by physics-based simulation is really making it accessible. You know, there's really been three key factors that have impacted accessibility of simulation for businesses.

The first has been having the right engineering skill sets to set up simulations. Generally, you'd require people with degrees, if not masters or PhDs, to execute the simulation process. Then there's accessing the right computing power to run the simulations. Yes, you can do simulation with small workstations, but as you get bigger and you start requiring massive high-performance computing installations to do so.

And the kind of third point is really having the right amount of time to fit all this in before you face the next business decision. And cloud computing, you know, of all these three factors, cloud computing has had the biggest impact today on the second one, having the right computing power available for organizations, as you mentioned at the start.

Simulation would normally require quite a significant amount of capital investment by a business to get started. For a local workstation, we'll give an example of roughly 64 cores, 128 gigs of RAM, could cost thousands of pounds. It would be enough maybe to run some small to medium-sized component level analyses. 

You know, it's this kind of small part, but as soon as you start getting into larger sub-assemblies or even system-level analyses, such as modeling the full car or modeling a full aircraft, then you start needing to access really high-performance computing infrastructures. And here we're talking about hardware in the hundreds, if not thousands of CPU cores. And you need that to make sure you can handle the size of the models and make sure you can get results back in a timely fashion, because time is the enemy here as well.

And here you go from a thousands of pounds of investment, to hundreds of thousands, if not millions. So it is a significant outlay for businesses today. It does give businesses an element of control, but at the cost of less flexibility. So, there's always a risk of, you know, when you have your on-premise hardware environments, of becoming resource limited. 

That can be very common if there's multiple competing programs of work where everyone's chasing results at the same time. And they all try and run their simulations at the same time. And there's only so much hardware, so at some point, someone has to make a decision on who has higher priority, someone has to wait. Waiting means delaying decision-making. That can then impact the business by bringing on higher secondary costs, such as delaying time to market. 

One of the real risks of this situation is, users then may start to try and avoid running certain simulations that they maybe need or changing their simulation models to fit resources. It's rather than being able to simulate actually what they want, like simulating reality, let's say. And this risks impacting the overall fidelity of the physics-based simulations they're running. And that can actually have a serious impact on confidence in the data that's being generated. And here, the advent of cloud computing and infrastructure-as-a-service has kind of changed this landscape quite significantly. 

There's now an opportunity for businesses to outsource their high-performance computing demands to these much larger providers who can minimize their costs by using the economies of scale and at the same time enabling customers to get access to the latest hardware without requiring these large upfront and capital investments themselves.

And it can mean for a business that they get access to increased capacity whereas previously they were limited by their budget, which means customers can lift that capacity ceiling for simulation. It means they're no longer worrying or waiting for jobs to run. It also means they can get access to on-demand resources they can elastically adjust to the priorities of the business.

It basically means that businesses developing products can spend less time now managing their on-premise IT resources, less time worrying about, you know, when they're going to get the simulation results back in time. And instead, put more focus on really what they do best, which is designing innovative products for their customers.

Arvind: Maybe the next question you would be able to enlighten us as to how those things are woven into the world of simulation. I'm really talking about the 3DEXPERIENCE platform. And you mentioned that it has got simulation capabilities on the cloud. But when it really comes to cloud and cloud computing, people are thinking of the Amazon AWS.

Adriano: I mean, the 3DEXPERIENCE platform on the cloud is a different proposal altogether. First of all, it's a software-as-a-service offer. And that means that we at Dassault Systèmes deliver software solutions that support customers to develop better products without worrying about the infrastructure behind it, and it's all delivered – it's a nice kind of all-fits-one package.

In terms of the SaaS, it means that, for a simulation user, they can get access to the full suite of SIMULIA's simulation portfolio and the applications directly from the cloud. That's all delivered on top of our platform-as-a-service, where we basically provide the 3DEXPERIENCE platform, which is a platform designed from the outset for collaboration.

And not only do we deploy that, we supervise it and operate it for you. And this is all then hosted on our own infrastructure-as-a-service. And that means we make the servers, the storage, the computing resources to support the software and the platform available to businesses with elastic resources all over the internet without them having to get involved at all or get their hands dirty.

Now here, SIMULIA’s software suite does cover a full suite of applications across the simulation domains. Our cloud solution delivers additional benefits on top of what we do already in terms of software. The first core benefit is the 3DEXPERIENCE platform is a platform for modeling and simulation, or a term we commonly call MODSIM.

And MODSIM is really transforming the way that simulation is integrated into the product development processes today. Whereas previously businesses would operate with separate teams, you’d have your design team, you’d have your analyst team… where they'd be working with CAD files and sending CAD files over email or on the file management system and someone then has to kind of pick it up and load it and then write a report and the report maybe gets sent out through email… It all gets a bit messy. Traceability becomes very difficult. MODSIM and 3DEXPERIENCE cloud brings that all together into a single unified interface. 

So the link between, a designer's CAD model and the simulation model is now intrinsically connected. It means that the CAD input to your simulation is always up to date. And here, there's multiple benefits to doing this. 

The first thing is that the automation aspect is really at the heart of MODSIM. It means that making a change to a design concept no longer requires a repeat of the full simulation workflow or the need for domain-specific knowledge to access these benefits.

You know, a simulation model is now automatically updated and ready to simulate every time a change is made at the CAD level. And there's little to no user intervention required. There's no meshing, no additional setup in the simulation.

MODSIM is really democratizing the benefits of simulation. And we're putting it in the hand of designers who can assess the impact of their changes much earlier than they would previously.

At the same time, we're also facilitating how a product can be understood properly, using MODSIM again. So a designer can now understand the impact each potential change is making well in advance because they now have access, or easier access, to a design of experiments type analysis where you can actually take a component that they're working on. 

The example I've got is maybe a bootlid spoiler. It's fairly simple, the little kind of gurney thing. And it's on the boot of a car. And you can easily parameterize this now, so, break it down into its height, its width, its length. Submit a design of experiments where you adjust each certain parameter by a certain amount and then you can actually view what impact these combinations of changes are going to have.

And this is great because you know, whereas previously you would work to focus towards a single optimum solution, the beauty of DOE is, with MODSIM, is it starts to open a discussion and a conversation between engineers and designers on the impact that any of their changes can make and allow them to actually make a real informed decision of what's really going to work when it comes to integrating this particular part with the overall product.

So for 3DEXPERIENCE, MODSIM is really allowing companies to simulate earlier in the process. It's allowing the opportunity to add more value at every stage of the product design, plus freeing up the time of expert users. Because these expert users really need to be focusing on critical tasks that require their particular skills and not going through this and back and forth, just running the setup process for simulations. 

There are other benefits when coming to cloud and the next one is the access to high-performance cloud computing and this really is an integral part of the SIMULIA offer and 3DEXPERIENCE cloud as well.

So unlike, you know, standard cloud computing, where the hardware is available on cloud, but you still need an IT expert to be able to log in, to set it up, to maintain it, to supervise it – our own cloud computing solution is fully tuned for the simulations and workflows that we offer as a business.

We provide cloud-ready user licensing, and that's all attached to a scalable, on-demand compute that can elastically adjust to what businesses need to run and it allows businesses to really assess the right capacity they need for their current problems. And that means they can access as many cores as they need for their problem.

And if required, they can add more CPU cores if, say, they want to get more results back faster or run higher fidelity simulation models with larger meshes. They're able to run multiple design variants in parallel if they want to get a better coverage of a design space, for example, for doing a MODSIM design of experiments.

And this is all provided as part of the SaaS offer, where in around a day, and maybe slightly more, they're able to get access to this HPC capacity out the box with no specialist IT skills required to set anything up. 

The third big benefit is that our compute offer supports all the simulation and domains of SIMULIA and it provides two really big elements of flexibility.

The first is customers can access it using an on-demand model, which gives them immediate access to all the cloud compute options. And this is probably more equivalent to a pay-as-you-go model, like they may find in other providers. However, customers can also access a sustained capacity model with us, where it allows them to pre-purchase a certain number of cores for a year and consolidate their HPC investment up front.

And again, this is all provided in a single SaaS rate, as a software-as-a-service rate. It includes both the software and hardware costs. There's no extra costs to consider. And it's a really convenient way for customers to get access to simulation and the computing power to run it without having to invest in any on-premise hardware.

I do have one last point to raise and I think it's important to raise a topic that isn't classically associated with simulation but does impact how productive we become when developing products. And that's collaboration. And what the 3DEXPERIENCE platform is, is really a platform designed for collaboration from the outset.

It enables better collaboration across the product development process, because it can ensure customers can track and share the project progress in real time. It allows them to use dashboards that can be populated with all the data from the simulations to help support decision-making. It can enable better collaboration with third party stakeholders. Making sure everyone's working on the same product, on the same parts, following the same project timeline and ensuring you get proper traceability of all your data and decisions. 

There's lots of benefits to be had from 3DEXPERIENCE cloud simulation, whether it's for MODSIM, to the on-demand compute, to the collaboration. There's a lot of value there for businesses to ensure they can optimize how their product development process is working and making sure they can make the best product possible. And again, it's all available as a software-as-a-service offer, 3DEXPERIENCE cloud.

Arvind: When Lifecycle Insights did a study earlier last year, we found that typically, large enterprise customers have no problems using simulation. They've got teams to handle simulation, they've got experts like yourself in their team, in fact, many of them, and they are aware of the different tools and how they can use simulation. 

But when it comes to small and medium-sized businesses, they pretty much face the same challenge, if not more, as their larger peers in their industry.

Adriano: For me, these businesses, one of their core strengths is, when it comes to accessing simulation, is their size and maturity, especially for startups. If you think about what they are, they have a certain level of agility on how they operate and that agility really gives them this ability to pivot to new technologies and opportunities very well. And that ability to pivot and react to these opportunities really can also be translated into being able to easily integrate the best practices around physics-based simulation into their processes. 

And obviously working with SIMULIA on the 3DEXPERIENCE cloud is a great place to start.You know, we have the automation integration with MODSIM, the out-of-the-box access, the right cloud computing needs. It's never really been easier to get going with simulation. So my best advice is obviously simulate more, simulate earlier, and join us with 3DEXPERIENCE cloud simulation to do so. 

For the startups, we do actually have an offer available today called 3DEXPERIENCE for Startups. This allows them to access, 3DEXPERIENCE cloud at quite a low price, under 500€ per user per year, with discounts of up to 90% on some of our more advanced solutions. And there are some criteria to meet that: they have to be first-time customers of Dassault Systèmes, and you have to obviously be designing or manufacturing your own physical products.

We require them to be mentored by an accelerator or a, you know, venture capitalist. And the companies have to be up to five years old with up to five million in funding or revenue. However, if you're not a startup, then we also have a three-month cloud trial available for a small initial fee per user.

In both cases, users can get access to all the simulation software they need, plus all the cloud-based infrastructure to get them going with simulation. So either way, you know, we have smart people at Dassault Systèmes, we're ready to help you get set up and going with simulation. So, if simulation has really kind of hit a nerve with you and piqued your interest, then please get in touch with us. 

Arvind: Simulate more, simulate early, and explore the offers that 3DEXPERIENCE has for you, seems to be the mantra. Thank you, Adriano, we've had a fantastic discussion. 

It's unfortunate that we are running out of time. Otherwise, there's probably plenty of things that you would like to share with us about simulation and MODSIM and the 3DEXPERIENCE platform. But thank you very much for your time and hopefully see you in another podcast!

Adriano: Yeah, likewise. Hopefully speak to you soon. Bye, everyone. 

Clara: Thank you for listening to Designing Impactful Innovation. To find out more, go to 3ds.com/cloud. Don't forget to subscribe for more insights and stories from our guest experts.