MBDA: Engineering Complex Weapons Systems for Modern Conflict

Modern conflict is evolving rapidly. How is the defense industry keeping pace? In this special bonus episode of The Third Angle in conversation, we visit MBDA's UK headquarters in Stevenage to discover how this European leader in complex weapons systems is meeting aggressive targets to cut program timelines by up to 50%.

MBDA employs over 19,000 people globally, designing advanced missile and defense technologies with more than 50 years of expertise. Host Paul Haimes sits down with Matt Beaumont, Senior Vice President for Mechanical Engineering at MBDA, inside their digital battle space facility.

Matt's philosophy is clear: "more haste, less speed." Cutting cycle times means maximizing valuable design time by eliminating non-value-added tasks. Operating in a single source-of-truth environment allows teams to work faster and with confidence, giving engineers time to get designs right with production and the entire value chain in mind. This approach has led them to double production output between 2023 and year-end, far beyond the industry-standard 10% annual growth. 

Throughout the conversation, one theme emerges: trust in data. Single source-of-truth enables parallel development streams where multiple disciplines work from continuously updating models rather than serial baseline handoffs, and it's the foundation for AI's future competitive advantage.

Find out more about MBDA here.

Find out more about Windchill here.

Find out more about Creo here.

Your host is Paul Haimes from industrial software company PTC.

Episodes are released bi-weekly. Follow us on LinkedIn and X for updates.

This is an 18Sixty production for PTC. Executive Producer is Jacqui Cook.

Transcript

Hello, I'm Paul Haines from industrial software company PTC. And welcome to the first of a series of some special bonus episodes of The Third Angle in conversation. Modern conflict is changing fast. So how is the defense industry evolving to keep pace? Today I'm at MBDA’s UK headquarters in Stevenage to find out.

MBDA is a European leader in complex weapons systems, with more than 50 years of expertise and a global workforce of over 19,000 people, dedicated to designing advanced missile and defense technologies. I'm outside their digital battle space facility, where innovation, technology and real world defense capability come together.

I've already had the chance to look around and see the teams here at work, but now I'm off to meet Matt Beaumont, Senior Vice President for mechanical engineering at MBDA.

So, Matt- it is wonderful to sit down with you. Thank you also for the tour of the facility earlier. I have a thousand questions based on what you were able to explain; the complexity of what MBDA do on a day to day basis, and the use of our tools to help support the business. Let me start with my first question. MBDA has some really aggressive targets to cut program timelines by anything up to 50%.

I'm curious, how are the engineering and manufacturing capabilities evolving to meet those standards all while keeping quality at the level that you need. 

So my perspective on this is very much it's about more haste, less speed. There's clearly a lot of pressure in the environment to do things faster. You have to look at the fastest route to the end game, which for me in my role, is about delivering valuable design time to the engineers. 

If you want to reduce cycle times across the board it's about reducing the non value added tasks. It's about reducing lead times but about maximizing the amount of time the engineers have got to get the job right. You'd love them to get it right the first time. The reality is they probably won't. But the more time you can give them to engineer the product with production in mind, with all of the other value chains in mind the better the product will be.

If you operate in a single source of truth environment where you know that the data you're using is being passed between different areas reliably and effectively, you can work faster.

And linked to that, so the embedding of as much information as you can put into the model, and then having a model based approach throughout the business, things like control characteristics and ensuring that they're going through and available to every department that needs them. How does that work in terms of perhaps some of the security sort of classifications that MBDA have to work with, and perhaps ways in which different engineers are working on alternative components, but working across different layers of security clearance?

So security is paramount in the industry to work in. If you employ good security principles as a culture inside your business, and then couple that with tools that allow you to containerize information that have very good access controls, and you marry two together, you end up with a culture where need to know principles apply. Where you have access to the data that you need rather than the data that you want. And you have a way of controlling who has access to that data. 

Then you can start to navigate the toolset in the right way inside the business. So an example of that, if you think about a technology that starts at low TRL, so low TRL technology is pretty much shareable. It's when you start getting into the application specific areas, that's when you start deploying technology on programs, when we start to get more controls in place around who can see what. 

So if you're operating in an R&D environment around MBDA, we try and share as much as we possibly can at low TRL. So one of the founding principles of MBDA is we act as a sovereign company in our domestic nations, but we also have an eye on cooperation. We lead in Europe on cooperation. So being able to be that national asset and protect sovereignty, yet also extend that principle to when the government say you can now share this, to be able to share that at pace and engineer at pace, means when we've got control of the data. We understand where the data is, who can have access to what. 

We can almost with a flick of a field, in Windchill, for instance, we can then say we now have a license in place. This data is now shareable. This can now be synchronized with this PLM instance in one of the other national companies. And all of a sudden you open up the engineering sphere to work to a cooperative model.

We can use tools like Windchill in cooperation with our culture and security policies, to really rapidly scale from a shareable idea to a sovereign program.

With the importance of collaboration for the engineers and the ability for you to hand as much time to them to deliver value added work. That ability to collaborate, to be able to switch on those and share data across the networks. That's obviously fundamental to MBDA.

It is fundamental. If I look back over my career, 20 years ago when we started to work internationally and when we would be cooperating with other countries to share information, you would burn information onto a disk, you would get an export license from the registry. You would travel overseas. You go to the registry. It was a very time consuming process. 

These days with the integrated PLM instances we have across group and with all of the data controls we have in place. When a program becomes shareable, when the license is available, we can almost instantly use the synchronization functionality within the PLM tool, Matrix in this instance, or Windchill to you, to then share that information.

So almost instantaneously now with the same controls in place as we had previously, we can really cut down the time it takes to share that information.

I'd like to return to this idea of technology readiness levels or TRLs, and link that to a discussion around simulation and the importance of simulation in MBDA’s development process. Because clearly, over the last 15 or 20 years, things have come on enormously. But perhaps you could just talk us through how you use simulation, how that's been able to drive better throughput for the business and, less prototyping perhaps?

Certainly. So simulation is at the heart of what we do within MBDA. When I look at what we talk about with simulation these days, with the advent of things like digital twins, it makes me smile because having been in MBDA 25 years, what I now understand to be a digital twin, we've been doing for probably the best part of 40 years.

Everything we do has an element of simulation associated with it, whether it's multiphysics simulations, whether it's system simulations. All of our products are built fundamentally, off a simulation background. And what that means is as tools become available that can use those simulations in different areas, as tools become more connected and you can start to share the evidence of things like single source of truth, the system becomes really, really powerful.

So every single one of our products has a very advanced simulation that sits behind it. We can now connect those simulations with other parts of the business. That again reduces the translation time. It reduces any human error that goes into it. More fundamentally, the approach we take with our products, we talk about how do we get products in service faster, how do we get products developed faster?

One of the slowest parts of any development is hardware lead times. So you can do a huge amount within the modeling tools to accelerate development. But you're quite often sat waiting for hardware to come in. Now, modern rapid development techniques are clearly very useful. They may not be what you want in a production environment, so ultimately you can use rapid prototyping techniques to get hardware very quickly into a development program, you need to have a very close eye on what you're going to do for production as well.

So the use of simulation tools, the use of multiphysics models, the structural models that we have. So you're building a set of really progressive assurance as you go through a program, understanding at each stage what the designer intended to do, what the product needs to do, what the part that you've manufactured is capable of doing as you move through the development lifecycle means that you can start to use these different tools and techniques, interweaving simulation with rapid prototyping, rapid development.

Ultimately, with that goal of getting the production article done as quickly as you can.

And we were discussing earlier, Matt, the importance of that known environment where you have the simulation because the reality is your products are in some extreme environments in which they work in. A missile could be strapped to the underneath of an aircraft flying at 50,000ft, and it is subjected to freezing conditions where things are icing up.

But then literally at the flick of a switch, that missile needs to be operational, and then you've got a propulsion system that's putting maybe 1000 degrees at the back end of the missile, and the thermal stress that's going through everything in that missile at that point, and the aerodynamic stresses- it's an extraordinary engineering exercise that MBDA go through.

It is. And each of our products is very finely tuned. So if you go back to what I said previously. It's the ability then to have a very, very clear design record of what the product is, what it's being qualified to. Knowing with a high degree of certainty what it's able to do and possibly how far you could push it, means that when a new project comes along and says, we need a really quick solution here, what have you got? Quite quickly, we can roll out 4 or 5 different product classes of a particular equipment. 

I apologize, I'm being very generic here, but you can kind of understand the constraints we're working in. And these days, it means we can very rapidly adapt things for rapid development, for rapid capability.

Matt, I’d like to turn to, obviously the area of technology that's probably growing the most right now across all aspects of life, and that is artificial intelligence. And in particular, maybe start with a discussion around how artificial intelligence, how you see it impacting the work of the engineers inside MBDA.

So AI is a huge topic. If I look at the toolset that's out there, it infuses me. If you look at the potential of it and it's really, really interesting. At the same time, it frightens me because, maybe I'm a control freak, but I like to know the provenance of the engineering that we've done. I like to know if you look at what we do as a functional organization inside MBDA, we spend a lot of time focused on the skills and the competencies of individuals. So to trust a computer to do that is really quite a big, big leap for me.

But ultimately, if you go back to, I guess, one of my anchors, I said at the start of this, is that anchor of to do things better and faster- it's about more valuable design time. Then maybe there's an opportunity for AI to take some of those non value added tasks away. You could see an example like the cost of change.

The engineering process is quite strict in terms of the way you manage modifications, particularly when we're talking about airworthiness and products that fly and are attached to aircraft and things like that. So control over changes is key. But wouldn't it be nice to understand better the cost of change if you're working through a development program? We've already talked about the need when you introduce hardware into that. 

If you make an engineering change you've got hardware set around you, the cost impact of that change could be quite significant. It can be very hard for an engineer to understand the total cost of that, but you could see some sort of AI bot go into the SAP system, interrogate the system, find out how much stock is around and generate a picture as to - you can make this change, but the cost impact might be this. That's a hugely valuable tool for the engineer to know, because that could take them a huge amount of time to find out. And again, I would like the engineers to be spending time designing and engineering products, not spending time trying to find information to help them make decisions.

So I can see AI really augmenting the change control process in that way. In other areas, we've played around with generative AI solutions as well, to with quite a degree of success I'd say. We've got one particular part, it's a reuse scenario again, where we've got a project come forward and said we'd like to reuse this particular piece of equipment, but it needs to do all of this.

And one of the key components was the system wanted the mass reducing, but it wanted the performance to be higher. So we needed a chassis being designed for this particular part that was both lighter, stiffer, you know, the magic mix of things that you rarely get. And we looked at a set of different tools. We took an initial concept for a chassis, put it through a topological optimization, put it through a general AI design tool. It was within Creo. And then actually then put it through the injection molding flow tools and came out with a product that's really quite different to what we've designed before.

That has limitations in terms of manufacturing techniques. So you're not going to machine something like that. You're into either additive manufacture or injection molding. And actually in that instance we went down the injection molding route. It's the learning from something like that is the part is exactly what we want it to be. The industrialization was harder. That injection molding tool, rather than taking 2 or 3 runs, took about five runs in the end to get it right, such as the complexity.

But we've now ended up with a part that meets the aspirations of the project. So we've got a delighted customer, our hands internally from something that we thought was going to be quite difficult to achieve, but we've actually done it and we've reached out to these new and different  types of tools to help us solve the problem.

It's a great example, I think, of where AI and in particular in this case, Matt, generative AI is finding its feet within industry because, as you rightly point out, the design was exactly as requested. The manufacturability was perhaps a little bit more challenging. Things will iterate and the solutions will get better.

I think from a PTC perspective, we very much see AI and the practical application of AI as key to our strategy for how we're going to implement it within our tools. It has to be secure and scalable and add real value to the engineer. So whether that's a capability to help, if you end up with a circular reference in your Creo model to how you unstitch that circular reference and put the geometry back together again in the right way.

Those are real value add tools that are going to, again, free up time for the engineer to do the work that’s needed.

I think they will. And what I do really like is the philosophy that was described to me when we visited Boston last year, the PTC headquarters. Where your approach to AI in terms of advise, assist, automate chimes really well with me. If I go back to my concerns again about what is the tool doing versus what's the engineer doing? Those three levels feel right to me. 

Of course, with what we're seeing with AI now, in my mind, I think we saw a few years ago with things like additive manufacture. So the notion of of making something in an additive method, which was quite far removed from our comfort zone of I understand the strength of my materials I'm using. I know how they've been made. They've been made to international standards. I'm machining metal away. I'm in control of that process to now I'm making that material in situ in a machine that I don't necessarily know or trust, and I want to use that part in an airworthy application.

That seemed like a hurdle at the time. Several years on, and we're now employing parts that have been made with additive manufacturing techniques in missiles, in structural parts that are flying around the skies today. So maybe, AI is going to be the same. It's a very new shiny thing today that looks promising on all fronts. But the nervousness we have is that control element of it. Maybe over time we'll develop techniques as to how we do control it, how we deploy it in a logical, sensible, risk based way. 

That again, brings us back to I want the designer to focus on design time, and I want everything else around to be as easy for them as it can be. But ultimately, this can't be about deskilling the engineer. The key thing for me, from a designer point of view is that the designer needs to understand the part they've designed, the design intent all the way through to how the parts are manufactured. So we go back to things like the digital industrialization platform we're putting in place, MPMLink, which is incredibly valuable for us.

You've now got the designer designing something with the ultimate goal in mind of the production part. Thinking as they are designed about how they need to control the part. What the key characteristics are. What the critical items are. How that manifests itself in control characteristics that will go into manufacturing either in our MBDA manufacturing environment or into the supply chain.

That's the skill of the designer.

It's a great topic, Matt, that. Because that ability to collapse the boundaries between engineering and manufacturing by embedding the designer's thinking into the model, that is then through MPMLink, inside Windchill, actually showing up on the shop floor so that your assembly technician actually has a clear set of instructions specific to that particular variant or instance. That level of assurance, that level of fidelity is, I think, a tremendous asset.

It's really important we have in an aerospace product you have certain, what we call grade A components, defined by a biodefense standard in the UK which have airworthiness attributes about them. So these are parts that are structurally significant that if they fail may lead to catastrophic results elsewhere.

The ability to control those parts through manufacture is key to our success. They're very difficult parts to design. Typically, they're very difficult parts to industrialize, particularly down to the control characteristics that are the part of that design. So something like MPMLink, the digital industrialization platform that we have that gives you that bidirectional link between the design bomb, the design intent and the manufacturing execution system, whether it be in an area where you're assembling a part or whether it be in a machine shop where you're machining a part, gives us that trust and confidence that what we're doing is right. And that can only be a good thing. The ability to use Creo Illustrate to manipulate the data, to give the operator better views of the part is really important. But those control characteristics, we know what the design intent was.

We know why the designer has done that. We know what's really critical. We come to inspecting the parts and making sure it's right. The flow down of that in a controlled way is bound to give us the quality we need out of the supply chain. And ultimately, when it comes to, we've talked a lot about how you develop products faster. We also want to produce products faster. We're in an environment where our customers want more, and if you want to double the rate of something, you have to build the quality into the process.

As you increase production rate, what you actually want to do is to decrease the number of defects that you're getting. If you don't do that, you'll end up with a quality system that's just out of control. So as we increase the production rate, our tolerance to bad parts arriving in the factory, things that won't fit together, things that won't integrate is significantly reduced.

So we need to operate in an environment where our supply chain is only providing good parts into our factories. And the way that we'll do that is by having control characteristics into the supply chain, making sure the industrialization has been done properly, really good surveillance from our quality team to make sure that our suppliers are operating in that environment.

And then you'll get good parts into a good factory. And when you push on that system and you want them to produce more, they will. And that's a lot of the success that we've had over the last few years. If I look at some of the years, we've doubled the production rate over the last five years. So we've talked a lot about development and rapid development, but actually as an industry, rapid production is also where our customers need us to be at a minute.

If we look at the pressures we get from our customers in terms of generating more products out the door, you have to do that in a controlled way. 

What we found really useful is what we call a digital industrialization platform, where we've used MPMLink to link the design, build materials to the manufacturing build of materials, and the manufacturing execution system to give us that flow down of control characteristics.

So this is taking the, it's really a very good way of taking the design intent of the designer and being able to articulate that very, very clearly to the people, actually making the parts or assembling the missiles on the shop floor. And that two way link means that we understand precisely what they're doing. They understand our design intent as well. But we're also able to manipulate things like CAD using Creo Illustrate now so we can very easily give them good model views of parts. Whereas previously they may have had a diagram that wasn't particularly clear. Now we can manipulate the model really easily.

And of course when the model changes, that update comes through automatically, something is flagged as being changed. So with that bidirectional link, when things change and things will change in the environment, we may have a part that's gone obsolete that we need to rapidly change. When we introduced that part in, the MPMLink updates all of that information to the manufacturer environment.

So these tools have really helped us accelerate what we're doing. And we're having great results. Between 2023 and the back end of last year, we doubled the production output within MBDA. Now, most companies would maybe think 10% a year was good. But to double that manufacturing output in two years shows you the phenomenal effort our teams are putting into responding to that pressure, but also that we're able to do that because we've built a system that we can trust, that we can put pressure into.

Yeah. And that system, I mean, as you say, 10% is considered good in the industry. What you've done is, is way, way beyond that. Way beyond that.

But we've had to. 

Yes. Yes. And Windchill, Creo, the digital systems that you've got in place are really the foundations for being able to scale at that pace.

Absolutely. Because it comes back to, I keep saying it because it's important, trusting the data. Single source of truth. You know you're reading off the same data. Whether it's leaning on a manufacturing system, whether it's rapidly designing a product. One of the hardest tasks within mechanical engineering is things like airframe design, where you are wrestling with the fact that you've got equipments that are maturing at pace, they're changing, their interfaces are changing.

If an equipment doesn't work, it might get bigger, it might get smaller, it might get hotter. There may be something else going on with it you need to understand.

The ability to have a single source of truth and have that model the same model you're using to integrate all those equipments, then being featured and used by aerodynamics as an outer mold line for their studies. Having your missile mass properties maintained in a continuous way, such that guidance and control and flight control algorithm areas are using the right data, means that you can now break your product development into a series of parallel streams. 

Whereas previously it would have been a very serial process. You would baseline something, hand over the aerodynamic configuration, the missile mass properties to aerodynamics, to guidance and control. They would go away and work on it, bring it back to you. But in the meantime one of your equipments may have changed and you're constantly wrestling with things changing around.

In this digitally continuous environment now, you can change things on the fly. The models are updating. Everybody's working off the same data. It does allow you to push that pace.

It does allow you to inject more parallelism into your development programs, but with the trust in the data that you've got. No one thing makes a real difference, but the collection of doing all of that allows you to engineer more at pace.

And that, you use the term trust in data, which I think is again, something that is fundamental for the success you're seeing today. But the trust in the data is what's going to drive your success in the future, because AI clearly relies on a foundation of data within the business. And so that degree of data that you're assembling today is really the foundation for the competitiveness in the future. I would say.

Absolutely. I mean, the AI, ultimately, if you think about the data that's stored within our tool, we've got a very long heritage to business. You know, 60, 70 years just in the UK of product domain knowledge. All of that stored in a business management system in technical libraries.

Which we try and make as accessible as we can to all the new engineers that come through. But that's really hard. The ability to use AI to help them find the right data in the right context, it's going to be really important. As a business, we're recruiting 2500 people across our European enterprise or global enterprise effectively every year. That's a huge growth. A lot of people to onboard. To be able to orientate them towards the right data is really useful. 

The other thing I think it could be really useful for is, one of the downsides of having all of this data is that it can constrain you in your thinking. So you bring a new engineer in and you look at the new techniques that are available now both in manufacturing and design. Some of the lessons of the past may not be as relevant as they were then, as they are today. If you have a set of design standards inside of business that actually give you a level of safety in what you're doing, your design standards almost mop up the lessons of the past and help the new engineers with, well, we've done that before, it didn't work. 

But they can also constrain you. So when you try and operate in an innovative environment, having a set of standards around you that almost handcuff the engineer, says you can only operate inside these parameters can be really constraining when you want to do innovation. So I like the idea that maybe AI could be used there, that says it can help bring that huge amount of data to life with engineers. It can maybe apply it in a context that allows them to innovate as well as constrain them.

Yeah. And I think it heads towards a future with a more intelligent product lifecycle, which is really where PTC are heading. But I think it has been a fantastic example of where these tool sets have driven real value in an incredibly complex engineering environment and one which MBDA are clearly excelling in.

I think it's been a fascinating conversation Matt. Thank you so much for your time today.

Thank you.

Thanks to my guest, Matt, and to MBDA for allowing us to visit their state of the art digital battle space facility here in Stevenage. For more information about PTC solutions, visit our website at PTC.com.

Please rate, review and subscribe to The Third Angle wherever you listen to your podcasts, and follow PTC on LinkedIn and X for future episodes.