 Boom, what's up everyone? Welcome to simulation. I'm your host Alan Sackian We are still at Kofa as the Congress on the future of engineering software for our second annual partnership with them We are now sitting down with Christian Thompson. Hello. Hey, pleasure to meet you Yeah, thanks for coming on to the show really appreciate it very excited to talk to you all the way from Denmark originally yep, and now lives in New York working on Entipology and topology. He's doing software engineering physics and optimization there Entipology is doing really cool work in engineering software for advanced manufacturing. We're gonna be breaking that down At technical University of Denmark Christian did a master of science in engineering design and applied mechanics and was also doing research assisting there So Christian, let's start things off with this with this Perspective on your journey. So how did you get interested in engineering and in design and software? How did you get interested in these fields? Yeah, so I've always you know been interested in in how things work and how you know We can describe Physical phenomena with with models and have been curious about how yet technology works And so I think it's it's natural to go into into engineering that I think that was in my night You know the most natural thing for me to do and so I have so after high school, you know, I went directly into the technical inversion of Denmark where I Decided to I was I wasn't quite sure whether or not I wanted to go with math or with Mechanical engineering, but ultimately I decided to go with mechanical because it also had some, you know, physical aspects to it So, you know, you model real life things and you know, I Saw that as being, you know, the way I could provide most value create more, you know, the biggest difference and so I Like to you know see, you know, abstraction is a good thing But I also like to you know, see that it actually, you know, the things I'm working on actually has an impact on You know the things that The work that you do has an impact on the on the industry and you know and so so after After I started at university, I still wasn't Sure, you know, whether or not I wanted to go with the traditional route or or I wanted to do more math Or I wanted to do optimization But after some time, you know, I was really inspired by you know The way you you know, how how we with compensation with we are able to set up models which Allow us to you know use computers to actually drive design. So take the manual take the manual Steps of a design process away and let the computer do the work because you know technology is evolving so fast And there's definitely an opportunity here For us to use that to our advantage and I think that's really what what got me interested in optimization in general and also, you know in order to make it it usable, you know and put it in into a real product and and have Solve real, you know solved real problems, you know, you also need to you know, it needs to be Available in you know in a cat system or in this, you know, you know In a in an engineering software way, so that I think this combination was really what sparked my you know interest for for Computer, you know software and you know simulation-driven design. Yeah, and all these things. Yeah Yeah, looking looking back at at you First picking up your interest based on your desire to understand how things work Yes This is a very frequent thing that we hear from the engineers and the designers and all and even just you know The sensational curiosity from young people is that if you have that and you can't ever quench it You you're always so excited to learn you'll go very far in life You've got to keep learning and keep understanding how things work And then if you really get to the level that you did which you also got into the mechanical engineering side of things the mathematical side of Things the computational side of things. That's when you can also Do things like you're like you're doing now with optimization exactly It's interesting to think about this and I'm interested to hear your perspective about it Which you know I'm sure we'll dig into this in a little bit later But having the computational power that's now developed over the last like really 70 years up until the peak that it's at now to to leverage that to be able to take away all of our mistakes that we make also to Be able to run all the permutations much faster than humans can to figure out which one is the most Optimal functional fit. Yeah, that's it's amazing to to see, you know, how much potential there is in Utilizing the computer to figure out, you know, how should we navigate this huge design space? Additive manufacturing is really has really, you know Advanced tremendously over the past years and now all of a sudden, you know We have this design freedom which we didn't have before and this is really rare Optimization and these computer aided technologies shine because now we have much more design freedom and puts it It gives the engineer so many more options to options to, you know More creative freedom, but it also comes with, you know a greater Responsibility in terms of knowledge and decision-making and so that is where I think, you know Using computer aided, you know Technology, you know using these advanced, you know mathematical models can really help the engineers and ultimately be able to deliver a product Take the product from initially initial phases all the way to, you know production ready In a much shorter time frame Yeah, yeah, and then now when you're when you're doing this this You're you this this optimization and you're you know, you're studying you're in your fit And you're figuring out where you want to go. How do you figure out that, you know, New York and Entomology and Optimization helping them out with that. How did you how did that transition end up happening for you? Yes, so during the the My final year as a graduate student. I was working on developing these Optimization algorithms, you know in more and a more academic sense academic sense and so but I've always, you know had this You know this desire to go out and solve real-life problems and I I think that is really what kept me, you know Interested kept my eye open and search for opportunities in the industry and so after I graduated I worked some time as a research assistant where I was writing a paper and doing some research and so that is where I You know Bradley, which is the CEO of Entomology He reached out to the research group and that's is where you know, I saw my opportunity And so I got in touch with Brad and two days after I contacted him, you know I was in New York and so it was you know an amazing experience and you know and I was you know very you know much fascinated by by by the tools that they've already developed But I could also see a potential for developing more sophisticated up to myself station algorithm using the The knowledge that I had acquired in degrees in my time as you know doing research And so I think that was really what got me, you know yet what transitioned me from from academia to to Entomology and I seriously considered doing a PhD, but it is ultimately decided to go with Entomology and it's funny because the stuff we are doing now is actually The bleeding working on breeding edge technology and actually developing new methods. So It's more, you know, almost like a PhD, but not just in the industry in industry You're actually and you're actually getting paid for it Exactly, you're all what you're working on is becoming immediately available to translating to industry This is this is what the whole we have so much conversation about what the optimal way to incentivize sciences and a lot of times It's not the PhD. Exactly. Yeah, yeah Okay Now teach us about the difference between the optimization algorithms that you were working on and then what Entomology was had already done with their optimization and then how you're you know, how that's coming together now. Yes So what they've done before my arrival at Entomology was lattice-based optimization so they could do optimization of of these, you know, lattice structure. So they have a a a I would say they have a great amount of tools Available to generate lattice structures and they also had some tools to simulate them and and also do the optimization But and lattice structures are mostly like material science Wise or whatever You can consider them on multiple levels. So what you can do is you can use various homogenization methods to actually consider them as materials And that is something I'm working on now actually But what they did beforehand was just considering it as a lattice structure So they when they model them they actually just look them as beam elements in similar way as you know conventional mechanical engineering Engineers model, let's say trust structure shall crane structures and so on. Yeah. Yeah, and so that by instead of considering, you know that you know use only considering the lattices as beam elements that you know puts a restriction on you know the the the Scale by which you you know you can let's if you if you had a structure of let's say a hundred million beam elements Then you wouldn't be able to to to model that with with those techniques in a very efficient way So we looked into okay, how do we make? How do we make a Is there a technology out there which allows you to simulate them much more efficiently and also allow you to simulate more general Material structure, so let's say gyro structures honeycomb structures these autotropic or these foam foams and and so in order to Simulate those you know you have to look to other methods and that's actually some of the stuff. I'm working on now But so that was one thing and another thing was we've also been doing Again, we try to develop a whole range of different generative Design process or Generative tools to allow the engineer to you know to to design parts and so all the optimization is another thing that we've been looking into so you know many companies do it out there and We recognize that you know We are not the company which goes out there and blows all composition away because yeah, that is not our purpose We more see to ball the optimization, you know as a part of the In part of it's a tool in the toolbox. So to say and it's not an entrant solution Yes, so we have that functionality, but it hasn't been you know our main focus and I've also been working on developing that You're really conquering a niche as the best aiming to be the best tool on the tool belt in that niche And you're not the whole tool belt as some people are trying to be just a little bit So so so what we are trying to do essentially so we are right now what we are working on at entomology is an engineering platform where we enable engineering workflows and engineering process try to capture engineering processes in Just in one framework all the way from Early design phases using generative design tools through simulation So we can also you know perform actually perform simulation and validate the performance of the structures To manufacturing so actually exporting them. So we are trying to develop develop right now We're trying to develop this this framework where we can capture the entire engineering process in one Unified system and ultimately, you know the the whole you know idea behind this is if you look at you know conventional ways The way people do design nowadays, you know, we have the workflow is so disjointed You know you do cat in maybe solid works and then when you want to do analysis You have to bring into answers or abacus and when you've done the simulation there Maybe you figure out I need to do some changes in in back in my cat model And so all these times you have to navigate or transmit information from one system to another And it's really a tedious process where manual labor where the engineer He uses a lot of you know valuable time He wasted waste a lot of valuable time Which he instead could have been used on actually real engineering work correct And so that is really what we are trying to do try to you know get rid of all this, you know Not this yeah This you know try to to to get rid of all this unnecessary work and try to automate You know these tedious tasks as much as possible By combining all the steps, you know into one unified system and then the One the main one that you just talked about a moment ago was the The cat with the simulation software Yes, so that when you want to make a tweak in the simulation software You don't have to go and make the tweak back in the cat and then re-imported the simulation How how are you then merging those two? So the way we do that is in terms of What we use is we use fields. So you can consider Anything in terms of fields. So What we have in entomology we have a unit or one of the special things about entomology is that we have an Implicit geometry kernel to represent geometry. So it's not the conventional way of represented geometry where you have boundary representations, but instead we what we do is we use math to represent geometry and what that really allows us to do is to Very you know to so that is essentially you represent geometry in terms of a field And also if you look at simulation data, you can also consider that as fields And so by having Almost this unified language as you know just one field we can very easily cover these different, you know steps of the engineering design process into Into an automated workflow and get rid of all the the Irrelate all the the very tedious tasks correct enable maximal creative engineering time. Exactly. So so what how do you? Label a you said it's a it's a geometry Kernel. Yes, that is it's so you're so like a geometry You would give numbers to and that would be and then when I would tweak that geometry the math the numbers would change Yes, so actually one way you could think of it is let's say you have a sphere So if you want to represent that in our geometry kernel The way you would do it is you would represent it in terms of you know the equation for a sphere So we did x squared plus y squared plus z squared Minus the radius and that essentially defines the the sphere and what this really also allows us to is store objects in a very lightweight format and And that is also you know brings us to maybe one of the other very interesting things about you know what we can do and that is generate Huge, you know huge lattice structures or these periodic structures in a very lightweight form lightweight formats So for example, you know in our system, we'd be able to generate lattice structures of Hundreds of hundreds of millions of beams and be able to load them almost instantaneously. That's really interesting. Yes And so I think that's really, you know one of the strong points of Our system and this implicit geometry kernel and it's orders of magnitude faster than you know what you can do elsewhere So it's yeah, yeah interesting and it's so it's So due due to the light weight Of your tech but the light weight of your tech is due to the geometry kernel Which then just is assigning the the x squared plus y squared plus z squared minus the radius Yes for us here for a sphere you're giving a value to the x the y and the z and the r Yes, and the final and the final answer. Yes, and you're giving and that that those five Numbers and the final this this final this this way of storing it. Yeah. Yeah, it gives you a whole sphere Yes, and that way you can load that And it'll give you the whole sphere in the and that's why when you make the change to it It'll change it mathematically. Exactly. Yeah, okay, okay, and it also has a number of other benefits. So for example, you know Traditional, you know in traditional caps, you know some typical operations are you know if you wanted to do Booleans or you wanted to let's say you have a box here and you have a lattice sitting combined with with the with the Your structure it could be let's say you have a sandwich structure So you have a solid plate and you have a lattice and you have a solid plate You could use that to maybe to make a very lightweight part. Okay And so for those cases, you know if you if you used conventional cat systems And you wanted to have a continuous blend between the lattice and the solid Then you would manually have to go in and click on each of the Each of the different beams the connection points between them and you can imagine if you have a solid Or if you have a lattice comprised of 100 000 beams that quickly becomes an impossible task You wouldn't be able to do but you can make a script right for that process. Yeah, but the thing is that with implicit With this implicit geometry kernel what we can actually do is we can do that It's it's it's natural to the system. So we can do by we can do the blending with on the underlying math So what what this means is that we can blend In a matter you don't have to do anything you just have to you know, the blending is defined in It's done on the math equations And so what this means is that you can do a blend which never fails and it's done fully automatically And so this really gives you again, you know, it removes the tedious, you know parts of the design process You don't have to write a script or you don't have to manually click on all the buttons And this really also means that you can you know in your design process you can quickly iterate through So let's say you figure out. I need to make it that is just a little bit bigger or a little bit wider Everything is done in an automated fashion. And so you don't have to go through that, you know, it just works And I think that is you know, one of the you know the powerful aspects of of the geometry panel And then your optimization that you were doing and then are bringing to anthropology is optimization within this geometry kernel Yes, that's correct. And so it is part so partly it is just conventional topology optimization as you see out there but we are also looking into Optimization using these These cellular materials. So for example, you we can talk about gyroid structures or or we can talk about lattice structures Or honeycombs of these various foams and such materials I've actually shown to have excellent performance Um They can provide material You know, if you zoom out sufficiently you can actually consider them as materials. So if you have enough of these Period this periodic assembly of gyroge, you can actually consider them as materials And what this means is that by going in and locally changing how the microstructure of these these cellular materials You are changing the the microstructure of these cellular materials You can actually locally change how they respond their material macroscopic material properties And what this allows us to do is actually tab into material regions which we Haven't been able to before so for example, we can make materials which are extremely lightweight with very high strength to rate ratio materials which isolate Have a very good isolation properties materials which Dampen the energy in a very efficient can absorb energy in a very efficient way. And so Obviously, you know these using these materials in mechanical engineering design is obviously, you know very It has a huge potential. And so what we're working on now is figuring out how can we Do optimization using these material structures? and so yeah, that's some of the stuff that we are working on now And you know the thing is with these materials that is that Compared to conventional materials, you just have okay, you have a finite number of materials that you can pick from But the thing is with these materials is that you have so many more so much more option Because now you can also go in and locally change maybe how thick should the beams be? What orientation should they be? How big should the cell size be? And so figuring that out is not trivial And so what we are working on now is trying to figure out ways in which we can guide the engineer through the The process of of designing with these and building out workflows which enable an engineer to do so. This is so cool Okay, all right. So how do you um Guide the engineer through The through the the process you gave the examples of okay used to be just different materials Then it became well now you can go You can kind of double click and go in and see all of the connection points between those two plates and the lattice in between So so how do you guide the engineer to then to see for themselves that? Oh, wow That's an interesting way that anthropology is guiding me to learn that I can make this more Functionable in the desire that I wanted to be so there are many different ways in which you can do this and one way would be to actually use The same methods as as we use for travalry optimization So automated optimization procedures to figure out how the distribution of of these Micro or the cellular material should be and so that would probably be the you know the most you know General the fastest way of doing are you constantly running the optimization of the cellular materials? In the in the you're constantly running the simulation in the background and then saying this would be more optimal Yes, so it is run in the exact same way as you you know run conventional optimization processes And so the way this works is that you set up your design space you say, okay This is my design region here. I have some here. I fix it here I fix it then I apply some external loads and then I say it must meet certain requirements So for example The stress must not exceed a certain value in this region or the displacement might not be Greater than some value in this region and you say it must also be lightweight and Yeah, you set up these various constraints to the optimization problem and essentially The once you have set up the optimization problem And defined, you know the overall constraints and objective. Yes, then the underlying Are all inputs into your exactly So this would be the user which sets up and defines the problem and the optimization algorithm then Figures out how the cellular Material structure should be interesting should be interesting. Yeah. Yeah. Okay. So there's not There's there's not even a chance for then An error really you're you're rolling with you're simulating out all of the possibilities and giving the best The thing is the thing is that if an engineer were to you know He has the engineer has so many different parameters that he can you know He can play with and it can really you know, it's really an overwhelming You know, it can really be an overwhelming, you know Experience because yeah, there are so many different parameters that you can you you can Control and so what we are trying to do is set up Workflows which enable engineers to to to actually Guide the engineer through the you know, this vast design space And really allow the engineers to fully utilize the potential that these materials have Because they can also be used in a you know, if you if you don't select your Your you know design the cellular material or the cellular structure probably Then It's not going to be bad. It's absolutely not going to be better than what you can do with conventional design methods So you really you know, there is a higher higher knowledge requirement using by designing with these cellular materials, but There's also a greater potential and what we are trying to do is just enabling engineers to actually tab into this this Yeah, yeah, yeah Increased potential. That's right. Yeah, and this is so cool. It it There's both the the the the lattice this this the cellular Distribution Matrix, yeah, it's a cellular matrix That can look like a lattice or a honeycomb. Exactly, okay Or gyro structures gyro structures So cellular matrix so which is that's where you see yeah the angles of the different beams that you're talking about and then that Being able to help with that process and have the engineers say, okay, I see why that's optimal It's going to have like we asked for all these variables. That's it's it's great. And then can you give us the Examples that you find most Relatable for that you gave a crane as as an example of where there's all those beams Yes, right. So cranes one of them then what are other examples where this can be like applicably understood so examples where you know I would say Any, you know, if you want to make something, you know, very lightweight, but also very, you know highly performative It makes sense to do this so it could be, you know structures Which have, you know Very high requirements in terms of weight saving. So for example in the automotive industry or in in the aerospace industry Yeah, and so I think those and those two industries would definitely be, you know solid Okay, let me see if I can get this right then so then the the cellular matrix in the automotive and the aerospace industries You don't want extra beams that are unnecessary. Yes You you can remove if you can remove beams In position them in in more solid ways to be able to hold Um to hold it together and have equal performance. Yes, you want that because it weighs less exactly So we want to optimally use the amount of it, you know We want to only make use of the material that we actually need and so we can use these optimization algorithms to figure out how How that is, you know, it should be done fascinating. Yeah, and so they also actually also use it quite a lot in the medical industry So, you know, these cellular materials, it's partly to do to you know So they use it in the medical industry for implants. So They can actually let's if you want you need your hip replaced or something like that What they do is you can have a customly made 3d printed, you know Hibbon plant inserted into your body and what you want is of course it has to be lightweight and that is really where the cellular materials You know make sense And and another thing is, you know, it also has to you know fit into your body and And and with these, you know cellular or these cellular structures, the bone can actually grow into the Yeah, yeah grow into the Into into the 3d printed cellular structure rather than read body reject it. Exactly Exactly you want waste for the body to actually incorporate it into exactly. Yep. Yeah. And so I think, you know Really, there is, you know, so many great Opportunities to to adopt this technology and I think, you know, the reason why we haven't seen it before now is because because Additive manufacturing has really evolved a lot over the past years. And so this Wouldn't really have been possible if you just go back 10 or 15 years Yeah, because you would need to Take and the the the piece of of raw material you have and try and actually carve out what And you can imagine how that would have been if it was just if you had an simply of Or you had a structure comprised of Just even 1,000 beams wouldn't even be possible. Yeah. Yeah. And so I think now with additive It's it's you can you can you can make the tweak. Yes. Yeah. Yeah. Yeah Yeah, additive manufacturing has really, you know, it's the driver for, you know, what we are doing now. Yeah. Yeah It's and it's also, you know, a great motivation because you know, 3d printing is one Form of additive. Yes, that's correct. That's correct. And the other popular forms Um additive. Yeah um So I'm not so much into I wonder what I wonder what else are like the other popular Additive manufacturing forms Yes, I mean 3d printing And then I don't know maybe something related to biology. It could potentially be like an Form of like an additive. Yep. Yep. Exactly. Potentially like cell division type stuff. Yeah, exactly. Yeah And so I think um You know We are in an interesting time right now because you know If you go back 20 years, it was manufacturing. That was the, you know, the limiting factor of The, you know, the evolution of technology, you know, cat software, you know, was on the forefront Whereas now over the past 10 years, we have seen that that the That the You know, the manufacturing scene has evolved much faster than the cat software. Interesting. And so now we are in a unique position where, you know We are actually lacking behind and that is also why I think it's so interesting to work at entomology because now we have a unique opportunity to actually, you know Enable design, you know, we have such great, you know opportunity We the design space is almost unlimited now. And you know, now we have the opportunity to develop tools Which actually enable engineer to to to have into this. Yes. Yes Christian a couple questions On the way out. One of the questions is, you know, we're here at cofaz and there's so much Cutting edge engineering software, including yours. Where do you see this all converging and going? Oh, that is a great question. Um, I think over the next Couple of years, you know, we'll definitely see improvement in, you know, the the technology So top all the optimization and, you know, just generative design tools, you know, those will definitely evolve because those are still, you know They they are not into insolutions now and they still I would say they're still in their infancy And so I think over the next years, we'll see, you know, we You'll see, you know, with, you know, the increasing, you know Yeah, the compute power is just Growing exponentially right now and obviously that helps a lot But I also think, you know, we'll see more methods development in these generative design approaches or strategies and I think that will, you know Move the technology to a stage where it's also easier for for Less experienced people to, you know, to get into the the industry and I think, you know That is probably also the biggest barrier of entry right now, you know, that you still need, you know Some level of, you know, technical understanding to actually be able to operate and make sense of these tools So I think, you know, over the next maybe 10 years, you know, we'll see a larger adoption due to Advancement in the underlying technology and the compute power Yeah, yeah, the democratization is fascinating getting it into the hands of younger and younger people Exactly now Last question is What would you say is an important or maybe even the most important skill For kids and for even adults to learn going into this exponential technology age So I think obviously have being curious is it's very important, I think Um, but even more so, I would say now in terms, you know, if it's just a technical skill I would say programming it it's really amazing to see, you know, how much Being learning this skill from a young age, how much it can do for you because, you know, you can Essentially solve all problems with with programming and then you can later on in in your, you know In your life, you know pick up on, you know, the more advanced skills such as, you know, these Yeah, mathematical skills and understanding the physics, but I would say even from a young age Just getting started on programming is probably You know, the best advice I can give, you know It's it's it's such a powerful tool and you can really solve so many problems with it and Pretty much all the systems that we see today, you know Have some form of programming, you know involved So I think having that skill or learning that skill from a young age would You know it Yeah, that's probably the best advice I can give. Yeah, that's great curiosity programming powerful too. Yeah, definitely Christian, thank you so much for coming out to the show. This has been a pleasure. Yeah, it's been a huge pleasure Yeah, and this is like, you know, the way you've taught us about about what you're working on with optimization and What what anthropology is doing is is is really beautiful. It's it's really well You did a great job explaining it and we greatly appreciate everyone for tuning in huge. Thank you We'd love to hear your thoughts in the comments below. Check out christians links and topologies links below Also, check out co-fes is links below as well Support the artists and entrepreneurs that you believe in our links are below simulations links are below get more conversations Thought-provoking conversations about things like what we talked about into our communities Inspire the next generation of people to build the future manifest your dreams into the world We love you so much. Thank you for tuning in and we will see you soon