0:00 Good morning everybody, thank you for being here. My presentation will be a bit different from you have been seeing so far, for two reasons. First of all, I, we, we apply computational design, so we are sort of clients of people like spine and and top, but the other reason, probably you never heard about what I’ll tell you, is that we were, we work on ceramics, and we don’t work on ceramics like toilets or plates, we work with Advanced Ceramics, and by Advanced Ceramics I mean something like this. This is a porous burner called, so it’s a it’s a burner in which the flame is constrained within its porosity, and of course to have this component running you need a material that withstands very very harsh environments, and by that I mean temperatures around 1300 degrees C, and Ox, oxidative environment. So these are highly engineered materials, both on materials themselves and on their topology.

Before going, I, I will, I line been my talk, my, I, I will present a little bit my University, I’m coming from Southern Switzerland, so I am in the Italian speaking part of Switzerland, and we we are within a University of appli Science, and I will tell you a little a little bit about that. Then I will speak about how I, from where I started, because it’s, this is very interesting also, and then why we choose computational design to go ahead, and then finally I show you some applications.

2:05 First of all we are in the in the very south of Switzerland, close to the Italian border, few kilometers, and we are in Lugano, it’s a town in in the south of Switzerland, and we have a new campus right now, just in in downtown Lugano. Within subsid there is the m Institute, The Institute of mechanics and materials, and I am heading the hybrid materials laboratory. So before saying what what where you can go with the topology optimization in Ceramics, I would like to tell you a little bit from where you, we started.

In talking about topology optimization in Ceramics, for many many years, I would say 20, 25 years, I was dealing with porous Ceramics, and normally before additive manufacturing you were using poral Ceramics by templating poly rtin Foams, that’s why we call them still Foams, not yet, not lazes or whatever, and we were using a very very old process which is called the Schwarz balder method, which is back to 1967, and it’s a very simple method. It’s, you take a a a poane foam, you cover it with with a slur, which is a a mixture of powders and water or other solvents and binders, then you heat treat this body, so you burn out all the organic part, so the poly retain foam inside and all the organic binders, and in the case of most of the Ceramics I’m working with, finally we do what is called silicon infiltration. So we, this body, which is comprised pretty much of carbon, it’s a, we put this into a a vacuum furnace at very very high temperatures, we melt silicon, and silicon by capillarity is sucked in within the micr structure of this, of this porous body.

4:20 So it’s very interesting to see from where we started, I mean we were trying to understand how these really random structure, by random I mean really random, because it’s a foaming process, you cannot control, at the end we were trying to understand which were the weakest point of these structures and so forth. And so we started doing computer tomography of the forms, we were making, making some, some let’s say loading, could be thermal, could be mechanical, could be whatever, and we were watch, watching where is this material able to fail first. And talking about meshing, this was a real mess for us, I’m talking about something, this is more than 15 years ago, so making the mesh, you see here on on the left, it was it was a completely a nightmare, because we were starting from a computer tomography, meshing it, fitting all the these hell of a triangles to to to to together, and trying to make something that was working for a simulation, so it was a real mess.

5:33 And the best way to work with bigger structure was this way, was producing cells easy to model with a cad model, like enx, replicate those cells which were pretty much working as a foam, as a stochastic foam, and then saying that that pretty much we were doing what a fo, we were, let’s say, simulating what a foam should do. In this case this is a European project for a protect of a sandwi structure for a re-entry vehicles, and we were simulating how how this structure bumps up, when when it’s heated up on on one side and it’s cold on the other.

6:23 And then, and then it was 2005, 20 four, I saw for the first time a t fell printed by Fus filament deposition, and for me this was mindblowing. I mean because I immediately saw a a foam, for me this was a foam, so I started having a a cartoon in my my head taking this foam and process it as I was usually procing it. So all of a sudden, I was, what I, what I was develop Ving to understand what was going on became the starting point for making what I want. So I mean it’s it’s a completely reverse way of engineering materials.

And besides all the cells, all the wonderful topologies that you saw so far, I mean it’s a, you cannot believe, I just doing a simple thing, like having a bunch of Cubes, CU cubic cells, into the space, okay, and flowing a gas, liquid or whatever into this Cube. If the pl, the face of the of this Cube are perpendicular to the flow goes, you you you will get a flow like this, okay, so it’s a straight flow. Now the same cubic cell, I just tilt this cell of 45° in the three Direction, and just the same cell, will the flow will behave like this, and this is amazing. I mean you just change everything, you, this is this is a mixer on the bottom, and in the other case it’s a completely different flow. So I mean this is really changing the game for for us, especially for for the burners.

8:29 So we, for for for many years we started, u, playing with these latices, at the time we didn’t even know that it was called computational design, we were just CAD drawing things and testing them, and of course we started with the with Matlab, we we were assigning different thickness to whatever, then grasshopper, rhinos, seros came out, so we were able to to grade the pore size in each Direction, and blending cellular structures to to have a different flows within just one pipe, for instance, which is this case. And together with my colleagues we were making a lot of simulation, therof fluido dnamic, thermal fluid, Thermo mechanical, whatever, and so this is really from let’s say 2010, this really flipped over the way we were doing these things.

9:31 Now few words on on what actually we do with those numerical models, we have, we have different ways to print these numerical models, on top left you see a so-called binder jetting machine, it’s a it’s a it’s a machine that deposit a a bad of powders with with a given thickness, and just printing an ink, you will create layer by layer a 3D dist struction. On the right, top right, you see a stereographic machine, very precise but with the optical limitation, and also on the bottom left you see a powder bed Fusion machine for a process in which we were preparing porous body for a further infiltration, okay. And finally, and this is the part more of my experience, we have a bunch of ways to thermally treat these materials and convert them from a say ceramic polymer ceramic composite to a fully ceramic material, okay.

10:37 So finally, let me give you some examples of many many years of experience, we we work, we have been working a lot in renewable energy, in Aerospace, in industrial heat management, and also in filtration. Let’s start with this, imagine that you have a tube which is heated from outside, and you want to heat the fluid which is flowing through. If you don’t put anything, it’s it’s it’s very uneffective, you just hit the the the the the first layer of of the fluid touching the walls, the inner walls of your tube. As soon as you put a a a porse body, you start increasing the the the the mixing of the fluid, on one side you more surface to have a heat exchange, and you have also guides for the heat to be transferred from the outside to the inside by conduction, but that’s very limiting, because if you make, if you put a regular structure like the one on the left here, you don’t fully exploit all them, all the ways that he is transferred, because heat is transferred by conduction, convection and radiation, in in in in a in a cell like this radiation, it’s somehow shaded, okay.

12:04 And then if you think of a tree, being inspired by Nature all the time, you you see that the tree has most of the leaves outside, and the lower you go the the less are frequent the leaves, okay, so there is a kind of graded, let’s call it cell size, of this the porous body which is the tree. And so if you reverse this idea, so, and this is let’s say the idea of this cell in which the we were grading by with Matlab the cell size. So here there were, there were, there was a big cell, and radially going outside the cells were getting smaller and smaller. In this other case, the smaller cell was inside here, and the bigger ones were here, and so this proved to be, I’m sorry, I’m using, yeah, in this one the smaller, and this one is the bigger, and so if you want to let the the radiation penetrate from outside to inside, of course this solution is the better, because the the outer cells are not sha, shadowing the the the inner cells, so you get a much more efficient heating of your tube, you see here from the temperature distribution. So that’s one case.

13:29 Another case was M, putting together mechanical properties and termal properties, it was a, this was a process with the DLR, DLR, it’s the German NASA, I say, they, we were developing a re-entry wing for for a re-entry vehicle, and this is the this Wing tested into a PL, my wind tunnel, so very high flows, very a lot of heating by friction, and we were making a sandwich structure like you see here in which the inside was filled with the ceramic material, and this the goal of the using this was twofold, on one side you have a sandwich structure which keeps Integrity, on the other side we were flowing a gas inside here, and we were removing heat from from the Leading Edge, which is, you see here, is is the most heated part of it. You see here the setup before testing, you see here the sandwich structure here, this is a CMC, it’s a ceramic Matrix composite, able to withstand high temperatures, you see this the our structure inside here, and you see the sample after the test, it survived something like 20 test, 20 reentry tests, and you see some oxidation here.

14:55 Of course another, another application is is a a heat exchanger for using a temperature to recover heat from a solar concentrated solar panel. In this case we were using a porous body inside here, this is a plate, and plate and something like a plate and fin heat exchanger, very Compact, and it was used in the in in a project for, it’s called hydrol, for to produce solar fuels. This is another application in which we were developing a porous body here to enhance heat transfer from let’s say a gas to a a phase change material for each storage, this was conducted with FFL in in Switzerland, and was part of a huge project, huge Swiss project, to recover let’s say industrial Heat.

15:59 Another very interesting process, project, was developed with, again, in Switzerland, and within a European project, it’s called The thermochemical Heat storage. It’s Kind kind of complicated, by sorbing and heat in the winter by adding more water again, and producing heat by this way. To do this you need a lot of surface to to to have this exchange, and so we were developing these weird structures, periodic structures, in which the goal was, to during the dripping of this solution, allowed more and more wetting of of of a surface, so you see, for instance, you start, we start here with one drop, and we come out on the other side with three drops, meaning that the flow was was spreading.

17:05 Finally, an application on this, is the the most let’s say Advanced application we are doing so far, is called in catalysis. You normally hit a catalyzer by hitting it from the outside, but if, in this case, if you put a porse body within your reactor and you heat it up by Jal effect, you can, by depositing the catalyzer on on this porous body, you can say have a a very efficient way to heat up your catalyzer. And so with the topology optimization we designed these structures which are let’s say heaters made of silicon carbide, with topology optimized both for flowing and for heating up by jeel effect.

17:51 This this thing was was mounted within a reactor, this is a mockup of the reactor, you see it inside, you see the current is Flowing from here to here, through here, of course, and in this case the reactor was a reactor in which we were cracking methane, so we were feding this Reactor with methane, and we are getting as two products carbon, which is very much used for batteries, and hydrogen, which in our case was a a byproduct, very interesting byproduct by the way. So you see here that we are producing carbon in this way, and this is a very interesting way to produce carbon, because normally carbon is used to to be produced with the methods which involves a lot of production, production of a lot of CO2, so this is very very interesting.

Finally, where, where to go. Well, yesterday Bradley was, I totally agree with him, the next step for for me is to go backwards from what we, where we are heading now, I mean we we start from a field and we get the structure which better fits with the with the with the with the fielded, whatever it is, termal, temperature distribution, loads, whatever, stresses, and so this is definitely the way to Pur, to to further increase the efficiency of these kind of structures.

19:30 Yeah, this is my team, which I would like to thank very much, and finally, if you are interested in ceramic structures, and ceramic structures produced by additive manufacturing, we just finished to write a a chapter of a book, of this a book of hand, Handbook of addtive manufacturing, so you can find more details of what I was saying to you so far. Thank you very much.