Fluid path optimization - Generative Fluids in Fusion 360

GILBERTO FERNANDEZ: Hello, guys, and welcome to the technical instruction class on "Fluid Path Optimization on Generative Fluids in Fusion 360." So first off, to start with, there's this safe harbor statement that we have in this presentation, so for forward-looking statements, et cetera, so please be aware of this.

So summary of the class. This class will show how to use generative design technology to optimize the fluid paths in Fusion 360. So I'm going to take examples of real class models. We will show the workflow to get a model ready, explore that, generate multiple CAD-ready solutions based on constraints, product performance, and finalize on a solution that is optimized.

So the objectives of the class will be getting introduced to the concepts of generative design technology in general, understand how a real model can be set up in Fusion 360 to run generative design, knowing the new technology of applying generative design to fluid mechanics, and finally, seeing how efficient 360 offer us multiple solutions for fluid paths, pointing the user to high-performing alternatives.

So in the agenda, we'll have summary and introduction to generative design, talking a little bit about, what is generative design, general, what makes it different? Then I'll talk a little bit about the workflow for the normal generative design, general workflow, a little demo video of that to show the technology, and places to go for further information.

Then we're going to focus more on the fluid path studies on generative design. There's going to be description, then set of commands describing how the user interface looks like before going into a fluid path study example. So this contains setting up of a model, of a flow control valve. Then we're going to be exploring the outcomes from that study. Then we'll see how this integrates and runs using Autodesk CFD as well.

After we-- we finish all of this, we're going to look a little bit into the future, into a wish list couple of slides to show what would be nice to develop this fluid-- fluid path within generative design, and that's going to be a set of additional resources. Finally, there's going to be-- in the live session, there's going to be a bit of a Q&A.

So without further things, let's start first a few words about who I am. My name is Gilberto Fernandez. I'm a technical account specialist in Autodesk in the customer success organization. I have an engineering background, and I have more than 15 years of experience in the field of simulation, computational fluid dynamics.

And my role within Autodesk is to be responsible for the technical support for Autodesk premium customers and, in particular, putting the focus on simulation solutions. I am based in Barcelona. Then, below, you can see the two main softwares that I'm an expert on, and those are Fusion 360 and the Autodesk CFD, so it's all about-- it's all about fluids.

I normally start my presentations with a quote on the subject. Even though the subject is optimization, this quote here says, "True optimization is the revolutionary contribution of modern research to decision processes." This links really well with our intention in generative design where what we want is to contribute to making the decision process better.

This, by the way, is a quote by Mr. George Dantzig, who was a mathematician-- he's really famous because he was the inspiration of the character of the movie Good Will Hunting, so good quote that reflects that the optimization is that the big revolution when it comes to taking decisions in there. This is exactly what we're going to be trying to apply.

So first, let me give you a little bit of an introduction to generative design. This may sound a bit familiar to some of you guys that are already into generative design for manufacturing. But anyhow, it's good to set the environment for them talking about the fluid path statics.

So what is Autodesk's generative design? It's basically a design exploration technology. So it simultaneously generates different options and solutions based on real-world manufacturing constraints and the product performance requirements.

As you can see on the right-hand side, this is a very simple example on structural simulations. This is a design of a chair with different constraints. So generative design explores-- iterates on different shapes running a simulation and achieving-- or trying to achieve an optimal solution for particular requirements. So it's a little bit of a-- for some of you guys that are more into simulation and stuff, it's more of a big parametric study based on constraints and based on optimizing performance. So it goes way beyond, sometimes, the idea-- the traditional ideas that we have in there for a design, as you can see on the right-hand side.

So what do we care about it-- about this? So let's talk about why generative design is talked about that much. So due to the fact that there's-- this notable impact in there-- not just the products but the people that surround them.

So, yeah, first, on the left, many of you may have heard about the work that was done by GM around this seat belt bracket. This highlights one of the aspects of generative design, where we're taking multiple paths and we were able to have the software help produce a single equivalent in there, that can optimize the mass, improve the efficiency, reducing supply chain costs associated with each of those unique parts. So it's really good optimization that was achieved in that.

And secondly, on the right that's the impact on people's lives as well. So this company DisruptDisability uses generative design to make the wheelchairs and optimized in terms of the mass, the versatility, making these wheelchairs customizable, tailored to people's measurements. It has already had a big impact on people's lives through this company.

So today, most paths to a final product may look something like this. So these few concepts are generated. Even fewer are evaluated or assessed for manufacturability. There may be some iterating back and forth, in there, between some of these steps for those selected concepts. Then there's a final validation for production.

So this time to market includes this iteration phase, this design to production. So how does generative design help this product development process? So let's overlay what generative design can do to help in this process. We get considerably more concepts that can be explored. Then, additionally, many of these are already validated against the manufacturing methods, allowing for the productivity increase. So we shorten the time to-- shorten the time to market.

So this is the overlay from the comparison before. So you can see that in terms of time to market, there's a productivity increase around the process. So how this is different than optimization? What makes the design exploration unique is its ability to analyze all possible variations to a solution, present the list of choices to the customer to do the trade-offs.

So using transporting analogy here, we can outline the difference between this optimisation and exploration. So we have a route here, from going to from A to B. We can have optimization there. So we use the shipping, and the optimization would find the shipping route that is faster or is most fuel efficient. So we have a faster route in there.

But then we have this optimization. Then we have the exploration, which contains these two, but then also contains and contemplates different possibilities-- different transit possibilities that go beyond what we were talking about-- the optimization of the shipping or going different ways and explore all the transit possibilities there to evaluate the potential options, the solutions, and pick the right one that meets the requirements.

This can be plotted according to different criteria and performance indicators in there like the speed, the security, the cost, et cetera. So we can find the best trade-off in there for the costs, speed, et cetera. So this is the-- a difference between the optimization and this design exploration, exploring different routes.

So where does generative design apply and where would it fit into our process? So first, it could fit in our new product design. We can use the results as a design guide to understand where material may be required and then manipulate the design-- manipulate the design from there.

Second, it would be-- as a form of part consolidation, taking subassemblies with many parts and combining them into one part that may reduce the assembly time or improve the performance. And then, lastly, it can be applied as a part enhancement, so basically taking an existing product and improving its strength-to-weight ratio, improving the manufacturability. So we could use, maybe, cheaper material, keep similar performance. There's a lot of areas where generative design does apply.

So basically, the goal for the activity is to have the right balance between the performance and the cost to produce for a given design. So sometimes we are limited, sometimes, in time. So we cannot be thinking in all the infinite possibilities there. But with this technology, we try to reach out a little bit more.

So you can have, also, examples in there, because there's this price on the-- price and performance curve there. Each one of the designs basically fits a different purpose. So there's the cost to produce. There's the four months there.

So it's not a matter of us saying, this one is exactly the best, because it depends on if we prioritize one thing or the other. So there's different criteria. So there's a big curve. There's a kind of cloud of options there. So we need to decide which one delivers the value or the balance that we basically need. And this goes for car, but this can go, also, for any sort of part in there.

This was a bracket challenge. There interesting thing in there is that many hundreds of designs were created, and all were satisfying their requirements, but all within different costs to produce based on the materials, the processes, and the other design parameters.

So the challenge for any design team is, how can you produce all of these options, which-- it's a big effort. It's a big engineering effort to produce all of these options in there. So this is why we need some sort of rapid decision-making environment. So the answer is to turn to generative design, where you can basically assess lots and lots of possibilities on the same thing, with knowing exactly what's going to be the performance, knowing what's going to be the cost, knowing what's going to be the resistance-- in this case, the stresses, et cetera. So this highlights, really, the importance of this generative design technology.

So there's a few examples here as well. So basically, we know we can target these things. With the help of cloud computing, we can now make more-- or increment the ability of the engineering teams to develop and to explore the full design space, the full range of possibilities for any of these problems that they may want to explore.

So to talk a little bit about the standard generative design in Fusion 360, there's-- let's talk a little bit about the workflow. And basically, what we do is we basically start from a design. We may already have an existing design, normally, that we're looking to improve or for new concepts.

From this known geometry, we'll need to project specific geometries that will control the generative-- generative space. So at the very beginning, in the car, we're going to model the appropriate Preserve and Avoidance regions or obstacle regions. So those are basically the parts that need to stay because, normally, we're going to be reducing a part that's going to be-- itself-- that's going to be part of an assembly.

So there's parts that are fixed in there, that are required, and then, again, there's some other areas where I don't want material in them. So those are going to be the Avoidance or the Obstacle-- or the Obstacle regions.

Then we're going to be generating-- so setting up the design options, so all the restrictions and the criteria and the multiple materials. We're going to be setting up the study with the geometry, with the constraints, and the loads, et cetera.

Then the workflow-- we're going to be hitting Generate. So we're going to be solving, and we're going to be generating different outcomes. Once we have the outcomes, we start the exploration through all of this design options to see which one best fits our project's price, performance curve, et cetera, before we move forward.

Once we finalize an outcome, we can export that-- the desired outcomes for the actual use. So all of these can be seen in this-- this demo. So this is going to be a part of an assembly, as mentioned there. So we start with the model. We go to Generative Design or Generative Design Workspaces.

So there is an area where I basically edit the model a bit to generate these areas that are going to be preserved, and also the Obstacle ones-- so you can see, in here, we basically define a few areas like this-- the whole of the part that's going to be left to software to generate.

But we need to define what parts we want to be present. So those are going to be Preserved parts. Then we're going to select Obstacle Geometry. So we don't want material to go there. Then, in terms of the design conditions, we put constraints in there. We're going to put whether the part is attached or not. What are the loads in there? What sort of resistance these part needs to cope with?

Then we set up criteria of weight, limits on weight, et cetera. Then we set up manufacturing criteria, as well, for production. So if we want to include milling, casting, et cetera, we set up also different materials to give the software options for building up this big range of options to compare. So we select multiple materials.

And then, finally, we check and then we generate the outcomes. So there's a lot of outcomes in there. Then we go to explore and to see which ones are performing the way we want. We can filter through the different parameters. We can see, as well, a list, and we can sort out the list in terms of the mass of the materials, the different performance.

You can see in here, for instance, the mass versus the displacement there-- maximum displacement for a particular load. So we measured the resistance, stiffness of our part. On the exploration, we can compare, different parts, different shapes, different materials, et cetera. So we'll end up deciding on an optimized one.

Then, when we have the optimized one, we can see all the parameters-- whether the manufacturing method is the right one, what's the mass, what's the volume, et cetera. So we end up with one, and we can basically create a design from that outcome so we can use it so we have it in there and we can use it, and we can basically put it back. So that's the way we do the generative design with the final product, Joining the-- the big assembly.

So after this demo, there's further information on Autodesk generative design. If you want further information, there's plenty. There was actually a class in Autodesk University, 2019, that-- I gave that class, and it goes through things more step by step. So that demo-- by explaining a little bit how to go around these things, step by step in the standard structure of generative designs.

There's also documentation on the Fusion 360 help. And that's also courses, certification in our Autodesk portal, so you have plenty of information in there. Then, needless to mention, there's also a lot of information in our channels in YouTube with-- where there are generative design meetups and things like that.

So once we have explained a little bit what generative design is, let's go, a little bit further, into the main subject for this class, which is the fluid path studies. So within Fusion 360 Generative Design, we are able to do fluid studies, so also known as generative fluids, meaning it is similar to the normal generative design. But instead of being a structural simulation, this is based on fluid simulation.

And the difference is, instead of taking care of the solids-- their round path-- we are now setting up a flow path. So we're going to be caring about the Fluid Path that's going to be optimized. So when we select this type of study, the outcome is going to be the actual fluid, not the solid. So it's going to be the fluid, and the path that this flows to for an optimum pressure drop, and that's going to be based on the geometry, on the performance requirements.

So this Fluid Path is going to be the shape of fluid as it flows through something, through a device. So it will help us reduce in time when designing fluid paths so we get an optimization. And it will help us understand the-- between-- the comparison between different designs, understanding trade-offs, and prioritizing our requirements. So this will really-- as it is optimized for pressure drop-- the place where it really comes to life will be for flow control, for pressure drop optimization. Will be applied for things like bounce.

So first, I need to mention that these Fluid Path studies are still a featured as in Preview-- Preview Features. For those of you that are familiar with Fusion 360, you may be familiar with this thing. But the Preview feature is a new functionality that's still in the conceptual stage. So it will still be susceptible to changes before it goes-- it goes fully public and belongs, fully, to the software.

So there are different types of Preview Features. There's public ones, which are close to being fully released and they're available for all the customers. There are Extension Preview Features, that are also available to all the customers, but may depend on certain existing features in there. So these ones are close to being released, but they're going to be part of-- part of an extension.

And there's, finally, these insider Preview Features, which are-- they're still in development. They could be released as part of Fusion or as part of an extension, but these features are private-- only available to customers who are part of the insider program. You need to be invited to-- for this because there's more features that are a little bit more in beta mode.

It's important to take into account the basis of Preview feature because you would be able to see that-- that there's still room for improvement. There's actually, at the end, one-- one slide on looking into the future and then a little bit of a wishlist from my side, from me to the developers, so see if we can keep on improving the software.

So talking about Preview Features, how do we make this available? So within Fusion, if we go to the actual profile, which is the little-- your little face or image there for your profile-- if you click in there, you can see Preferences. And then within the Preferences, there's this-- item in here that says, Preview Features. In here, you can see a Fluid Path. And the Fluid Path is cataloged as a public Preview Feature. So it is available for all customers. So you can experiment with that. You can play.

So this type of study will be available through the Generative Design workspace. No surprises in there. And you can see here for those, of you that are not fully familiar with Fusion 360, it does work with Workspaces. So it uses two different environments to allow for different tools, different commands.

There are basically types of actions that you can do to the mobile depending, on the workspace. So there is a workspace that's specific for generative design.

Once we have a model, we access-- we click on generative design. You can see here that you have these two types of studies. One of them is the structural component. This structural component is the one that I mentioned before, as I mentioned that as standard generative design, because it's the one based on structural simulation.

And this has been developed way earlier. It includes more manufacturing requirements, alternatives, criteria, et cetera. So it generates these design alternatives for designs based on manufacturing and structural constraints and on criteria.

Then the other one is the one that we're going to focus on, which is the Fluid Path. So here, what it does is generates a flow path optimized for pressure drop based on geometry, performance requirements, et cetera.

So if we click on Learning More, this takes you directly to our product documentation, where it takes you exactly to the chapter that talks about Fluid Path studies. So it's good. If you want to learn more, you can read through that. There's comprehensive information on generative design studies.

So in terms of the commands, I'm looking at the user interface that we are talking about. This is Fusion 360. That is a flow-control valve. We have gone to the Generative Design Workspace. There's going to be a workflow sequence there that-- it's good to follow it. And that is going to be a lot of commands that need to be fulfilled with information given to the software so we can run and we can generate the-- generate the optimized outcome.

So we're going to see, roughly, a little bit of what these patterns are for. And then we'll see how they apply to this flow control path.

So first, you have-- the very beginning of the sequence, we have our model, and we can edit the model. This is basically an editor on the model that applies only to this workspace. The good thing about the workspaces is you can play around and modify your model a little bit without having to change the overall design that we have in the full design workspace.

The Edit Model enables you to generate little things, create the bodies to represent this [INAUDIBLE] the starting shape geometries in our design problem. So it is good because it gives us quite a few-- quite a few tools. So we can modify, and we can generate the volumes that we-- the volumes that we are after.

We assign the geometry that we want to preserve. So mainly, the areas that are going to be connected with our fluid path are the areas that are going to be preserved. So then we assign this geometry to incorporate some of the bodies into the final shape of the design. So those will not be changing in the generation of the outcomes. So they will always stay the same.

And when we select them, they display in green, as you can-- as you can see. Then we do the same with the geometry that is going to be obstacle geometry. So those bodies-- they represent spaces that need to be empty, that will not have any fluid.

So mainly, we can use these to fix, really, the outer parts where we have solids there. So we definitely don't want the fluid to go into-- or certain obstacles there. In the case of valves, it's going to be important to set-- if we have a proper valve, like this one-- to select the proper the obstacle geometry because no fluid can go through there. That's a solid that's going to be part of my fluid study.

We can assign starting shape.

We're going to give the software a little bit of a guidance on what the initial shape is so then software can apply the optimization and it can apply the study and the fluids to it, but then start to achieve the right criteria for optimization and for checking out the criteria for the fluid. That's going to be also like reducing the volume. The software will start from that shape and then start trying to get closer to the actual target.

Having said that, the strategy is recommended, but it's not really compulsory. There's also Obstacle Offset in there to increase the size of obstacle geometry, sometimes. So sometimes this-- this is OK if we have-- or if we want to model thicker obstacles.

Then we will set up the conditions of the problem, which are-- we need to define where the flow source is-- the openings, so what we call, normally, in fluids-- the boundary conditions. So we apply a flow source-- a flow path entry that simulates the fluid behavior. So we set up the kind of force that drives-- that drives the flow.

So we can set up inlets with flow sources, whether there are flow rate, velocities, et cetera. Or we can set up openings where we set zero pressure, so zero pressure gauge-- so meaning atmospheric pressure, things like that. We can set fluids-- fluid inlets and outlets because that's needed. That's required to define my problem.

After that, it's defined. Then we fix the objectives. We fix the target. So we just optimized. The outcome should be minimize the pressure drop. We specify, also, the volume limit that the outcome should satisfy in there. We specify the volume limit as a percentage of the design volume in the very beginning.

In order to set up a fluid run, we need to set up the fluid material to use in this process. Default material is water. But you can select, also, air as a default material. Or you can select, also, custom material that you want-- that you can use-- that you can use some can tailor yourself.

To compare the performance with different materials there, we need to have different studies, so we can-- we'll see it later. But we have-- we can have clone studies and so on. So we can compare to different materials.

Finally, we get to the Generate setup of commands. First, we have a pre-check-- checks that all the conditions parameters et cetera are going to be set OK to ensure the setup is the right requirements. Then we basically generate the outcomes. So that's the Running button.

We check the job status. There's the list of studies there. You see the generative jobs that have been completed, the ones that are in progress, the ones that fail, if some of them fail, et cetera. Then you have information about the generation of these outcomes.

Finally, we have the post-processing parts, where you have the exploration of design alternatives. The outcomes are using different tools to identify the optimal one. So that's basically the post-processing environment.

So to summarize, we have added the model. So we do the model prepping with geometry tools. We prep for the Generate the regions. Then we have the area where we do the setting up, preprocessing, where we define the characteristics of the study, regions to define where the model will be set up, the boundary conditions and its outlets. We fix the criteria, and we define the materials, so we do all of this.

We generate the runs. We do, first, the sanity checks. Then we set the run parameters. We can see the log of the runs, the progress check. And finally, we explore the outcomes. We take analysis. We take decisions. So we browse and select the options to find an optimal design and the exported.

So we go over to see the example study. First, we're going to-- we're going to take this valve example. So it's a real model there, and we're going to be setting that up. So first thing to set it up is to start with the model and then go through Generative Design, as we saw there.

So we get into the Workspace. There's plenty of tools. We do extrusions. We're going to generate the parts that, later on, we're going to set up as preset geometry. Do that for the outlet, then for the inlet.

You can see that these are standard tools-- standard tools in Fusion. And we can go finish in the Edit Model. Here, we didn't do-- we didn't do that that much because we didn't need to. A lot of things were already defined.

So we then set up the Preserved geometry, Obstacle geometry, and so on. So for Preserved geometry, we are going to get the inlet and outlet extensions. The obstacles are going to be, basically, the solid around our valves.

Also, don't forget, we're going to have the [INAUDIBLE] as well. We'll see that, and select. So there's two bodies. And then we make sure that we have those two selected and they are colored in red.

There, we select the starting shape. You can see the [INAUDIBLE] in red. In yellow, you see the starting check. So we have all of our regions already defined.

Then we go over to setting up the fluid inlets and outlets. I'm going to be selecting the inlet, so setting up 2 meters per second. Then, really easy, we just select outlet. We go for pressure, so zero pressure. So we have our things in there.

Then the next thing will be setting up our criteria. It's going to be objectives. As you can see, the objective is, minimize the pressure drop on the limits. There's the target volume. There's the percentage of design volume. I set up 40%. And we can select materials as well. I have selected the water material.

During the pre-check, you can see that everything is according to plan, and then we can go to Generate. So this thing is very simple. It's very similar, actually, to the standard generative design. Only, the changes on the criteria-- fluid conditions, et cetera-- those are the parts that change because, essentially, the type of study is going to be different. It's not going to be a structural one. It's going to be a fluids one.

We now have everything selected.

OK, we can see that you can clone studies, as we can see there, and that it's really easy to clone a study and change some of the parameters. In this case, in terms of the percentage volume, I changed from 40 to 15, and then I can actually change it to a different thing.

So you then have quite a few studies based on the ones that you cloned. And then you click on Generate. The whole thing starts working, starts running on the cloud. So you get all your studies in the mode of batch running.

And that is pretty advantageous as well because it is nice as everything can run at the same time and they can do the iterations on everything in the cloud. We basically leverage the power of cloud computing.

So yeah, we have our whole model completely done. So then we see the outcomes. So to see how the exploration of the outcomes work-- so we can see an outcome in there. And you can see the shape. The shape is funny there. It has minimized the pressure draw. It has adjusted to the percentage of the target volume that we wanted.

You can actually browse in there through the iterations. So you can see, in the iterations, the different shapes that the software has iterated on. We can show a lot of velocities, pressures, et cetera. You can also see the flow path in there-- the streamlines of particles going through it. You can display the different parameters and volume. You can see the xy plot of different parameters around, and you can see, also, maximum pressure-- pressure, maximum velocity minimum, so a lot of parameters in there that you can study in your outcomes

There's another valve study that I run myself there. And it's important to notice, here, that you can run-- you can run with a coarse image-- finer image for studies. So you can implement that a little bit. You can implement that in different clone studies, as shown in there. You can clone different studies.

And once you clone different studies, you can compare like to like. And you can also compare how the designs look for different parameters of-- well, different targets, actually, for the percentage of the volume that we're going forward.

In this particular case, where it's more of-- it's a different, more elaborated kind of valve, I did run a few different studies-- like three different studies with 40%, 25%, and 15% volume change, always with the target of minimizing the pressure drop, optimizing the pressure drop.

And I found out that, at the very beginning, the things that I set up are the same. Then I set up clones of them, and I would find the image on those studies did work. So you can see that this can definitely make-- can definitely make a difference. You can see here the comparison, really, between the 40% and the 25% of the target volume final outputs.

You can set up different iterations to run, as I showed before. But in this case, what we want to do is we're going to set up different parameters. You can see that the percentage of design volume is a different value.

Then you can right-click and clone again, and you can clone, again, the number one and number two. Then what we're going to do is, for number 3 and 4, we're going to take the study settings, and then we can find a solution, meaning the mass it's going to be better. So you can go there, and you can immediately set things up, even after you run the first one. You can immediately set up and run different-- different studies.

You can click there. It starts running. So it's really useful to do this type of-- to this type of batch running. It is pretty good because it allows you to-- it allows you to do different kind of scenarios there for the target volume and then to see the importance of the mesh-- of the mesh refinement.

Then we are going to explore the design from the outcome. So once we have this final outcome, we can generate. We can export this and generate a full Fusion model from that. You can see that we can use that outcome as a full design. You can see, at the very beginning, we are the outcome exploration, left-hand side. And then, on the right-hand side, you can see already that I have my final design as a model of its own in Fusion.

You can basically see that, if you bring-- if you bring the other parts in there, you can reassembly. Then you can even-- you can even set things up to actually run into-- to run into CFD. If you go into CFD, go to Simulation, Simplify Model, and then Tools.

And then you can launch this into CFD and do a different, more comprehensive CFD study on it, where you can have an already optimized Fluid Path. But then, at the time, you will have-- you will have that CFD so you can apply different set of conditions, et cetera. You can see that part. So that's cool.

Then we can transfer that to CFD, and you can run that. And then you can check the results. These ones are similar to the ones that I get in Fusion 360, but you can see that we are already starting from a path that is fully optimized in there. So our initial model is already an optimal one. It's a pretty good thing.

So looking into the future, what should we-- well, it's not really, what we should expect. But basically, what would be awesome to have-- what we wish to have, looking a little bit to the future?

The fact that you have a single outcome per study doesn't allow us to compare a whole range of options. But this is due to having the criteria pretty tight because we only target the minimum pressure job and we only have one limit in there in my design volume. And also, we have only one material per study. So we cannot-- for per study, we have only one outcome. So the comparison can happen but between different studies.

Then, also, the types of studies are little bit restricted so far because the fluid inlets and outlets and the conditions are still something that could be improved. Again, for the materials, you can compare. But you need to select different studies. It could be something-- you could select multiple ones and do the same as we have in the structure of generative design.

Then, finally, also, in terms of more wishes, it will be good to run with a heat transfer, particularly that came out of the right-hand side, or I have a light in study. It would be good if we could include criteria to maximize or minimize certain areas or-- say, heat conduction and things like that. It would be really awesome for [INAUDIBLE] things. And if the conditions were to be different, we would be able to run more parametric CFD studies.

And then, finally, of course, the very final improvement would be to go for full fluid and structure interaction, where we do a lot of fluids and then we link with the structure. And then we get to run both things at the same time, so combining the structural with the fluids with certain parameters, which will be possibly, maybe, too much to ask, but who knows what can happen in the future?

In terms of further information, just pointing you out to things like the class handout. You can go to the-- you can go to a link on the Autodesk University and look for the session, and there's going to be handout.

There's online help both from Fusion 360, CFD Online help as well. There's YouTube channels both for Fusion 360 and simulation in general. And overall, there's the Autodesk Knowledge Network, which is our main portal that can drive you to different places as well.

With that said, my only-- my only final say would be to say, thanks a lot. Thanks for attending these classes. So, thanks, in a lot of languages in there. So that said, the live session-- there's a Q&A. There's a Q&A here.

So again, thanks. That's it. Hopefully, this has been useful to introduce you to Generative Design, Fluid Path studies. Thanks again.