Description
Key Learnings
- Learn how to get started with Dynamo Visual Studio and Project Fractal for MEP design engineering and MEP disciplines coordination by using generative design
- Learn how to use generative design for automation of air-terminal-devices allocation and exploration of thousands of design variations
- Learn how to use generative design for automation of lighting-devices allocation and exploration of thousands of design variations
- Learn how to use generative-design output from Project Fractal into Revit 2018 via Dynamo BIM
Speaker
ALEJANDRO MATA: I think we are ready to go. I hope that you are having a great day at the AU. I'm Alejandro Mata from Ramboll, Denmark. And today, I will be talking about MEP-generated-- MEP Explore for a generative design.
So I'm BIM consultant, and my background is HVAC engineer. And I work with the digitalization station, BIM development, implementation for large project in Ramboll, Denmark. And normally, I work with the buildings, but sometime I work also with a pharma project for pharma production.
So this is the learning objectives. We are going to be talking-- why generative design for MEP? Because everyone talks about it, and it looks really cool, but why we haven't used it yet. Then we are going to have a quick introduction for Project Fractal. Had you used it before?
AUDIENCE: [INAUDIBLE]
ALEJANDRO MATA: How many people in the room are MEP designers? OK. So mostly-- had you used Dynamo before? Is it basic, medium, advanced?
AUDIENCE: Basic.
ALEJANDRO MATA: OK. Great. So then it's going to be really good because my presentation is a getting-it-started guide for Project Fractal. And it's not something really advanced. And then we are going to have the instruction demo, a step-by-step.
And the last part, I would like to use some minutes to show how to utilize and implement, generate, and design in your organization. Because it's not about the technology. It's more about the people and processes. And it really important, also, that we have a look to it, so we have an action plan when you are back to your companies.
So why generative design? I am pretty sure all of you had you seen these really cool videos from Now Today using computational design for simulating a lot of design options for manufacturing drones, cars. This is a door for an Airbus plane.
And then a couple two years ago, after they released this demo, the study case, it was the first time that we started using generative design for architecture, engineering, construction. And here, you can see the different steps. Basically, it first-- generating the design, our design criteria, then evaluating what our work constraints, what our element and simulating all the different options.
Then learning from this result, optimizing-- optimizing the algorithm and [INAUDIBLE] again. And then at the end, we can find the optimal solutions. And basically what it does, generative design-- it gives you the opportunity for having the high value with the lower production cost.
So now, why not-- why not generative design for MEP? For example, we use Dynamo for it because we like-- we would like to automate design processor, why not too generative design? I am pretty sure a lot of you have been dealing with this issue before in project like rooms are always updating.
Ceiling plans are changed again. Then ceiling plans are provided really late. And always, when we have to coordinate all the ceilings, and now we're all MEP installation is really late in the project. And it require a lot of work. So the idea with generative design, it gives you the opportunity to find our best solution, or best guess, as early as possible to have a solid foundation.
So basically what generative design it gives for MEP design and also for architecture and unit in construction is like-- it take us away from time and material business model. Because normally, when we do project, we bill our client by hours. And then also, it's a methodology for surviving in a burning platform.
We are in a burning platform. Like every time our tenders are cheaper, we have to work more higher requirement. So this generative design-- it will give us the opportunity to be in business.
And as I said before, the main three keys that you-- I hope, like when you are finished-- finished from the class, you get back to your companies. Generative design-- it provides higher quality deliverables, optimized solution. Then a speed in design processes actually is delivering fast with confident.
This comes from the agile methodology that is used in the software company and basically is done to be too afraid to take a decision in the early process. Because our application has this knowledge, intelligent, that we have paths to going forward. And the last part is for lower production costs, because it's an automating process.
So now we are going to getting in started with generative design for MEP designers. So as I said before, I hope you to take this thing to your conference-- is people and process first, technology second. I hope like when you-- after the instruction demo, you can see it's really simple to automate this process. But we don't have a standardized process, and we train people with the right mindset and skills. Technology will not help us out.
So here, I have an illustration like how is generative design. Basically, to have a success with generative design, we need to have a solid foundation of standard processes, digital content, digital standards. And then when we have a solid foundation of information our models, so then we start using this information for automating, exploring with generative design, exploring again, optimizing the solution. Because everything that we have in the model, all the data is a structure.
And then, as I say again, it really important we will talk about the people, because we have to-- it has to be integrated. So now here, I have a video for you understand. This is Detroit in the 1950s, how it was-- the production of cars.
You can see this guy is trying to-- trying to-- really manual to try to tie the bolts, and then here are people that start working. [INAUDIBLE] are really busy. And then in the [INAUDIBLE], you can see this is from Japan, from Honda. And you can see, there is a lot of robots that working by themselves. But then you can see people working in collaboration with robots.
And this is kind of the same with generative design. I am pretty sure if we give this robot in the 1950s, they would totally fail. Because to have success with the robots and with generative design scripts, our process has to be standardized. And our design engineers, project leaders, technical drafter, they had to-- they had to work according our design process.
So now, the idea with this generative design is normally when we have the different project phases, we lose information, because we get a lot of design changes. And then here, what we want to do is, as early as possible-- in the conceptual design is where generative design can give us value. And then when we get in the next phases and main project, it will give us also-- we would have to use automation based on the foundation that we have gone.
But in reality what we are experiencing is every year, after AU, there is a lot of automation classes, everyone going to automate everything. But the end what we do is-- we are working even more because we try early. And then we get to them in project. And we [INAUDIBLE]. We are not finished. So this is what we shouldn't do.
So now, I'm going to start with introduction with Project Fractal. This is kind of [INAUDIBLE] generative design Cloud platform for generative design. It's free. However, you need to have a Dynamo Studio license. It is really cheap in comparison with Revit. And this project has been available since 2016.
The challenge before-- it wasn't too much knowledge, so it was difficult to start. But during the last year, there is a lot of architects, engineers that have started using it. And you can log on with your Autodesk account.
So here, this is the process. Dynamo Studio-- the difference between Dynamo Studio and normal Dynamo is Dynamo Studio doesn't have any connection to Revit. So normally, the first part is you have Dynamo Studio. You install it. Then you sing up to Project Fractal. You send a mail to the account, and then you can-- you can start using Dynamo Studio and send in your generative design application to the Cloud, to start exploring your solutions.
So here, this is the framework between Revit models and Project Fractal, and then Revit model again. So normally, when we have a project, what we do is we need to export some information, some geometry, to Dynamo Studio. And then from Dynamo Studio, we can have the opportunity to structure this data. And when we have it structured, we can send it-- and parametrize also-- we can send it to Project Fractal.
We do all the design solution, the design generation. And then when we have found the best solution, we can download CSV file. And then with normal Dynamo, we can start-- we can start generating this solution within Revit. Basically, in Project Fractal, what we do is visualization and analysis. And then we documentate in the model.
So this instructional demo is going to be focused for a really-- a simple prototype for allocation of air terminals, which is the-- also the same workflow for our electrical features and lighting devices. So normally, here you have all the workflow that we are going to go through, step-by-steps. I'm going to show the different structure of the script.
I'm going to show also the videos. So normally is export from Revit to Dynamo. Dynamo will make some generative design families. And then when we have these families, we can start playing with our surface geometry. And when we have the design that we want to simulate, we send it to the Cloud, to Project Fractal. And then when we have found our solution, we get it back to Revit.
Here is the first step. Normally, we can use Dynamo Player. And basically is what it does is it gets an space for-- from the project and export the ceiling, the surface. We don't need to have architectural ceiling. It get just the top-- the top surface of the space. And then, also, it export as a [INAUDIBLE] information. And the information that export is the air supply requirement.
So here you have the structure-- is first we have the relative spaces in the active view. You don't need to open this screen. You can run it from the Dynamo Player. And normally, it's a good idea sometime to start-- when you start using the generative design, approach to start small. And sometimes, it's better just to export one room and try playing with that. And when you feel comfortable, you can export the whole level or the whole building.
Then when we get this space, we get the [INAUDIBLE] file, which is the one-- only the one that is compatible with Dynamo Studio. And then from this space, we export the information that we want to have to make some flow calculation. So even though the Dynamo Studio is not linked with Revit, we can play with that.
So here, the first-- the first step for doing generative design-- is creating our generative design families is like-- and this is-- I call it-- like the-- is really through the [INAUDIBLE] we go. What we do is we take a specification sheet from the producer that we want to simulate. We get all the formula from the get distribution flow, and then we can replicate this behavior in Dynamo Studio to find out how much [INAUDIBLE] I have and then in relation to the flow for this terminal.
And here on the left-hand side, you can see how it looks in [INAUDIBLE]. So here, you can see the structure of the script, is first, we replicate the different types of air terminals. We do one way, two way, three ways. It can be different. You don't need to do everything, but it depends how many different variables you want to simulate.
And then when we have this surface, we play with its conditions, because we don't want to visualize everything. We want to only visualize one design solution on the time. Otherwise, we cannot use the web servers to understand what is the impact of the solution.
And then the last part is just to put some color. It's not important, but you can play it also with that. Here is the really kind of geometrical problem, is getting from the center of their terminal and calculating the different points, drawing some linear lines, and then creating an arc. And as you can see, I have done these with a designer script. It a programming language behind Dynamo, but you can do the same workflow using the standard Dynamo nodes.
The only thing is that sometimes, it can locally be really messy, and it difficult to manage them. But for starting out, it totally fine. So here, you can see when we are finished, how-- this is actually our generated-- our computational design family is from a manufacturer sheet, and you can see basically what you simulate-- but it doesn't simulate anything. It does have the different air terminal device ways, air rotation angle at a number of ways, and also the surface.
But this is not a generated design family, because it doesn't move. It just only change the geometry, so that's why-- the next step is which feature makes an object be generated design friendly? It's not just to making a computational design object. It had to be moved, because that's why it's going to simulate the different design options. So basically, in this part, we have taken the same code that we have defined before.
And the only thing that we had changed here, you can see before, it was a origin 000. And now it's based with variable, which is a point. And these unknown points, it will come from our ceiling. And this is what makes to move, because we don't know which point it's going to be. But once that we get it, it would simulate a location. So it really important our family can move.
So here is the same family than before, but only with the point. So you can see now we are changing the rotation, the distance, and the number of weight. And this is where it makes to make generator design for MEP. It has to move within spaces. And now you can see here, you have all the design parameters. And this family gives you 122 design options, and then you can start simulating.
This is a really good practice to do it, because you cannot expect to just generate all the solution when you are finished with your whole script. Is that a good idea like every time that you develop something, you start trying and checking? Is it good or not? And then you do try to do better again. And then when it is good enough, you go the next step.
So now, next step is because the idea of a generated design for MEP is to create a ceiling. It can be greater ceiling or can be anything; either to give a solution, a design solution as early as possible with the less information that we have. So the idea is we don't have to wait for the architects. We can give our best proposal. And then when we get the more information, we will put it in the script, and we will update our solution. By doing that, we are telling that we need some space for our terminal.
So basically, what it does this part of the script, it takes to top surface of this space, and then it start generating our best, our optimal ceiling plan. What would be the best ceiling plan for us? So this is an inspiration I took from our design processes in Ramboll, Denmark. Normally, when our engineers, they model air terminal device, what they do, they do model manually with MagiCAD. And actually, you can see here, they use on a space or sometimes, they draw an idea, and then they got--
This is an idea actually and then you can define you are ready for that terminal device based on minimum distances, such distances to the center or number of distances. This example will be only for in relation to the exact distance to the center, because that was our prototype. We didn't want to do everything. We wanted to test it out to see what is the opportunities. And then when we do this with the planning, it generates their terminals.
But they think this process can be really annoying, because you need to check in. You place a terminal. If it's not good enough, you have to remove it, do it again, remove it to take in. So you have to maintain a solution, which is not the final one. So here, what we are going to do is the same, but we are going to do it on the cloud without modeling this element in the model until we found the best solution.
So basically, what we do is now you can see we are in Dynamo Studio. So the first step in Dynamo Studio is we need to import our basic geometry, because it doesn't talk to grid. So when we have imported, we can start doing that were our grid for our ceiling. And then these grid values, it's going to be a variable, because this is what we want to find. And then the rest of this grid is just a mathematical challenge.
It basically is because we are calculating our maximum distance from the center of the room. So we have first, we do something. The first that we do our way, we need to find the center, and we do our coordinate system from there. And the last part, we do the point generation.
So now here, you can see this is a solution with the script. You can see the first value. I have a fixed distance from my wall, because this is our wandering condition. And then from there, the program factors start fluctuating between x-value and y-value, and they start generating all the random solution. You can see here sometimes, it gives you, like, a [INAUDIBLE] and that's-- [INAUDIBLE] is OK. When you get [INAUDIBLE] it means the solution is not believable. It doesn't comply with your constraint that you have defined in the program.
So now, when we have this point, what we need to do is to explore the layout of air terminals because now we have our ceiling. We have our point where we are going to place the terminals, but we need to find out the distribution error based on the flow on the number of devices that you have on this space, and to see how much area it occupies from your roof-- from your ceiling to find the best solution.
So basically, the top part of this script is the one we did before. So now we are combining-- we have the one from the ceiling is here, then we have our generated design families that are going to be placed on the point that we had generated before. And then we get this CSV file that we-- it was the one thing that we export in the first step, together with our geometry. And this CSV file is going to be used to find the optimal flow for each terminal based on your supply requirement for that room.
And the last part is we are going to have, looking in relation to the performance analysis-- and this is really flexible. It depends on you, based on your design criteria and what you like. It's basically to quantify-- to give a score of what you think is a good solution or a bad solution. And we will have a look when we get there.
So now we have our ceiling point. So basically, on our terminal devices-- so the first step is calculating how much flow should be-- the terminal. So basically, it's just a really simple formula here on the right hand side, and then joining this part with the point.
So next step is what is a good solution? A bad solution? What I want to achieve for my project-- what the client wants. So normally, the first part when we do a generated design, before we start doing anything is that-- to think to yourself, what I want to accomplish. And normally-- in this example, basically what I want to do with this prototype is I want to get the high performance for my space. I want to have the higher comfort and I want to have the material production of the better price for my room.
And then here, you have-- the metrics are the parameters that you are going to measure to achieve these goals, and then elements are the geometry objects that are your constraints, which in this case, can be air terminal devices, lighting features, ceiling, walls, and MEP spaces. These are our constraints for this example.
And the last part is here-- is how we do our-- how do we convert this diagram to a specific value? So the thing-- the approach that I follow is I can have the flow from the step before to see how much is the flow, and then I have here, also, the impact area ratio is basically my distribution error divided for the total surface of my ceiling. So what I want to have is to find my air terminal to give the air as much as possible with the less element.
And then what I did instead of a negative remark, I did, like, a bonus score and say, if my result is less than this particular limit, so then I get 5 points. It can be 10. It can be 100. It doesn't really matter. It just to-- to know because in this formula, what we do is we make, like, a sum of this element so that we have the flow multiplied by 0.1, then the area ratio multiplied by 0.1. We summate it, but the thing is, when you have a lot of design solutions, the result that you are going to get for your score is going to be really similar. So then, you need to have-- what is it? What do you want to achieve? Because if it's your result below these design criteria, it will give you a really good score or a really bad score. So it's easier to filter out this information in your result.
So now here, we are going to have the demo from the whole process for these prototypes. So actually, you can see here, what I am doing-- I'm defining my wall offset value, which is-- right now, is 1,500. And you can see now, I have it, like, 1,400 design. But if I define by default, this wall offset value-- it reduces to 486. So that's a good idea, like, to know if you have some fixed design criteria. So you're defining before running this [INAUDIBLE] because it's going to be faster.
So now you can see here, I'm starting getting a really weird solution because my distance between the two [INAUDIBLE] is really small. So you can see this is not really buildable, and this is the thing that I showed before. It's when we do generated design the first time, then we have to tune it again. So actually, what I have done here-- I have increased my minimum grade value, so I don't really close the component. So it helps the program to know where are your boundary conditions-- your limits. So you don't get really bad solution options.
Here. So now, you can see I start doing this [INAUDIBLE] before because I don't want to have 1,500 options. And then, I start doing-- to focusing in just number of way-- 3. And then the program-- it will start generating all the design solutions. Normally, it takes, like, 10, 20 seconds for each solution, and you can leave the computer. And then after one hour, you can have a look at it.
So you can see everything is starting generating, and here in the top, we have our app design criteria and values of our performance indicators. They will give-- it will help us to find the best solution. So now what you can do here is you can start filtering out this information. And it's really simple-- you have drag and drop, and then you have the solutions that are within that range.
So now, I'm looking to the area ratio, performance score, average air flow-- lower than 50 liters per second. And then, I start-- here, you can also define your favorite, and then you can sort it out by time, by some particular indicator. And here also, you sometimes-- here, what it makes-- it makes, like, a cross product. But you can also say, OK, I don't want to use time to define these design criteria. So I can put it, like, random, and the program will find any solution, which is sometimes not a good idea because you also really get really bad-- something that doesn't really make sense.
And the last part is here-- is when we have found this favorite component. You can see all the other parameters that make the solution feasible. So then we have the opportunity to download this data to our local computer, so then we can model these solutions in our models.
So you can see here is just a really simple file. You have the data-- the values of these parameters. So now is what I said before is from-- you can see here, we are going on the way back to our model. It's from fractal to prior documentation, and the thing is now, I am switching to normal dynamo because the one who has the connection with Revit. And here, I'm using the same script that I used before for generating these grid lines and this point.
And the only thing that I am doing-- I am using this part-- it basically is inserting a family by point and selecting the family type. And when I have defined those two parameters here, I need to-- here in this example, I put it manual-- the result that you got from the CSV file. But you can also do it kind of more automated. You could have the opportunity to read the CSV file and say, OK, I want to do the whole process automated.
So here is-- this is really simple. It's how it looks in Revit. So what we do is we have the dynamo player. We don't need to do any-- open any program. And here we have our design result that we got from prior fractal. And this is the one we had to define. So here, you can see on the top, we have our file path to the CSV file, while you can also do it manually and then define here-- no, this one is the geometry. But also, you have, here, the axis that we got here from this file.
So what are the findings from this process? The finding that we found is a generative design can also be used for MEP, and it doesn't really need to be a really fancy solution like the Autodesk video. It had to be something really simple, like we show here. It can be also used for normal calculation, and then you can also visualize it for in MEP spaces. And then also, what we found here is, like, this process is-- it should be taking-- it should be adopted by you guys-- by the designers. It's not like a tool like a technical drafter can use. It's a help for your design process to gauge this speed-- and confident to take decision in early phases.
Then we also found, like, some solutions that-- it was not really good. It's like right now, the technology is really not robust enough. So it is really difficult to, if you are a company more than 21,000 people, it's really difficult to scale up these solutions because you cannot really compile these scripts. So you cannot protect your property-- intellectual rights. And also, the thing is, the process is really manual. You can see, like, you have a lot of steps. And the-- also, thing that we found is if-- the basics-- they need to be in place. I mean, if people don't really know how to model that space or how to put the right information on how to make the boundaries condition. So if you don't have a model, which has been modeled right, so then you cannot really use it.
So here is like a-- I don't know if you noticed before, but what we did before is we didn't show-- we didn't, in our [INAUDIBLE] we didn't model a precise air terminal. What we did is just a place holder because what we want to do in early phase is we want to reserve-- we want to make our airspace reservation for the other disciplines. And we don't know yet which, exactly, model we are going to use. We know the functionalities.
So that's why we use this box. And the idea with a place holder is, like, you make your level of information higher and higher when you are going through the different phases. And also, the idea with this place holder is, like, your place holder occupies more space than a real component. And then when you have defined the real component, you remove the placeholder. And this is going to be the last solution. So we can go more like a progressive enhancement of the level of information.
Here, you might be thinking, how-- we are going to use this place holder if changes happen a lot. And the thing here you can see, like a generated design, is for early phases. The ones that we are getting more in detailed design-- generated in design-- it doesn't really give you the opportunities, and it makes it difficult to generate a design for detailed design. So here, it's more about using MEP automation with a normal dynamo or can be-- some scripts can be whatever.
So here, for example, this is-- probably, you know, this is Cesar Scalante. He made, like, a really interesting script for automating this change management process for a terminal. Here, you can watch the video, and also download the script. And the thing-- what he did is-- what if I have generated my ceiling solution, then if something changed, should I remove it or should I update it? What should I do? So basically, what it does is instead of going back to the generated design process of finding the right solution, it finds what is the right model to place my air terminal, sprinkler, [INAUDIBLE] lighting feature. So the idea with this process is like a gateway. So when we go on the airport and we go to this gateway, we cannot go back. So we don't waste time.
Yeah, this is also like a development project that we have been doing. We have done some in Rambolls and workshops with Autodesk for exploring advanced MEP design. And here, what we did was just to automate in detailed design phase the placing of our [INAUDIBLE]. So basically, what we found is we found the Windows that we have in our phase-- we had the heating lows on the spaces and then we can find which radiators are the right one.
Here, now, is the full-- placing lights over the desk. Now we are in detailed design. The architect had defined where to read all the furniture. So this is what we did. So now where we are going to look here is for the-- we call it our fixture, and relative to just switches-- lighting switches. And what it makes-- it looks to the doors and it looks in which way the door is open, and then you define how much distance should be from the door and the height, and then you place it. And the cool thing is you can do it-- kind of a simulation. If you don't close dynamo, you can change the height on the offset and all the lighting switches in you project will-- it will be updated here. You can also change the type, and now you can see it has updated.
So now is what are our findings from this process is. As I said before, it's something that can be used, but also, process and people come first. And this is something that happened to me every time that I come here to Autodesk University-- is the first week after you are back to work, you are really motivated. You have a lot of ideas. And then after two weeks, you forgot almost everything-- the motivation, and why? Why, if we have all that amazing technology, the really cool presentation from Autodesk, the program-- why don't we implement it? Why don't we utilize it? And it's because what we see here is more like a proof of concept. It's what is possible-- what are our opportunities. But when we go to our company, where we need to do this foundation and solid process of standardizing, making sure we take one step at a time, and we're going up-- solid. Otherwise, not sustainable.
And the thing, that I also said before is, like, it's really a challenge now to scale up this generated design process because the Autodesk, for some part, in the last two years-- period factor is the same. Right now, it's the same as it was in 2016, probably because I think what they are going to do in the future-- they are going to move. They are going to integrate Project Fractal to Forge, and then you're going to have the opportunity to do the same workflow, but without exporting and importing to your local file. So it will have the opportunity to talk directly from Project Fractal to BIM 360, and everything is on the same environment.
So now, the last part-- we have 12 minutes, and the last part-- I'm almost finished, but I would like to show this part. It's implementation and moving forward. So now, what I show here is just the first step-- it's the prototyping. So we are in this prototyping phase. And now we need to start using a more-- like how we use this prototype in our pilot period so we can monetize it. We can have some value in our project-- get some feedback, make it better, and going moving forward.
And now, in this year, in the last six months, there are some startups that are doing a lot of really cool things in this technology. And I think when we will see the benefit of the design process, they're going to be around two, three years, I would say.
AUDIENCE: [INAUDIBLE]
ALEJANDRO MATA: Yes.
AUDIENCE: [INAUDIBLE]
ALEJANDRO MATA: OK. Thank you. And here, you have the opportunities-- the implementation road map opportunities like generated technology-- generated design technology and design process that can help a lot to improve our quality and project cost.
But to get this value from the generated design approach, we shouldn't follow this waterfall traditional process. It's normally in the construction industry. We don't tend to use a tool or a new design process until it's really, really robust. And we have really small risk. It's more about building the application, testing out, learning, doing better, and going forward. And what we did with this prototype-- it was instead of looking to the wide side and saying, how can I achieve everything was OK. Instead of trying to cover all the functional stuff, we will cover a little bit from functional, reliable, usable, and [INAUDIBLE] design. So it gives some value. It's more about getting happy step by step right until the end.
So now, here, I will show some start ups, which are doing really, really cool stuff. This is Hypar from Antony Hauck. He was the former generated design director from Autodesk. Now, he's doing his own company, and here, basically what he's doing is the same concept of Project Fractal, having, really, cloud computing. He is trying to-- speeding up these design generation options. But also, what this is, which is really interesting about this platform is you own your script. You know you own your code. So basically, the platform is for generating the solution. But all the script-- all the protection of intellectual property rights is yours, which we don't have now on Project Fractal. Everything is open source and everyone can download your script, and that's not very good.
And here, this is the next step for generated design. As I showed before, generated design worked really good for conceptual design, but when we get to the detailed design, it gets really tricky. It's because we have a lot of design construction. We need to find a specific design, and actually, for example, which is this company doing-- Entazz-- they are doing synthesis design. Synthesis design is the next step from generated design. Synthesis design-- what it makes is, instead of generating all the solutions that are possible for your room-- only generate the one who complies with your project requirements or your building code. So actually, you can find the best solution faster and use less time exploring and filtering the solution because you only get solutions that follow these rules.
Now, for example, now, the prototype they have-- they have, usually, a lot of this methodology for hospital and prefabrication models because it's really easy to code the design process, which is not flexible for generative design.
And here-- last part-- if you'd like to know more for generative design, these three classes are really good. It was two years ago-- it was from the crane-- from the construction site, finding the optimal location and finding where the trucks should go in, should go on out. Last year was this one, also from the company from Denmark, MTH. They were using the same concept to generating automatically a parking place. And here, this is from the last year-- it was really good. It was the prototype from the exhibition halls. Actually, Autodesk used this design process to find the best design solution, and optimized the square meters for the sponsoring in the exhibition halls. That's it. Thank you a lot. I hope that you enjoyed.
[APPLAUSE]
Downloads
Tags
Product | |
Industries | |
Topics |