We've talked about using the IoT to design buildings and we've talked about designing a bridge, houses and a motorcycle swingarm with generative design. Let's take that to the next level and look at designing a car!
Introducing the Hackrod!
With the Bandito Brothers, we noodled on the idea that three kids in a dorm room could start a car company and showed off our progress at Autodesk University.
The chasis you see above was wired up with sensors to gather data on the forces that the car goes through as it's being driven. Just like with project Dasher, we started with a scan of the object so we can plan where and how to attach the sensors. All of this data is captured and visualized for the next part of the plan - generating a new chasis with project Dreamcatcher. The chasis comes out looking like an alien skeleton - some people like this and some people don't.
From a design standpoint, Dreamcatcher is handling all the complex math to make a good structure and then the designer can get to work making a "cover" that meets whatever aesthetic criteria is important.
Over the last several decades, generative design techniques have enabled designers and engineers to broaden their exploration of topology and performance of human-scale structural forms in Architecture. Autodesk is collaborating with Lawrence Livermore National Labs to extend this exploration to micro-architecture and how to design materials at the microscopic level. The researchers intend to generate and analyze the performance of very large sets – thousands to tens of thousands – of different structural configurations of material microarchitectures using generative (aka computational) techniques. Helmet design is an excellent example of a multi-objective design problem where constraining for weight, cost, durability, material thickness, and response to compression and sheer within the range of impact conditions will produce multiple high-performing material configurations.
Likewise, helmet design stands to advance considerably from additive manufacturing. The internal structures of helmets not only need to be lightweight, but also must absorb impact and dissipate energy predictably. Advanced additive manufacturing techniques can produce complex material microstructures that will dissipate energy more predictably and repeatedly than what is currently possible with traditionally manufactured helmet pads such as foams and gels. When paired with advanced computational design methods, additive manufacturing opens up the opportunity for a functionally graded multi-material design that integrates the helmet shell with its cushioning element. A fully validated, 100 percent additive helmet is an audacious goal, yet this collaboration expects demonstrable progress toward a prototype.
Erin Bradner from the Dreamcatcher team explains more about this exciting project in the following video from Wired exploring the future of football and dealing with concussions.
John asked some interesting questions in exploring this project:
Can design be automated? What does that mean for the role of designer?
And then he found an answers to his question in the designer letting go and letting the software do what it does best. Going on to say:
It's going to be an amazing dialog between humans and machines - figuring out the best possible solution for a design
John started of with sketching his project and and goals for the design.
He scanned his arm and hand with reality capture to define one of the goals so that there would be a custom fit. Other parts defined below include a space for the arrow and the upper and lower limbs and the bolt holes to mount them.
With this set, Dreamcatcher can begin the simulation. At each step, the bow is tested for strength.
Final the digital part is sent to the CNC milling machine. John says:
This is like watching the evolution of bone structures or plant structures
The completed part gets removed from the machine.
You can learn more about how John did this on Instructables and through the video below.
For those attending Autodesk University this year in Las Vegas, Autodesk Research will have a booth in the “Central Park” section of the Exhibit Hall where we’ll be showcasing a number of exciting projects.
The projects represented at this year’s conference will include:
The Bio/Nano Research group will be showing the current status of their research on how to fold DNA to create functional nanostructures as well as how to grow artificial bones.
Autodesk Within Medical, which allows implant designers to create porous coatings to aid bone and implant fusion (ie. osseointergration), will be displaying a number of their 3D printed medical components and explaining how their technology works.
When you enter Sands Hall B & C, just walk to the Central Park and Autodesk Research will be on the right!
In addition to the booth, look for the Hive Project near the Exhibit Hall where Autodesk University attendees will build an architectural scale pavilion guided by human/robot interaction.
A number of team members will be giving talks at AU:
Composite Materials and Manufacturing Processes for Automotive Applications
Massimiliano Moruzzi presents an end-to-end solution for the automated composite manufacturing process. This class will cover advanced lay-up design strategies such as fiber placement, tape layering, and robotics lay-up which are utilized when programming automatic material layup equipment. High composite production rates will be covered through automated robotic material nesting and taping.
Cultivating Innovation and Developing Intrapreneurs
Wednesday, Dec 2, 10:00 AM - 11:30 AM, Location: Zeno 4704, Level 4
Cory Mogk will be doing a talk on Cultivating Innovation and Developing Intrapreneurs that uses the tools from the Innovation Workshop. This class will talk about how Autodesk is helping intrapreneurs develop their ideas and we’ll provide tools and guidance that attendees can use on their own or in their organizations.
Composite Manufacturing Solution for Optimum Material Nesting and Ply Layup
Thursday, Dec 3, 2:45 PM - 4:00 PM, Location: San Polo 3405, Level 3
Massimiliano Moruzzi will lead this two-part class where attendees will utilize Autodesk TruNest Composites to show the complete process from import to nesting to NC part cutting of ply materials. Special focus will be given to optimal nesting for efficient material usage. During the second half, we will utilize Autodesk TruLaser to perform laser projection for showing composite ply lay-up.
Once again, the Design Research team will be conducting user research sessions. This year’s focus will be on collecting feedback for Withinand Dreamcatcher. Look for the OCTO Airstream in the AU registration area.
We hope you’ll make some time to come by and meet some of the team.
The Dreamcatcher team is helping MX3D with their design for a bridge that will be 3d printed in place by robots.
Autodesk CEO Carl Bass says that one of the really cool things about this project is that it will happen in public - not behind closed doors in a lab. Doing this project in public makes it more complicated and risky which increases the chances of learning new things.
You can see the novel printing process that MX3D has developed below. They have a multi-axis industrial robot hooked up to a robot via custom software.
Before MX3D developed their metal printing process, they perfected a resin-based method. This super fast curing resin neutralizes the effect of gravity during the printing process - the structure keeps it shape without drooping or sagging.
This may take a couple years to complete but should be fun to watch.
The finished bridge may end up looking like this model made with Dreamcatcher. The organic, tree-like structure fits nicely into the natural environment of a park.
ABC7 News in San Francisco put a nice story together on how the Dreamcatcher team is teaming up with Lawrence Livermore National Laboratory (LLNL) on generative design and material science. The team at LLNL is working on printing materials 1/10th the width of a human hair. Together the teams are considering what this could do for bicycle helmets.
Question: When does a motorcycle swingarm look like a pelvic bone?
Answer: When it's designed with Project Dreamcatcher!
A swingarm is the main component of the rear suspension of a motorcycle. It attaches the rear wheel to the motorcycle. The swingarms you see below are designed with Project Dreamcatcher and get their organic shape as the system iteratively tests the strength of the piece and removes unnecessary material as you can see below.
To set up for this simulation a designer needs to specify their objectives. In this case, the objectives include the forces, the bounding space for the swingarm (as seen in the initial state above - effectively stating that the finished solution must live within this space), the connection points (where the swingarm connects to the wheel and motorcyle) and objects that must be considered in the space (the wheels and chain).
Connection points for the swingarm
Obstacles for the swingarm - a chain is placed on both sides to create a symmetrical result
Dreamcatcher can produce many options for a designer to choose from. Here are some alternative swingarms.
From these options a designer could then decide to do further work, such as:
change the shape if they want something less organic and more traditional looking
develop wings for a footrest or saddlebags
add decorations like an embossed logo
Dreamcatcher is a collaboration between the Design Research and Computational Science groups at Autodesk Research. The Computational Science group is looking at the simulation and generation of these shapes using high performance computing options like GPU's and the cloud. The Design Research group is exploring the user experience for designers and how to push beyond the limits of what is possible today. This makes for a lot of exciting possibilities with Project Dreamcatcher - what would you like to design?
The design of houses has become a highly mechanized process with few houses having direct involvement with an architect or licensed design professional. A few home plans are created and then copied, mirrored and rotated to create standardized communities.
The Housing Agency System as outlined would take design criteria such as family needs and climate conditions as goals and create homes from components defined in a parametric Building Information Model (BIM). This would both increase the number of options available and increase the viability of mass customization, it may even help to avoid the standardized blandness that has enveloped our suburbs and exurbs.
With the HAS, the following criteria could be considered in the generation of home plans:
Site Model: Topography, obstructions and trees, market value of adjacent homes and location of utility inputs are examples of required information for the site model.
Planning Strategies: Planning for one or more houses simultaneously and account for interaction between the various units, the frequency of pedestrian pathways, parks setbacks and density standards.
Construction Systems: Construction systems such as light frame, heavy timber and light gauge steel construction systems.
Building Components: Pre-fabricated or manufactured elements like windows, doors, stairs and mechanical equipment.
Once solutions are created, shareholders can review and decide if they are happy with the results. If they are happy, planning documents can be created.
In the simulation phase outlined above, there are a number of things that could be evaluated, including:
Structural Simulations: Determine if the solutions meet requirements for gravity, wind and seismic loading.
Cost and Schedule Simulations: Cost simulations evaluate the first and approximate lifecycle costs of a building.
Flood, Fire and Code Simulations: Analyze the risk of flood and fire in addition to tests for code violations.
Stakeholders - including designers, contractors, and clients - could adjust design criteria to explore the space as shown below.
The HAS is not a simple system to create but it can be helped through simplified mobile computing, public interest in design, increased industry collaboration through BIM and elastic cloud computing. Such a system could help to inform the general public of the value that architects add to the design of homes and provides a venue for interaction and advertisement can help architects regain some of their involvement in this market.
The Research team will be displaying some of their work and views on the future in the Exhibit Hall and you are cordially invited to come by, have a look, be inspired and share your feedback.
In the Exhibit Hall, you'll find people and displays for the following projects:
Draco and Kitty
Autodesk is researching how design tools can be applied to synthetic biology, problems like fighting diseases, such as cancer, and improving drug discovery.
Draco and Kitty
Answering the challenge to make animation (Draco) and authoring interactive content (Kitty) as easy as drawing, you not only see this in action but try out it out for yourself.
Showing that computers can help you design - not just produce design documentation - structurally sound and interesting pieces based on your specified goals.
If you missed Hy-Fi on display at New York's MoMA, you can get a little taste of it at AU. Haven't heard of Hy-Fi or its creators The Living? Check out this video showing Hy-Fi and some of what The Living are doing.
The Autodesk University Exhibit Hall will be open at the following times:
Tuesday, December 2: 6:30 p.m. - 9:30 p.m. for the Community Reception
Wednesday, December 3: 11:30 a.m. - 3:00 p.m.
Wednesday, December 3: 6:30 p.m. - 9:30 p.m. for the AUGI Reception
Thursday, December 4: 11:30 a.m. - 3:00 p.m.
Beyond the Exhibit Hall, there will be a number of presentations from Research team members:
The Design Computation Symposium will explore how advanced firms are bridging the gap between Computational Design and Building Information Modeling. Speaker topics will include both pragmatic aspects of digital design in daily practice, and forward thinking ideas and research. There are three main areas of interest under this theme:
Performance-based design, simulation and analysis.
Would you like to get your designs out of the screen and into your hands? While 3D printing has become an exceedingly useful tool for demonstrating and prototyping design ideas, preparing files for 3D printing can be frustrating and time consuming. In this 90-minute course we will generate a complex surface in the Fusion 360 3D CAD design app that takes advantage of the T-Splines modeling technology. We will bring this model into Revit software where it will serve as the base for a panelized solid form using the Dynamo visual programming language extension. Once we have generated the complex parametric model to the required specifications, we will export the model to a STL file for 3D printing. A 2-step process of healing the mesh for optimal printing is described with the meshmixer tool and Project Miller. Finally, we will inspect the mesh and prepare it for output to various 3D printing platforms.