Interview with Autodesk’s Peter Rogers on Generative Design and the Future of Additive

Peter Rogers is easily one of the most recognizable people in 3D Printing. Underneath the flamboyant hairdo sits a very hardworking business development and sales professional who flits tirelessly from airport to airport throughout the world. Rogers is mainly active in Japan as an award-winning executive for Autodesk there. Speaking and presenting often at conferences on generative design and 3D printing he champions the marriage of software and hardware to usher in the future of manufacturing. Autodesk through its 360 and Netfabb platforms is taking a leading role in trying to make our manufacturing toolchains digital. What is happening right now in software on the factory floor? We interviewed the congenial exec on the future of the enmeshing of the virtual and real.

What brought you to 3D Printing and when did you join?

I have walked a bit more of a strange path to AM than most. After completing my double degree in International Finance and Japanese (the language), I started working in investment banking then digital advertising. After setting up the Japanese operations for the digital marketing platform I was working for, I was invited to join as an international business consultant for a 3D printing startup in Tokyo, which led to me moving into the head of international business for them. After helping them raise close to $10m in capital investment and joining full time, I negotiated a partnership with Autodesk’s Spark platform team, which then lead to an offer from their newly setup Additive team to setup operations in APAC. In total I have been in AM for 6 years now.

What role does software play in 3D printing? 

Software is more and more playing a pivotal role in success with AM. The machines are of course the workhorses in the whole process, but as most companies are coming to understand, the key lies in how to get the best out of these machines. Whether this be through feeding the machine higher value part designs and simulation to eliminate build failures, or build prep to reduce post processing and improved laser strategies, the software has a large role to play in this. All of these help to take AM to full capacity levels. I have purposely avoided the term Mass Production here, as I feel that AM allows companies to move away from this somewhat dated concept, and more towards mass customization, or the ability for such frequent product iterations that most products will move from mass production to short run, distributed manufacturing.

Why don’t we have one click 3D printing yet?

One click printing is slowly becoming more of an achievable goal. Until now, the knowledge that is needed to run a job in AM is very high. This is becoming a bottleneck as the number of highly skilled engineers that can run machines at near perfect rates is unable to keep up with the needs. However, the trick to achieving one-click AM revolves around knowledge of the process and the ability to transfer this into scalable software solutions. This knowledge is growing by the day, and we can see the automation within software starting to bring us closer. As this is an advanced manufacturing technology though, there are still large elements of human intervention. As we have seen from CNC, casting and injection molding, which all still require the human touch, it can take a while to reach the desired “one-click” manufacturing process that we are all keen to see come into fruition.

Why should I get Netfabb?

Netfabb is the Autodesk platform for Additive, with not only all the prerequisite functionality needed in getting the machine running, but it also comes with a bunch of technology to make the hardware run better. Some key elements of this are metal build process simulation for stress and distortion prediction and compensation, latticing and design optimization tools, and automation through scripting of processes. Having all these tools within the same software, and allowing scripting to optimize of process efficiency is one of the key reasons why customers are transitioning to Netfabb. Netfabb also comes with access to Fusion 360, the gateway to the Generative Design product, which is getting a huge amount of traction in the market.

Lattices and distortion simulation in Netfabb

What happened to Within?

The algorithms of Within have now been moved into Netfabb Optimization Utility, and we have coupled that with the integration of Autodesk’s Nastran FE solver to allow for speedy simulation and optimization of lattices and geometry. By integrating these together, the massive delays in transferring large data sets between tools is removed, and the possibility for data loss/misalignment of load cases can be avoided.

Is generative design a silly marketing tool or the future of design?

To be honest, I was skeptical initially as well. However, the idea of generative design is one that is going to cause a paradigm shift in the way we think of design. Currently CAD is computer aided design, where the human does 90% of the work, and the software documents it. What if we could flip that so that the human does 10% of the work, and the time consuming manual design creation and drafting is done for them? Human aided design, if you will. Imagine in the future you are having the software design a part and the human is only required to do some small adjustments to that design? Then the connectivity of that design links into the various manufacturing methods to create it. Automatically updating the molds to injection mold the part, or iteratively adding and adjusting the AM support materials for new part features? I strongly believe that this is where the technology is heading, and I am looking forward to the day where anyone can design things as easily as Tony Stark did when creating his Iron Man suit!

Why isn’t FEA completely automated? 

One of the key reasons is because we need to have information about load cases and understand all the forces that are going to be put on the part. By importing full assemblies and improvements in FEA load case setup UIs, it is getting a lot easier for the engineers to do this. However, to your point of automation, the tricky part is getting the various load cases and then lining them up with the actual design. One thing to look forward to, is more IoT integration. By having sensors embedded in various parts, load cases can be captured while using the product and automatically be plotted. This can then be used for further iterations of the part, and in effect, the part will be using sensory data to develop the next generation of itself.

A series of generatively produced designs.

What advice do you have for companies who want to try out generative design?

You will never hear a company saying “I started too soon”. Considering the opportunities that this technology offers, and the way that it flips traditional part development on its head, getting started as soon as possible will be key. With the costs of getting started being incredibly low too, there really is no excuse. If for some reason it will take a company a lot of time to get started, the best preparation that I can suggest is getting engineers to work closely with simulation teams and manufacturing teams. By understanding how they setup simulations and the various manufacturing technologies limitations, they will be setting themselves up for success with Generative Design once they decide to start using it.

What advice do you have for companies that want to use 3D printing for manufacturing?

Additive is one tool in the toolbox. Part screening is incredibly important in finding the right parts to start producing. If you do not have the inhouse expertise to do this, work with external companies (OEMs, consultants or us!) or engage with universities to develop this knowledge. Manufacturing can be done with service bureaus and OEMs directly without the need for large capital investment in a machine, until you know that you have some applications that make sense to use AM. Use the best tools for the job at hand. Failure often occurs when companies try to hammer in screws.

Can we automate QI and QC processes using scanning and software?

QI and QC has never been more important, so having the right processes and software is vital. As product lifecycles shorten and mass customization becomes more prevalent, the manual elements of connecting the digital designs and physical parts becomes a giant wall to progress. By having an inspection software that can check physical parts and then, if needed, automatically connecting that to iterate CNC tool paths in real time to get the end product within accepted tolerances eliminates a lot of these manual processes. This is what we are doing with some high-end applications developed using our Power-series of products (Powerinspect, Powershape and Powermill).

What role can software play in certification and qualification of parts?

One of the key aspects of certification is having a repeatable process that ensures parts come out the same each time. By using automation functions within our software, it can apply the same process and logic to parts in a repeatable fashion. Considering a lot of the certification isn’t based on software as such, but rather the process, the ability for users to automate almost all activities in Netfabb and apply the same process to various parts, deploy it across various facilities and transfer these processes to suppliers, the process chain remains unbroken.

What is holding back 3D printing? 

Understanding. I know a lot of companies in AM will point the finger at the end users, it is a two way street. While the understanding of how to leverage AM within end users is still quite limited, I think that the AM industry insiders need to also look at the challenges of the customers and better understand the actual hurdles and painpoints of the customer. Bringing an understanding of AM into the product development cycle as early as possible will allow engineers to create designs that best leverage the AM capabilities. As Autodesk, a company with just shy of 700 million designers, engineers and 3D experts using our wide portfolio of software tools globally, I feel we have a strong responsibility to lead the way in this education process.

Can we kill STL already?

Netfabb’s founder, Alex Oster, is also the technical lead for 3MF, which is hopefully going to achieve that. Netfabb itself works mostly with CAD files directly which is great, but unfortunately the need for STLs is still here to stay for now. This is quite strongly driven by hardware though. The software side already has a lot of far better alternatives.