Chandrakant Patel serves as HP’s Chief Engineer and Senior Fellow, and is well known around Silicon Valley. If you have any doubts to how important he has been to the company, take a stroll through HP Labs, where one wall contains plaques for all the patents awarded to employees in three incredibly long rows. The top row? All Patel’s.
Patel recently discussed his thoughts on the future of manufacturing, how 3D printing will change how and where we build things, and more. Here are six takeaways from the brilliant and engaging mechanical engineer.
- Manufacturing will continue to evolve. “I started on the drafting table in 1983,” Patel said. “I’ve gone through all the CAD, the evolution of CAD, gone through evolution of CAE. We are now entering a new world where the whole computer-aided design will change. It will get democratized. More people will have access to design by the tools we use. And design will occur in a traditional way—as an engineer, you may design a 3D part, you may analyze a 3D part. Design may occur by scanning [an existing part]. You make big parts that have been printed already and a parts library exists.”
This concept leads to some interesting ideas. It’s not hard to imagine a future with millions of parts in a learned parts library. People have built them, characterized them, and these parts are available—likely via the cloud.
- 3D printing allows for flexibility. Patel is passionate about HP’s Multi Jet Fusion (MJF) technology, which makes use of powder as a raw material. “As we are building a part, we can build an array of parts. We can orient them in very different ways to get the maximum number of parts on a single bed. You can build a different array of parts. The print part goes and scans back and forth, putting droplets where we want that area to fuse. The bed itself is pre-heated to a given temperature, then we come in and the areas we want to fuse, we deposit the heat-absorbing ink, or agent. The parts we create have structural qualities of a part created by other means, and that’s what we want to assure.”
- The coming cyber-physical age is going to be exciting, but data management is key. Patel explained that everything from the Internet to collecting real-time data from systems as disparate as airplanes or offshore oil platforms requires a multi-disciplinary approach.
“We are operating at the intersection of data management, machine learning, and domain knowledge,” he said. “If you don’t have domain knowledge you can collect all the data, but you can’t get to causation. In this world, machine learning requires domain knowledge.”
- 3D printing is going to take more and more computing power. “As a whole, you have trillions of voxels we are printing,” he said. “And if each of them can be characterized in different ways, think of the computing problem. It’s a trillion-voxel problem. What does the compute engine look like?”
Patel explained that HP’s large-format printers send so many bits that the digital pipeline is very complex. “We used to use a rack-scale computer, but we went from that to a board-scale computer. We have ASIC designers, we have chip designers, who made a special-purpose computer so we can take this digital pipeline and print on 2D objects. Now you’re telling designers, instead of a pixel, you can worry about a voxel. You have to carry all those properties and add changing voxel by voxel by voxel. It’s a very complex compute engine. So, in the future, when we start to engineer at that level, we have to think about, what does the compute engine look like?” he said.
- Don’t forget about security. Patel said that we have to assure that a part being 3D printed is secure.
“It came out of our printer; does it have the right properties? How do we learn from what we have done and create secure, resilient entry of data—even with tomorrow’s much richer actuators and sensors? Today, I won’t say that this is all done, but this is how we have to think.”
- Think about the engineers of tomorrow. “When I advise a young person, I tell them, as I tell my children: get a degree in a silo, like mechanical engineering. Broaden out into computer science. Get dual degrees. Strengthen your fundamentals, couple them with computer science. The world of just having data science is over. The cyber-physical age is a return to fundamentals.”