HP Labs and HP’s 3D Printing Software Business Collaborate to Deliver Customer Value Through Technical Innovation

By Simon Firth, HP Labs Correspondent — May 22, 2019

Photo by HP

HP 3D Printing and Digital Manufacturing business unit’s Barcelona Software and Data team. Front row, from left to right, Annarosa Multari, Jordi Roca and Jordi Sanroma.

A research collaboration between HP Labs and the company’s 3D Printing & Digital Manufacturing Software business group promises to help HP’s 3D Print customers maximize the value they derive from 3D printing.

The collaboration draws on a multi-year effort within HP Labs’ 3D Lab to model the complex physical processes at the heart of HP’s industry-leading 3D Multi Jet Fusion(MJF) print process. Insights derived from this modeling project are now informing the development of HP 3D Process Control, a new HP software solution that aims to assist customers in optimizing 3D printing performance to meet the quality requirements of final part manufacturing.

“In building a virtual model of the 3D printing process, we’ve created a framework that lets us explore a variety of open research problems in 3D printing,” notes Jun Zeng, the principal investigator at HP Labs for 3D Printing software research. “Also, importantly, it’s providing the technological foundation for us to collaborate with our colleagues in HP’s business units to developing higher-performing solutions for our customers,” he says.

“HP Labs’ deep expertise in modeling and optimization is a unique asset for the company,” adds Annarosa Multari, Lead for HP’s 3D Printing Software Center in Barcelona, Spain. “My team adds deep understanding of customer needs, and together we are showing that we can really accelerate the journey from idea to innovative, customer-focused, technology-driven product,” she says.

HP’s MJF 3D printing technology works by placing small quantities of a fusing agent very precisely onto a thin layer of preheated plastic powder. As heating lamps pass over the powder, the jetted agent captures and distributes the energy to fuse the powder onto successive layers of the final printed part. At the same time, the printer manipulates temperatures at the print bed’s edge, helping give the final part the exact shape and material properties that it was designed to have.

The HP Labs investigation has been to recreate this entire process in software, which requires both a deep understanding of the physical processes of MJF and the ability to draw inferences from very large sets of data registered by sensors located in the printers.

“In building a virtual model of the 3D printing process, we’ve created a framework that lets us explore a variety of open research problems in 3D printing. Also, importantly, it’s providing the technological foundation for us to collaborate with our colleagues in HP’s business units to developing higher-performing solutions for our customers.”

Jun Zeng, principal investigator, 3D Printing software research

Photo by HP

HP Labs’ 3D Printing Software Research team. From left to right, Israel Figueroa, Fabiola Leyva, Juan Catana,

“We are able to record temperature data, for example, for every individual voxel (volumetric pixel) when it is exposed to light for fusing. We can also record the heating activity of the print unit itself,” Zeng reports. “That kind of information, when added to the process physics insights we are developing, is letting us reproduce the complex mechanical processes involved with increased accuracy, which in turn is allowing us to run ever better simulations predicting how future print runs will go.”

Zeng assembled a team including deep learning specialists and computational physicists to turn these advances into technical assets. While the predictive power they developed proved invaluable in terms of advancing the HP Labs researchers’ fundamental understanding of the Multi Jet Fusion process, Zeng, Multari, and their colleagues quickly realized it could also be deployed through HP 3D Process Control to help HP’s 3D print customers achieve unrivaled end-part quality and manufacturing efficiency.

One avenue that the resulting joint research project is pursuing focuses on “fit,” which aims to optimize printing so that a finished printed part matches the precise geometric specifications of the part’s original design. A new technique explored by Juan Catana from HP Labs and Jordi Roca from the Barcelona team deploys a very large cloud of coordinate data points across the part’s entire geometry. It applies deep learning to analyze those points to model and then predict “parts as manufactured” based on “parts as designed.”

Photo by HP

HP Labs’ 3D Printing Software Research team. From left to right, Sunil Kothari, He Luan, Jun Zeng.

This predictive capability can also be deployed to help ensure that printed parts have the specific strengths that they are designed to possess. Objects’ properties, such as their mechanical strength and toughness, can be affected by how they are heated and cooled during the printing process. In a second collaboration with the HP 3D Process Control team, the HP Labs group is refining a predictive model they have built for the entire 3D print process, which encompasses some 50 hours for each batch of printed parts.

“By mining data from the thermal sensors embedded in our printers and feeding it into a software model guided by the laws of thermodynamics, we can predict what we call the “thermal journey” of any specific voxel over the entirety of any potential print run,” Zeng explains. “The predictive scope of deep learning or artificial intelligence is fairly limited because we can only sense a very small fraction of the entire thermal journey. But when we combine artificial intelligence with the physical laws of thermodynamics, we are able to build a simulation engine that not only provides comprehensive thermal coverage in both space and time, but also continuously learns and adapts in real time.”

The new Labs’ technology can simulate the heating and cooling journey that newly designed 3D objects will undergo before they are ever printed. It can already address batches with more complex geometries than current commercial tools can handle.

“Process engineers can potentially use these simulations to adjust the thermal process as needed to ensure that any finished part exhibits the strength properties it was designed to have,” notes Jordi Sanroma, HP 3D Process Control Chief Engineer and a principal collaborator with the HP Labs team.

Some of these thermal modeling capabilities will likely be integrated into HP 3D Process Control, where it promises to improve end part yield and the manufacturing efficiency of any single print run. Zeng attributes the early success of this thermal modeling research to his talented colleagues. “We were purposeful in bringing deep learning specialist He Luan and computational physicists Carlos Lopez and Fabiola Leyva into the team; this research requires us to bridge the data / theory divide.”

“We were excited to have our HP Labs colleagues come to us with these ideas that grew out of their fundamental research,” says Ryan Palmer, Global Head of Software and Data for HP’s 3D Printing business group. “It’s a great example of how our 3D printing teams are working proactively together to achieve the innovation velocity we need in order to revolutionize digital manufacturing.”

“The relationship is just as valuable to HP Labs,” emphasizes Lihua Zhao, HP Distinguished Technologist and Head of the 3D Lab at HP Labs. “We feel very lucky to have such a great partnership with HP’s 3D Printing business group,” she says. “They not only have a strong, collaborative R&D team, but their business insights help us maintain a clear focus on projects that promise to make a powerful difference for our customers.”

Tags: Innovation / HP Labs / Science
Posted In: HP Labs Blog

HP Labs Blog

HP Labs

HP Labs and HP’s 3D Printing Software Business Collaborate to Deliver Customer Value Through Technical Innovation

Recommended Stories

HP Labs spatial audio research adds powerful presence to VR and other environments

Building an engineering curriculum for the era of digital manufacturing

HP Labs officially opens its HP-NTU Digital Manufacturing Lab in Singapore

Select Your Country/Region and Language

HP Worldwide

Select Your Country/Region and Language