Cloning Metal Parts for Space and Earth

Subheadline

Software for modeling metal parts before 3D printing speeds development, cuts costs for NASA and business

Compared to cloning a sheep, copying a metal part should be simple. It’s not, though, especially if that part is to be built by a 3D printer.

Additive manufacturing – or 3D printing – can save weight, time, and money and often produce sturdier parts than otherwise possible. For these reasons, NASA is interested in using the technology to make rover components, ion engines, and other hardware, but the agency needs to know the parts will be reliable. One way to prove them is with software that creates a digital version, or “clone,” of the part to be printed.

It’s a complex task. Machines for 3D printing with metal can work in a number of different ways, and factors like power level, laser intensity, and alloy properties affect the final product, often in ways that are hard to predict. So NASA invested in a company that already specialized in creating digital versions of gears and other components. Now DigitalClone for Additive Manufacturing (DC-AM) can perform virtual testing of parts at multiple stages of the process from initial print to final part quality before a single piece of hardware is printed.

NASA’s Jet Propulsion Laboratory in Southern California worked with Sentient Science Corp. of Buffalo, New York, to build a modeling program with the help of Small Business Technology Transfer contracts.

Physics in Real Life

“When you’re dealing with expensive alloys or complex shapes, you’d like to have a deeper understanding of what will be happening to the parts during the building process,” said Bryan McEnerney, materials technologist at JPL. Computational material science can minimize the risks connected to printed parts. That could be financial risk, schedule risk, and technical risk, according to McEnerney.

Conventional manufacturing can be time-consuming and expensive and add more mass to a system, but the parts are proven reliable. That’s essential when lives of astronauts and billions of dollars of investment are at stake. Additively manufactured parts are still too new to have such a track record of success, making the use of a digital clone important.

“If we model parts, it will save a lot of time and maybe even predict where we might have issues as we’re designing them,” explained McEnerney. Modeling brings together the part, its use in real life, and the impact of the conditions it will experience over time.

DC-AM includes three modules for designing and testing a part, which rely on the principles of physics to dependably represent what takes place in any 3D printer. Process modeling takes into account printer parameters, such as power, layer thickness, and laser beam dimension, and their effects on the metal. This contributes to residual stress and distortion predictions for a part. Microstructure modeling predicts the grain structure and porosity necessary to achieve the desired properties. Finally, fatigue modeling forecasts the potential damage caused by extended use, such as the onset of cracks, that a part will experience in a specific application.

This approach, called integrated computational materials engineering, yields high-fidelity data that’s particularly important for NASA when fabricating unconventional parts, according to McEnerney. “The entire agency often has challenges with certain high-complexity parts, because a lot of what we do is hard.”

Trial and Error

Makers of additive-manufacturing machines provide some recommended print settings for specific types of metal powder feedstock, according to Jason Rios, senior vice president with Sentient Science. But “there’s a lot of uncertainty” when creating a new part, he said. “You have to try to search for that right combination of print parameters to get the part quality that you’re looking for.”

Before DC-AM modeling, that meant creating and tracking different settings for various materials before subjecting each part to physical tests. That’s an expensive trial-and-error approach. It’s been worth the effort, given the opportunity to create more efficient components by using new geometries, but it’s not sustainable.

“What’s really unique about DigitalClone Additive Manufacturing is the ability to model the process, predict the microstructure, and predict how long the part’s going to last in its intended application based on the stresses and the loads it’s going to see,” said Rios. While some of these functions are possible with other programs, this combination is the first of its kind.

Thanks to the Amazon Web Services high-performance computing that runs DC-AM, anyone can develop and optimize any part, building and testing it in a virtual environment. When it comes to replacing an existing part design, the program can quickly determine whether the new part is going to meet the design specifications and the intended lifespan for a particular purpose, explained Melissa McReynolds, vice president
of automotive operations with Sentient Science.

The company has a number of automobile manufacturers putting the program through its paces to identify how 3D-printed parts might fit into existing and future car designs. The program can be used in any industry that relies on metal parts, and the Japanese company Matsuura Machinery is using DC-AM to provide quality control for its 3D-part-printing arm. And the company is working with Sentient Science to include the program in a new 3D printer that will monitor parts while they’re being created. If a flaw is detected that will result in a defect, the system will automatically cease production, remove the damaged layer, and resume the process.

PhD-Level Experience

This new modeling program builds on the success of the first DigitalClone software, also built with NASA’s help, which predicts the lifespan of machine components (Spinoff 2016). New versions, in addition to DC-AM, support specific industry needs: DC-E for engineering, DC-OM for wind energy, and DC-RM for rail.

Businesses that 3D print parts need to accurately predict how long a part will function so they can create maintenance and replacement schedules without compromising safety. For them and their customers, the expertise NASA contributed to the program adds another level of quality assurance.

“Through that collaboration with JPL, we were able to include their knowledge in creating a product that allows a customer to do that same type of analysis without having to acquire the PhD-level expertise our team has,” said Rios.

The agency benefits from the significant cost and time reduction compared to conventional physical testing for new component designs, materials, and processes. McEnerney noted that the partnership with Sentient Science has created a new resource for “the greater engineering community.”

“We’re looking to continually evolve and hear new voices, to come up with new technologies,” said McEnerney. “By leveraging small businesses, we’re going to get a better outcome for the agency and for the country.”