Building the future with robotic additive manufacturing
Additive manufacturing (AM) is changing how engineers and part designers think and robots are enabling the technology by making AM machines and processing faster and more accurate than they would be on their own.
Additive manufacturing (AM) is not only transforming the way we make things; it’s changing how engineers and part designers think. They have to forget limitations imposed by conventional manufacturing methods and open their eyes to new design possibilities. These possibilities are expected to catapult the AM industry to $17 billion by 2020. There are several types of AM processes, including selective laser sintering (SLS), stereolithography (SLA), and fused deposition modeling (FDM).
All are digital manufacturing methods where computer-aided design (CAD) data is used to fabricate a 3-D object by adding layer upon layer of material, whether it’s liquid, powder or sheet, or some other type of material. Even human tissue can be used. AM is used to create myriad structures, from dental appliances, to advanced aircraft components, an entire bridge, and even works of art.
Robots help make it possible. Robots are not only enabling additive manufacturing, they’re tending 3-D printing machines (which are also robotic), automating AM post-processing, and allowing architects to envision new ways to build.
Layer by layer
At Midwest Engineered Systems Inc. (MWES) in Waukesha, Wisc., they are using laser AM to create complicated metal parts that would otherwise be extremely difficult, if not impossible, to manufacture. A six-axis articulated robot drives the process, combining hot wire deposition and a laser to build metal parts layer by layer on an existing substrate. Exotic metals are deposited with precision and speed to build prototypes and small batches of high-value complex parts.
MWES brought its 25-plus years of expertise in complex systems integration to bear in developing this process, which was unveiled at the International Manufacturing Technology Show 2016. In a show floor demo, a propeller took shape during the layer-by-layer process.
MWES named their system ADDere, which was derived from the Latin word meaning to add. The process is similar to wire and laser additive manufacturing (WLAM), where a metal wire is fed into a melt pool generated by the laser beam on the substrate. The wire and substrate consequently form a metallurgical bond. The difference is that MWES uses a hot wire process.
"We heat the wire to the point that it’s molten at the tip," said Scott Woida, president and founder of MWES. "Since the wire is already molten, we then use the right amount of laser power to melt the substrate underneath to form a strong bond. You’re able to use less laser power when you’re not trying to melt the wire as well as the substrate. The hot wire allows you to get higher deposition and put less heat into the part."
The process always starts with a substrate. In the show floor demo with the propeller, it was a cylinder.
"We can either use the substrate as part of the final part, or we can cut the substrate away and just have the part made of weld bead," Woida said. "But we have to start with something. It can be as simple as an eighth of an inch thick piece of steel."
Wire and laser, plus robot
The primary elements of the system include a high-precision industrial robot, the laser system, an integrated MIG wire and laser head, and the MWES controls system. The process includes active head control and dynamic deposition measuring to closely monitor the process before, during, and after the build.
To begin, CAD data is imported into CAD/CAM software, where it is prepared for the additive process. The part is then "sliced" into layers and the robot path is generated offline. Process information can be added automatically or manipulated manually. The generated path and process information is translated through a post processor and automatically transferred to the robot controller. Then the robot executes the program and builds the part layer by layer.
- Small batch production runs
- Replacement parts
- Rebuilt surfaces
The ADDere system uses a six-axis long-reach robot, which provides for path flexibility and a large working envelope. It’s merged with a multi-axis part positioner. Woida said 2 x 8 x 40 m working ranges are possible. Achievable tolerances are +/- 0.5 to +/- 1.5 mm, depending on deposition rate. Post-processing usually requires some machining. The additive process creates a hardened form of the material, so soft metal also requires annealing.
Freeform fabrication leads to less waste
System advantages include rapid development of new metal parts, quick design changes without adding tooling costs, and low initial cost to production. Woida said one of the main advantages is the ability to take multiple part subassemblies and combine them as one unit.
As an example, GE Aviation took this concept to a whole new level with the AM process for its Advanced Turboprop engine. GE designers were able to reduce 855 separate parts down to just 12. More than a third of the engine is 3-D printed.
With MWES’ ADDere system, solid freeform fabrication allows the use of different metals on different areas of the part to create engineered characteristics specific to an application. This is particularly cost-effective when you want to clad a less expensive metal with a more exotic metal for particular properties like high wear resistance. The process also can be used for repairs by first machining a part to a stable structure and then building up the part to its original state.
"We’re getting properties similar to casting, closer to forged," Woida said. "Compared to subtractive methods, you waste less base material because you’re building to near net shape."
Wire AM also results in less waste than powder-based AM processes. Woida said they achieve 99% utilization. "When we’re running the wire for manufacturing our component, all that wire ends up getting used to make that part," he said. "There is very little waste of the wire (as opposed to powdered metal AM processes where the excess powder falls by the wayside and needs to be recycled). The only thing that happens is that you’re machining the outside of that component to get from your near net shape to your net shape. Typically you only machine your mating surfaces. You don’t have to machine the whole part."
For the propeller in the demo cell, Woida said you may only need to machine about 5% of that part after the AM process. He said it’s also 10 times the speed of powder-based AM processes.
"We can put down 32 lbs. per hour of stainless steel right now, and that’s with a 14 kW laser. Soon we’ll have a 20 kW laser. When the material has a high dollar value and it’s really hard to machine, this process makes sense," he said.
Not suited for this process are small components, parts that have low manufacturing costs, and parts that require little machining from billet.
"When the part is done, it has a casting-like quality to it," Woida said. "You can either machine the part or we can use a laser to smooth out the outside for a better surface finish. But a lot of our customers are less interested in surface finish as much as they are functionality."
High-value parts, exotic metals
The ADDere system is available as a turnkey product for purchase or as a manufacturing service.
"The parts we are working on to date are basically validation for customers that we can make the components to their specifications," Woida said. "Mass production hasn’t started, but we are providing sample sets to customers to verify the capability of the system. They are evaluating them for quality and then they will be buying them in larger quantities from us, or buying the system."
One of those parts undergoing testing in MWES’ R&D system is an 1,800-lb. bulkhead for an aircraft carrier. Rather than having to waste valuable space with spare parts inventory aboard the ship, imagine being able to use AM to create or repair parts, on demand, while at sea.
The ADDere AM system has application for aerospace, drive train, suspension, naval, military, oil and gas, construction, mining, and agricultural equipment. Materials best suited for these applications are typically exotic metals, such as stainless steel, aluminum, titanium, cobalt, Inconel, and tungsten alloys. Woida said their experience in laser welding is paying off.
"We typically get involved in highly engineered systems, so we have a lot of exposure to the latest technologies, whether it’s the latest laser technology or robotic technology," he said. "On a daily basis, we design systems that don’t exist in a catalog, that are highly engineered. You need a lot of diverse experience. You need mechanical engineers because these systems are fairly complex. You need software people to make this easy and viable to sell on the open market. You need robotic engineers to then integrate all that. You need weld engineers that can verify and make sure the metallurgical properties are what they’re supposed to be. You need a whole lot of people to bring this together."
AM and robotic automation also are ushering in a new digital world for the metal casting industry. 3-D printing was born in the 1990’s at the Massachusetts Institute of Technology with student Jim Bredt, who was working on his doctoral thesis investigating the creation of inkjet on powder technology.
The term 3-D printing originally was used to describe how intermittent layers of powdered materials and liquid binder are dispensed in a programmed pattern to form a 3-D object. Bredt said the term was coined by his thesis advisor, and eventually adopted by industry at large. Now, 3-D printing often is used interchangeably with AM and encompasses many different types of processes.
Bredt, research and development director at Viridis3D in Woburn, Mass., has nearly 30 years in the 3-D printing industry, but he never lost sight of his first love. One of his specialties is ceramics, particularly mold making for metal casting. He has been interested in metal casting since he was a child. That may have helped him get into MIT. The university was intrigued with Bredt after discovering he built a foundry in his backyard while he was in high school.
Bredt helped start Z Corporation in 1995 after graduating from MIT. Z Corp. is credited for the first commercial introduction of inkjet-based 3-D printing technology. Z Corp. was then acquired by 3-D Systems, whose cofounders invented stereolithography.
Bredt left 3-D Systems to launch Viridis3D in 2010, where he was eager to get back to metal casting. The goal was to build a 3-D printing machine that was more versatile in the types of materials it could process and more hardened industrially to handle the rigors of working in a foundry. The Viridis3D team focused its sights on the sand casting industry.
"A lot of the components that you really need a 3-D printer for are things you can’t make by conventional processes, like cores especially, which can be very intricate," Bredt said. "It increases the capabilities of your process in such a way that you can take more risks in your design. You don’t have to spend all that on tooling. If you 3-D print a part and it’s a failure because the design is too fragile, then you didn’t really lose that much. By being able to take more risks in your design, it expands the gamut of geometric shapes that you can create with the technology."
Breaking the mold
Among reimagining ways of making things, Bredt questioned how 3-D printing machines were designed. "A 3-D printer is basically a robot with a material dispenser attached to it," Bredt said. "When we created Viridis3D, I asked why should I try to build my own robot. My background is in materials, not in machine design. Why don’t I just buy a robot? Then I can focus on material dispensing."
For Viridis3D, using an off-the-shelf industrial robot in its 3-D printing system was a major break from the competition.
"Our use of a commercial robot is an interesting distinction," Bredt said. "Our competitors by and large use gantry systems to lug their much heavier printing engine around. By using an arm instead of a gantry system, our print head is designed to be light, rugged, and reliable."
He noted no motion control system comes without its complications.
"They each pose their own challenges. Having worked with gantry systems through my entire career before Viridis3D, I’m actually very happy with robot arms. Our cost is low because the robot is very economical. The accuracy and the load limit is very generous relative to what we need for our system."
3-D printing with robots
While development was underway on its robotic AM system, Viridis3D partnered with EnvisionTEC, a provider of 3-D printing solutions. Now a wholly owned subsidiary, Viridis3D can continue to fund further development. Earlier this year, Viridis3D commercialized the first robotic 3-D printer, the RAM 123.
The system uses a standard four-axis robot to create sand molds and cores for casting metal parts. The robot is equipped with a powdered material feeder that distributes the sand and a print head that dispenses the liquid binder into the sand. Spreading sand and dispensing binder intermittently, the robotic 3-D printer builds the mold layer by layer.
The print head can be heavy, especially when loaded with sand. Bredt said a four-axis robot, rather than a six-axis robot, is preferred because it has a larger load capacity.
"These printing elements really only work if they are held horizontally," he said. "The print head travels in a plane and very gradually rises up. Four-axis is ideal because they are constrained to always travel in a plane, at least the wrist. It has a high-load capacity and stays accurate."
Viridis3D’s RAM system has an open architecture. The molds are built on a stationary table. The tabletop is a pallet that can be used to move parts on and off the machine with a forklift.
"Ours is an open table, so you can build parts of different sizes without having to fill the entire box with materials," Bredt said. "One of the reasons we went to a stationary table is that it was clear to me during the later years at Z Corp., when we were building larger and larger machines, that pretty soon the substrate would weigh more than the machine. Some of the heavy foundry sand we use, if you fill up the box, it weighs more than the machine."
How it works
The process begins with the designer creating a CAD model. Once uploaded, it is opened with the desktop software, which waits for the print head to request data for each layer of the mold in sequence. Then it queues up a few layers in advance and sends data to the robot on a layer-by-layer basis.
"Our desktop software is similar to your print manager program that sends each page to a printer," Bredt said. "Except this is in 3-D and there’s a lot of algebra involved."
Viridis3D’s proprietary desktop software is responsible for taking the data created on the CAD system and converting it into machine code used by the robot and the print head to actually build the mold. After the mold is built, you may need to wait a little while for the chemical reaction to cause it to solidify. It’s then removed from the other sand piled on the table, brushed off, and taken to the foundry.
"You can pour molten metal right on top of it and it works just like a conventional mold," Bredt said. "The mold itself cannot be re-used. The sand material can be heat treated to clean it up (to separate it from the binder) and then used again. The sand left on the table that doesn’t get made into the mold can just be scooped up and fed back into the machine."
Short lead times, space savings
For users of the RAM system, the main advantage is they can quote parts with a short lead time. Molds and cores with complicated shapes can be made without a lot of tooling. Going digital is also easier.
"Sand casting is still dominated by the old wooden pattern technology, where you take a box and you squish sand up against a wooden pattern to make a cavity of a certain shape," Bredt said. "Then you assemble it into a mold to make a casting. It’s a relatively quick process for making a mold. In fact, it’s quicker than a 3-D printer, but you have to have a warehouse somewhere where you keep the patterns."
Patterns become damaged or misplaced. The cost of warehouse space to inventory all those patterns continues to rise.
"Pattern-making is a dying art," Bredt said. "Companies with these 50-year-old patterns send in a guy with a can of Bondo to try and fix it. In some cases, all they have is drawings or maybe they have to reverse engineer existing product. People that buy our system really like the option of switching over to digital manufacturing because you get rid of the overhead for storing those patterns. This is a perfect example of disruptive technology taking over old technology."
Bredt said they are widening their sights on other materials beyond sand that offer higher resolution, including plastic powders, ceramics, and even powdered metals. Robots will continue to shoulder the load. Together, AM and robotics will transform the way we think about manufacturing.
Tanya M. Anandan is contributing editor for the Robotic Industries Association (RIA) and Robotics Online. RIA is a not-for-profit trade association dedicated to improving the regional, national, and global competitiveness of the North American manufacturing and service sectors through robotics and related automation. This article originally appeared on the RIA website. The RIA is a part of the Association for Advancing Automation (A3), a CFE Media content partner. Edited by Chris Vavra, production editor, Control Engineering, CFE Media, email@example.com.
www.controleng.com keywords: robotics, additive manufacturing
Additive manufacturing is transforming the way things are made and has opened up new design possibilities for engineers and part designers.
Robots help additive manufacturing become more precise and efficient and allow it to perform more sensitive tasks.
Additive manufacturing and robotic automation are changing the metal casting industry by making it more digital and precise.
In what other industries could robotics and additive manufacturing be useful?
See related articles from the RIA linked below.