Advances in metrology are key in reducing expensive, wasteful scrap
According to Marco Torsello, a product manager at Faro, Lake Mary, Fla., some aerospace companies produce so much expensive scrap, you can see the piles from space. Naturally the incentive to reduce this waste is enormous, Torsello added, noting that recent advances in metrology play a big role in achieving that goal.
Let’s start with laser trackers, an essential tool in aerospace manufacturing given their ability to accurately measure distances over a wide area. As explained by Joel Martin, North American director of product sales management at Hexagon’s Manufacturing Intelligence division in Novi, Mich., these units send a laser beam to a reflective target—usually a spherically mounted retro reflector (SMR)—held against the part being measured. They measure two angles and the distance to establish the measurement point in space. But none of this is new. Such systems have been commercially available for more than 30 years. What is new is the degree to which these systems are now easy to use.
Older laser trackers required a constant line of sight between the source of the beam and the SMR. Any break, and the operator had to start over. Likewise, earlier systems were physically cumbersome. Hexagon and others have created trackers that address both frustrations. For example, Martin pointed to Hexagon’s new AT500 as “ultra-portable,” thanks to “the controller and exchangeable batteries” built into the unit itself. “And we’ve increased the PowerLock capability, which is our technology that allows the laser tracker to lock back onto the reflector automatically,” he continued.
This allows the operator to concentrate on placing the SMR. In other words, the operator can focus on what needs to be measured whereas, previously, he would typically spend half his time worrying about the laser tracker and whether or not it was actually following and locked to the SMR that he was using to measure, according to Martin.
Another notable feature of the AT500 is its integration with a tactile probe, which Hexagon calls the B-Probe+. Being able to combine optical distance and positional measurements with probe data is critical, because it’s common for part features to be hidden from sight, or otherwise unsuited to laser tracking (or optical scanning).
Like laser tracking itself, the integration of probing capability is not new. But Martin stated that the AT500 offered the function at a significantly lower price point than what had been available. “Spending $100,000 when you’re solving million-dollar problems” might be a “no-brainer” for the OEMs, Martin observed. But things are much more price competitive for the smaller shops making tooling and fixturing, and that’s the market the AT500 addresses.
Torsello added that probing is often necessary to achieve the required accuracy when measuring critical aerospace features, especially when compared to the surface measurements achieved by laser scanning. Such a scanner projects a laser line onto a surface and then interprets the distortions in that line to create a 3D point cloud that mimics the surface.
These renderings can be quite good, but in aerospace the QC threshold is very tight 90 percent of the time, Torsello noted. As a result, probing is used about the same amount of time, he said, because optical measurements introduce uncertainties over large distances.
“Even the air between the sensor and what you’re scanning plays a role,” Torsello explained. “And you don’t have a completely controlled environment.” An aircraft hanger would help, he agreed, but “you don’t have very sophisticated air conditioning. It’s not a white room.” So the best systems, like those from Hexagon and Faro, make it possible to seamlessly switch between the three measuring techniques as needed.
Laser tracking, laser scanning, and probing can be combined to aid assembly, improve QC, and (perhaps more surprisingly) lower tooling costs, and speed machining. For example, Martin said the AT500 has proven itself in rocket assembly “in a dirt field.”
Imagine trying to perfectly align 30’ diameter, 60’ long [9.14 x 18.29 m] barrel segments. “You need to be able to understand where each barrel segment is in space. And you need to be able to then identify their positions as you drive the two segments together before you start welding them.” Beyond that, the AT500 measures any weld distortion that occurs, and helps the team determine if adjustments need to be made to ensure the barrel segment is perfectly straight, linear, and where it’s supposed to be, added Martin. “Because at a certain point, these things become so big that you don’t want to micro move them after you’ve constructed them.”
The AT500 is rated IP54 for protection against dust and liquid contaminants, and Martin boasted that they have customers using them “literally in the rain, in the sun, doing detailed precision alignment work in conditions that we would never think you would want to be in.”
Faro offers the unique combination of a laser tracker working in conjunction with a laser projector. Thanks in part to a gimbal system, the latter projects nominal weld positions, drill locations, and so forth, in 3D onto the target part.
As Torsello put it, this kind of sophisticated guidance contributes to a zero-defect manufacturing process. At a minimum, it drastically reduces the level of rework that is often required on certain parts, because without such a system, assembly personnel follow handheld CAD drawings as they work along a fuselage or other large components.
“It’s very easy to miss something, and it’s very easy to misplace something,” Torsello acknowledged. But Faro’s latest projectors can triangulate with the camera systems to understand if the required item (e.g., a bolt), is actually there, and warn the user if it’s missing.
“Depending on the application and the required accuracy, these projection systems can be used on their own,” Torsello said. “On the other hand, they can also be used with traditional metrology devices such as a laser tracker,” he continued. “The laser tracker would be used to inspect certain features on the fuselage or any other assembly or tooling, and the data coming from the laser tracker would be interpreted by the projector in order to project even more accurately on the part. So the combination of the two further reduces the room for assembly error.”
Measuring wing stringers that are up to 120’ long [36.58 m], while only a few inches wide, has typically required a 200’ [60.96 m] gantry style CMM, explained Martin. Plus you needed a fixture as long as the part, with integrated fiducials. “The fixtures become incredibly expensive, they’re difficult to align, and they’re difficult to keep aligned. Because the fixtures are so big that unless you’re in a temperature-controlled area, the fixtures are constantly moving, and they’re never in a temperature-controlled area.”
Hexagon’s solution is to put an AS1 laser scanner on the end of a robot (or a cobot), and mount the robot on a rail running along the stringer. It uses a laser tracker to understand where the AS1 is in space—eliminating any reliance on the robot’s own positioning feedback—and the AS1 scans the stringer. This effectively creates a “very large volume inspection device for automated measurements that meet the specific requirements of the part,” concluded Martin.
The published accuracy of the AS1 is 50 µm over its full measurement range, which is well within the tolerance for these parts. What’s more, Martin added, Hexagon’s Robotic Automation software suite “allows us to pull in a CAD model of the part that we’re going to measure, pull in the inspection program of what it is we’re measuring, and use the inspection program to build the robot program automatically.”
This approach can also be used to greatly simplify tooling for machining applications, offered Torsello. “In the last few years, manufacturers have wanted to use more flexible machining centers, to reduce the need for specific tooling for each part, or the need to setup a specific machining center for each part.”
Take Cirrus Aircraft, in Duluth, Minn., which uses a robotic machining center to drill holes and mill surfaces on large airframe components, but switching between parts required expensive retooling and long delays. With a laser tracker, the company is able to quickly recalibrate the robot for each new part, according to Torsello.
“The robot can interpret the data from the laser tracker. And we also offer the software solution that drives the tool,” he said. “So in a matter of only a day of configuration, the part can be changed and the same machining center can be used to manufacture something completely different from the day before.
This has cut tooling costs by 60 percent. In fact, because of the flexibility offered by the laser tracker’s ability to drive the robot, the tooling cost was removed for most parts.”
A laser scanner/tracker combo can also be used to lower the cost to produce tooling, as evidenced by Janicki Industries, in Sedro-Woolley, Wash. Among other things, Janicki, makes tools for forming large aerospace sheet metal components. These molds start as oversized castings, which then need to be machined to their desired geometry. But inconsistencies in the amount of “over material” makes it impossible to create a milling program that avoids “cutting air” over large areas that don’t actually have such material. The solution is to use a Faro laser tracker to orient the part position relative to known references, and then orient a Faro laser scanner—mounted on a robot arm—to the same references, “so you know exactly where the arm is in space,” Torsello explained.
“You would scan the surface of the forming tool and load that data into your CAM system. The CAM system then knows exactly where there is excess material in each machining path.” The machining program then proceeds, going down in Z as far as needed to remove excess material with each pass.
The result at Janicki has been a 35 percent reduction in the machining time needed for such parts.
In addition to airframe parts and tooling, there’s also the problem of measuring the toughest part of the plane: the turbine hot zone. As Blake Spaman, senior application engineer for aerospace projects at GOM Metrology, a Zeiss company based in Braunschweig, Germany, explained, the traditionally accepted method of measuring turbine blades relies on a CMM with a touch probe.
“They do a contact scan in which the tactile probe drags across the surface and creates a section,” Spaman said. “Then they repeat that 10 to 15 times at different heights on the blade, yielding a number of sections for each blade.” For a full blisk, this process can take about eight hours.
GOM and its U.S. partner, Capture 3D, another Zeiss company based in Santa Ana, Calif., offer a better, and much faster automated solution: the ATOS 5 Airfoil ScanBox. Unlike a laser scanner that projects laser lines across the surface of the part, the ATOS 5 Airfoil uses an LED to project a very fine fringe pattern. As the patterns rapidly shift during each scan, two cameras (working in stereo) observe the part surface and triangulate the points in space. Spaman referred to the interaction of the cameras and central projection unit as achieving a “triple scan—three measurements in one shot,” in contrast to a laser scanner that uses one light source and one camera to receive the surface data via reflection. With ATOS, the end result is a more accurate scan of the entire blade surface in 3D, (with the exception of the turbine blade root form, which is a different measuring task with different tolerances).
What’s more, added Johan Gout, senior director of operations at Capture 3D, the ATOS 5 Airfoil sensor (paired with GOM Software) is a solution with many of the industry’s requirements “baked in” to make it “intelligent.” For example, there are specific angles with which to approach the leading or trailing edges of any particular blade in order to collect the “best, most reliable, full field data,” according to Spaman. “The GOM Inspect VMR [virtual measuring room] software automatically recognizes an airfoil’s geometry, and calculates the optimal positioning of the robotic sensor to digitize the blisk.”
A wise user would also model the fixtures, of course, and likewise the VMR software would create a measuring sequence that avoided any collisions.
Gout added that the software automatically adjusts the density of the measuring points to provide more detail of any contours. “You’ll see all the polygons that make up the surface. And where there’s a radius, you’ll notice a higher density. The software uses intelligent chordal-based sampling to give you a higher definition of polygons on the leading and trailing edges than on the flatter surfaces, while bigger polygons still represent the surface accurately.” Spaman said this also serves to limit the file sizes as much as possible, while still being true to the part.
All this technology adds up to producing a 3D scan of a blisk that’s accurate to 25-35 µm in only 20 to 25 minutes, versus eight hours on a traditional CMM. In certain configurations, the technology can achieve accuracies of 3-4 µm. As mentioned previously, ATOS produces a full as-built digital twin of the blade, not just 10-15 segments as with the traditional method. In fact, it’s so good, Gout enthused, that the technology can be used to reverse engineer a product, or record a part’s digital twin for Industry 4.0 considerations.
And while it has taken numerous correlation studies against CMM measurements, Gout reported that “the good majority” of OEM jet engine manufacturers now accept ATOS measurements for part validation. “The OEMs feel confident they can use the ATOS solution and get more information in their measurement, especially on the first article inspection side.”
Applications go beyond quality control. Spaman recounted a case in which an engine manufacturer needed to dial in its robotic coating process for airfoils and nozzles. “Many of the single crystal alloy coatings used on these components have magnetic properties that make it impossible to use eddy current analysis to measure thickness. So, thickness measurements typically require a destructive test,” he continued. “Or you have to weld a number of coupons onto a part, and then you get localized data on a very select number of areas. You end up sending parts to a lab where they evaluate the cross sectional thickness of the coatings.”
Each iteration took six weeks. But ATOS technology made it easy, and the customer was able to “have a blade stripped, coated, scanned, and then evaluated for thickness in about an hour and a half. They once did seven iterations in one day.”
Measuring creep fatigue is another important use case. “Typically, the life expectancy of aerospace investment castings is based on the geometry, the metallurgy, and the heat in the engine,” Spaman noted. But in many cases, no one actually tracks the condition of the part during an engine interval. They simply switch out parts according to the expected life.
One company Spaman worked with set an engine interval of two years for the hot gas path turbine components (stage 1-4 blades and nozzles). “It took a long time to collect all the data. But we measured all these high-value components when they were made, and then after each engine interval to figure out the change,” Spaman said. “We were then able to create a very accurate model of creep fatigue. And we realized that components we thought might last two intervals, could actually last five intervals. That ended up saving the company $950 million over the cost of the fleet.”
That huge success led “everyone to start taking the technology seriously, and to start asking ‘If we’re already scanning these for creep fatigue, are these scans good enough to eliminate some of the other inspections that we’re doing?’ And the answer was ‘Yes!’”
Connect With Us
