One of GE Aviation’s lesser-known success stories, the GE Bromont Global Robotics and Automation R&D Center, is developing robotic technologies to automate internal and external inspections of engines arriving at GE maintenance shops.
Founded in 2011, the automation center of excellence emerged as an adjunct to GE Aviation’s compressor blade- and vane-manufacturing facility in Bromont in southeastern Québec. Opened in 1982, the facility is GE Aviation’s most automated production plant, and GE established the R&D center to find ways to automate the factory’s procedures to make production more efficient in terms of unit time and cost.
The world average ratio of robotic production systems to machine operators stands at about 1:100; in Singapore, where production is more automated than in any other country, the ratio now equates to about 1:10, said Eric Dube, director of GE Aviation’s Global Automation and Robotics R&D Center. But GE Bromont's production has become so automated that it boasts a robots-to-operators ratio of 1:2.
GE characterizes the R&D center’s mission as a success. Dube said that the Bromont blade-production facility has averaged a 6.4 percent reduction annually in unit production cost, but from 2018 to 2019 the cost reduction grew to more than 10 percent.
He attributes the improvement to GE Bromont's application of many of the automation solutions developed by the R&D facility—which works closely with the GE Research Center in Niskayuna, New York—for GE Aviation and other GE production subsidiaries. One such project involves automating the production of hybrid-electric engine systems, according to Steeves Bouchard, chief engineer for GE Bromont.
Bouchard told AIN that the automation center has now developed about 120 successful robotics and other automation systems installed at GE facilities globally and offered to external customers.
Most of the center’s initial R&D work involved robotics projects, according to Bouchard. Within the past five years, it has used that expertise to automate internal borescope inspections of engines inducted for maintenance before and after engine strip-down and rebuild. The R&D center is studying the use of automated ground vehicles (AGVs) and drones to perform external inspections, and AGVs to conduct internal borescope inspections.
If GE can automate inspections fully, it could create a big advantage for GE’s MRO shops, said Bouchard. “If you digitize your inspection, you can automate the repair," he explained. "You can get the part, find the defect, weld it, blend it, inspect it, and ship it [without any human involvement other than oversight.]"
Another of the center’s MRO-related solutions uses digital cameras to perform the visual portion of fluorescent penetrant inspections (FPIs) of engine metallic parts for signs of cracking. Mechanics require a lot of training to perform FPIs properly and, for some parts with convoluted topologies, operators need complex arrangements of tented mirrors to see the entire part, said Bouchard. Using a digital camera provides a better view.
Now the center has begun work on artificial intelligence-based tools to create a fully automated inspection system for non-destructive testing of engine parts. It aims not only to create the capability to detect and locate cracks but to do so with a better detection rate than possible with human inspection.
One automated inspection tool the center has already developed uses a 3D sensor to detect and measure the depth of visible defects in parts. Not only does the sensor inspect turbine and compressor blades, but it can also inspect any piece of engine hardware, said Bouchard. The tool measures defects more reliably than human inspectors and has reduced the rejection rate of inspected blades by more than 30 percent.
Dube expressed particular pride in the approval of GE Aviation’s investment review board of the center’s proposal to install a newly developed robotic system for handling large parts weighing 400 to 500 pounds at the company’s factory in Lynn, Massachusetts, he said.
The board determined the project provided “instantaneous payback” in costing less to install than traditional production machinery. The system has also lowered production costs by $2 to $3 million per year, according to the company.
Another development project involves an automated way—using robots with pressure sensors, controlled by artificial intelligence algorithms—to solve the “bin picking” challenge. Simple for humans but not for robots, picking highly reflective parts with complex surfaces from storage bins and placing them exactly where needed for the next production step has proved a challenge for GE Aviation.
The center planned to demonstrate the first such system at GE Bromont in early June. If that proves successful, the GE Bromont production facility could install 20 systems, said Dube.