Generic robots for shop maintenance
Maintenance shops perform many inspections, sometimes very detailed and sometimes very repetitive. They require highly trained technicians with a sharp eye for defects because some defects are very small and difficult to detect. Royal NLR develops generic technological solutions to automate repetitive tasks, to inspect objects with a very high accuracy, to automatically detect defects, and to qualify these innovative inspection tools. This helps maintenance shops to improve the working conditions for their technicians and their productivity.
The challenge
Component and piece-part inspections are a large part of the shop maintenance activities. These inspections are sometimes detailed, repetitive and time-consuming. The first challenge is to develop technology that can perform these detailed, repetitive and time-consuming inspections more efficiently.
And, shops inspect many different parts, each with their own unique shape, material characteristics, inspection criteria, and so on. The second challenge is to develop technology that is suitable to inspect as many of these parts as possible, and find the part-specific defects for each of these parts.
There is not a single sensor to scan all parts and detect all thinkable defects. And from a productivity point-of-view, it makes no sense to hand-carry these sensors through the aircraft or to manually analyse the scans or images. The third challenge is, to robotise sensor systems that can perform these inspections.
Inspecting parts in a shop requires a generic (or universal) and autonomous robot that can inspect many different parts and detect a multitude of defects. The forth challenge is the use of machine vision technology and advanced defect detection algorithms.
Before these robots can be applied in a shop environment, it is necessary to prove that these robots are actually good enough to replace their human equivalents, the technicians. The Fifth challenge is to qualify the robot for maintenance work.
The solution
Royal NLR has worked on maintenance technology to perform autonomous robotic shop maintenance inspections (in parallel to our developments on autonomous aircraft inspections). NLR is developing autonomous robots for inspection purposes, comprising of:
- One or more sensors to scan the object
- A robot to move the sensors to the right locations around the object
- Advanced defect detection algorithms
- Automation technology to autonomously perform the prescribed inspections
- Digital Twins to store, access and report the sensor data and inspection results.
One of our test rigs is the GERDA, a Generic Engineering Robot for Defect Assessments. The working range of this particular test rig is suitable to inspect different parts, as long as they fit in an oversized shoebox. GERDA is for example suitable to inspect engine airfoils, blades and vanes. The test rig can be fitted with multiple sensors to inspect the outer shape of cog wheels, axes, shafts, brackets, bearing races, and so on. We use machine vision and advanced algorithms to detect defects, and to classify and measure the size of these defects. Our routines compare the actual damage sizes against the allowable damage limits.
What did we do
The test rig is part of our efforts to develop maintenance technology to improve productivity and to make the working environment for shop technicians more interesting. We focused on the automation of inspections. We reckoned that we needed sensors, robots and automation technology to inspect aircraft and components for damages. And most importantly, we needed to qualify the robot and prove that it is as good as highly trained and skilled technicians.
NLR studied these areas over the last years. We looked at the suitability of sensors, the capabilities of robots, autonomous systems and the integration of these building blocks into an autonomous inspection robot. In the process we learned a lot, especially about the complexity and the qualification of these inspection systems. It goes way beyond the installation of a camera at the tip of a robot arm.
It also means that there is a lot of work that needs to be done. The inspection of one part of an aircraft for one defect is relatively simple and it seems feasible to bring these systems to the market in a relatively short period of time. Now, we look at universal solutions with a large degree of autonomy.
GERDA is a joint effort with KLM Engineering & Maintenance and the BrightSky consortium.