Taking airframe related composite repairs to a new level, part 3: Fan Cowl Door repairs

In the first article of my four-part series, I provided a basic overview regarding the capabilities of our new scarfing and milling robot for Airframe Related Components (ARC). In the subsequent part 2, I went a little deeper into detail regarding the first part-specific repair process for radomes. Although widely comparable to radomes from a process perspective, I would like to use this third article to shed some more light on the benefits that the robotic approach brings for the repair of fan cowl doors in daily MRO operations here at the Lufthansa Technik base in Hamburg.

Fan cowl doors such as those of the widespread IAE V2500 turbofan depicted in the title image are prone to a number of typical damage types. During ground handling of a commercial aircraft, these ARC components frequently suffer various unintended mechanical loads, often resulting in delaminations in their outer skin and disbonds between the honeycomb core and the composite skins. These damages often occur to such an extent, that operators and MRO providers are left no other choice but to scrap the entire part due to the size or location of the damage for which no repair is defined in the manual. As a consequence, the resulting exchange and acquisition of a new part pose a heavy financial burden on the aircraft operator.

To spare our customers from the costly replacement of a fan cowl door, our engineering team at Lufthansa Technik already introduced a first repair solution back in 2013. However, this manual process did not entirely free us from the risk of scrapping a part. Back then, if the inner skin would have been damaged during the excessive required manual grinding, our repair solution could not have covered the damage and the part would have been written off as unsalvageable.

This is where our new robot system really lives up to its potential by offering precision and repeatability that is far superior to manual work. Even though the level of required precision is less than described for the radome process in part 2, the accuracy of the robot still completely removes the risk of damaging the fan cowl door beyond repair.

Just like in the radome repair process, the fan cowl door is mounted on a special handling tool and placed in the robot cell. In the adjoining control room, our colleagues initiate the scan process using structured light, scanning the entire part (Fig. 1). As already described for radomes, the team can also select and confirm the areas not to be machined (e.g. service doors within the fan cowl) as these can vary from part to part. Afterward, the rest of the process runs fully automated with no supervision required.

Fig. 1: Model of the part superimposed (middle only) with the geometric and visual scan data

When repairing fan cowl doors at our ARC workshops in Hamburg, we now use the robot’s adaptive scanning/milling capabilities with accuracies in the range of ±0.1 mm (radomes: ±0.06mm). This level of accuracy is more than enough for the consecutive repair steps and even more impressive considering the various challenges that had to be overcome during the development. When we conducted our first test series with fan cowl doors, we found out that the part geometry of fan cowl doors can vary significantly (compared to the required accuracy), as these parts sometimes experience significant deformation during their extensive service life. The deviation of the different areas not to machine can easily reach magnitudes that go far beyond the required accuracy and strongly affect the reproducibility of the automated process for (supposedly) identical parts.

Additional deformation can occur during the actual milling process, when the plies and honeycomb material, that compensated thermal stresses during the manufacturing, are removed. The resulting deformations here also surpass the required accuracies and therefore had to be carefully taken into account when we developed and industrialized the robotic process. As already described for radomes, balancing the geometric variance of a part versus the required support for its machining process is also important for the fan cowl doors.

Fig. 2: Repair teams only intervene in test cases such as the one depicted: Once the scanning results are confirmed by a colleague, the entire milling process runs fully automated with no supervision required.

Another commonality between the fan cowl doors process and the radome process is our aim to relieve all employed personnel from uncomfortable and potentially unhealthy working conditions. During the manual repair process for fan cowl doors, the amount of manual grinding was so high that my colleagues often occupied our grinding room for several days. In addition to the fact that a full protection gear is always required due to the high amount of dust, the grinding work moreover had to be carried out in mostly unergonomic working positions. With the grinding of fan cowl doors now fully taken over by the robot, both the protective gear and the uncomfortable working positions became a thing of the past, much to the delight of my colleagues on the shop floor.

In this regard, it is important for me to highlight that our new robotic process never aimed at replacing our ARC colleagues. Instead, the idea is to significantly improve their working conditions and to allow them to concentrate on other, more rewarding tasks like challenging repairs instead of the “dull” and repetitive tasks connected with the time-consuming grinding and cutting. In fact, the feedback from my colleagues on the production floor was unanimous and many of them loudly expressed how happy they are that these stressful tasks are now taken over by the robot.

Nevertheless, the huge amount of dust and chips still resulting from the sometimes-extensive milling proved to be an important factor also for the automated robotic process. In some cases, the chip flow can easily reach more than 100 litres during the entire process and produce a mixture of various materials such as carbon fiber reinforced plastics (CFRP), aluminum, and core filler material. Consequently, we not only had to choose a robot system with sufficient dust protection or waste removal systems but also had to implement an effective explosion protection for the entire robot cell.

Even more special factors that need to be taken into account for the automated repair of Airframe Related Components will be highlighted in my concluding part 4 that focuses on Inlet cowls. Please stay tuned and feel free to ask questions in the comments!

#keepyouflying #mroexperts # avmro #lufthansatechnik #composites #compositematerials #robotics