Inspection Sensor and Evaluation Criteria for Assessing the Inner Surface Quality of Holes in CFRP Parts

Informations projet

HScan

N° de convention de subvention: 686355

DOI

10.3030/686355

Projet clôturé

Date de signature de la CE 3 Decembre 2015

Date de début 1 Janvier 2016

Date de fin 31 Mars 2018

Financé au titre de

• SOCIETAL CHALLENGES - Smart, Green And Integrated Transport

Coût total

€ 425 950,00

Contribution de l’UE

€ 370 637,88

Coordonné par

PROFACTOR GMBH
Austria

Periodic Reporting for period 2 - HScan (Inspection Sensor and Evaluation Criteria for Assessing the Inner Surface Quality of Holes in CFRP Parts)

Période du rapport: 2017-07-01 au 2018-03-31

Résumé du contexte et des objectifs généraux du projet

While being a common manufacturing process in the aeronautics industry, finding the right tools and parameters for drilling in composites remains a challenge. In today’s production, quality of the bores is ensured by keeping the drilling process within tight process tolerances, after qualification of the process via destructive testing.

The extended use of carbon fiber composite parts in aircraft construction (e.g. Airbus A350) requires improved process technologies for the drilling of holes in carbon fibre composites. For the assembly of hull components more than 10.000 holes need to be drilled. The quality of the drilling is important, because it has impact on the mechanical strength of the parts, when they are assembled using fasteners (rivets) that are inserted into the holes. Existing methods focus on the entry and exit area of the drilled area outside of the hole. If defects (especially delaminations) are detected in this area, ultra-sound is used for further inspection. The inner surface of the holes, however, is currently not inspected. Experiments with roughness measurement systems did not lead to conclusive results, because the inhomogeneous surface texture of carbon fiber materials inside the hole, does not lead to useful results, when standard roughness measurements are used. Consequently, better methods for assessing the inner surface of drilled holes are required.

This HScan project developed a machine-vision-based inspection tool for the inspection of the inner surface of bores. The sensor works with an endoscope equipped with light sources that allow illumination from eight different angles. The inspection result is based on set of quality criteria, Rating the Surface Quality and defects.

Travail effectué depuis le début du projet jusqu’à la fin de la période considérée dans le rapport et principaux résultats atteints jusqu’à présent

Work on HScan project started, by conducting a literature review. First results and conclusions could already be discussed at a meeting at the Airbus site in St. Nazaire. This was also the starting point for work on industrial use cases.
FACC provided seven test panels that covered UD and fabric material and hybrid stacks and all hole sizes, relevant for the project (diameters 4.8mm to 8.0mm). Defects on the hole walls were produced by varying drill parameters such as spindle speed, feed and driller wear. Three of the panels were cut open, to allow for a direct view of the drill hole walls.
Profactor compiled defect catalogues based on the cut open panels, by manually inspecting each drill hole and noting down defects found and surface quality. The visual appearance of defect types, had to be derived from examples in literature. No visual inspection of the inner hole walls is currently done and no expert based ground truth was available.
Based on the defect catalogues and findings from papers, Profactor compiled the mathematical formulation of the inspection rating, starting with a 100% score for a perfect surface. Defects found on the surface are weighted and the score lowered accordingly.
Profactor based the sensor design on experiments with different endoscopes and illuminations, resulting in two sensor prototypes. The first prototype worked with a standard endoscope and seven additional light sources, one of the light sources was a LED opposite from the lens. In a second improved model, the opposite LED could be dropped. Profactor ordered a custom endoscope without in-built illumination. This increased the optical diameter and significantly improved the image quality.
Main components of the HScan sensor are:
• Optical sensor with camera, endoscope and illumination via optical fibres
• Linear axis and motor for automated lowering into the borehole
• Fixture for use with drilling templates
• Casing and handle
• Graphical user interface for sensor control and results view
• Image analysis software
Two sets of ~140 drill hole samples (50% fabric and 50% UD) fabricated by FACC underwent destructive mechanical testing to correlate drilling parameters and resulting surface quality. The conducted bearing tests gave insights on durability of a rivet fastening. During the last stages of the project, extended tests in production took place at project partner FACC. The sensor handling was found to be good and the images informative. FACC also views the HScan sensor as a possible opportunity to generate feedback to production process from automated digital inline quality control. Further development activities are planned to enhance the accuracy of the image segmentation and the classification of the defects.
The sensor was presented at JEC fair in March 2018 in Paris and a patent for the HScan sensor was filed.

Progrès au-delà de l’état des connaissances et impact potentiel prévu (y compris l’impact socio-économique et les conséquences sociétales plus larges du projet jusqu’à présent)

HScan sensor provides beforehand unavailable means to measure drill hole quality in CFRP parts in an objective and non destructive way. The foreseen impact is increased productiveness and cost reductions for the European Aerospace industry based on:

• Cost reduction by prolonged life time of drills
Currently drills are replaced after drilling a certain number of holes. The number of holes is determined through experiments, but obviously has to consider a safety margin, because the actual quality of the drilled hole is not yet known. Due to the fact, that the materials to be drilled are particularly difficult (e.g. titanium + CFRP or aluminium + CFRP) drills in general do not last very long and need to be exchanged often. The HScan sensor allows a more regular inspection of the drilled holes and an objective assessment of the quality. This enables a more accurate decision when the drill needs to be exchanged, thus reducing unnecessary safety margins and extending the lifetime of the tools.

• Optimizing drilling procedures
Setting up a drilling process for composite material is difficult, especially when the drilling involved two very different materials such as aluminium or titanium and carbon fibre. The optimal drilling parameters are found through experimental work. Up to the HScan project, there was no suitable technology to generate objective information about the hole quality, where different sets of drilling parameters can be compared at a quantitative level. HScan provides this opportunity and thus enables the much quicker definition of suitable drilling parameters and their optimization.

• Automated Inline quality control
Manual inspection of drilled holes is time consuming and even if tools such as endoscopes are used, the results rely on human interpretation. Such manual inspection processes cannot be used in series production, especially considering that the drilling of single holes takes just a few seconds. HScan offers the opportunity to perform fully automatic inspection during the drilling process. In the case of robotic drilling systems, the sensor can be integrated through a tool changer and will allow the automatic quality control for each hole (if needed) or a sample of recently drilled holes (if process stability allows). For manual drilling processes the sensor is equipped to fit into the drilling template, so that a quick manual check is possible. Possible problems can thus be detected early and the need for re-work of the manufacturing of scarp parts can be avoided.

HScan Sensor Prototype

Project Logo

Defects found on a test sample