Post on 17-Mar-2020
NDT 4.0: Robotics and defect recognition for automated quality
control
Philippe Meynard
1VisiConsult X-ray System & Solutions GmbH
Brandenbrooker Weg 2-4
23617 Stockelsdorf
Germany
Tel +49 451 2902860; fax +49 451 29028622
e-mail p.meynard@visiconsult.de
Abstract
Industries are living today a change called Industry 4.0. After the version 3.0 based on
the electronics, computers and automation, this new evolution is now oriented with
cloud, connected devices, Internet of things and networks. Actually, on aerospace
companies, the need is to maintain a high-quality production despite higher volume, less
time for control. Industry 4.0 could be an opportunity to match this new challenge.
In this paper, we present a description about NDT 4.0 and examples of this new concept
on aerospace companies. NDT 4.0 is the declination of industry 4.0 dedicated to the
NDT world. It gives new tendencies, key points for this market linked to Industry 4.0
with a description of the cycle (Autonomous robot; simulation; system integration;
Internet of things; Cybersecurity; Cloud computing; Additive manufacturing;
augmented reality and big data)
We will give examples of applications and customized systems applied with the NDT
4.0 philosophy. We will mention new challenges and developments using x-ray
technologies.
1. Introduction
Industry 4.0 will be the new standard for companies. It includes several processes linked
together. The idea is to perform an internal real time feedback loop providing a fast and
precise diagnostic in a way to correct process and issues.
Figure 1: Industry 4.0 loop
More
info
about
this
art
icle
: htt
p:/
/ww
w.n
dt.
net
/?id
=25045
Copyright 2019 - by the Authors. License to Cofrend and NDT.net.
2
NDT4.0 is the use of this industry 4.0 philosophy dedicated to NDT technics
2. Cloud inspection
Based on the IOT (internet of things) and Cloud Computing, an example of NDT 4.0
could be the cloud inspection. Several applications can be done using this feature. You
can, as example, share images collected from several x-ray systems using cloud
capabilities. Every system is inter connected and can manage the new information flow.
Another example could be found on the maintenance area. It provides reliable
information for predictive maintenance.
Figure 2: multi-modality inspection
3. Additive Manufacturing
Additive manufacturing is more and more used by aerospace companies. The capability
to develop new parts with complex geometry and internal structure opens new field in
optimization of products. Unfortunately, the AM process generates different defects like
cracks, porosities. An NDT technic used for inspection is Computed Tomography (CT)
Figure 3: CT Scan of blade
3
4. Simulation and augmented reality
Simulation provides an effective and costless solution to analyze the system and see
results on change of parameters. With simulation, the customer can appreciate the
influence of changes and decide to implement the optimized solution.
Figure 4: Example of simulation on X-ray inspection system.
With augmented reality, additional information can be shown on the scene
Figure 5: example of augmented reality
4
5. Robotics and automation
Automation and robot are seen as the most important source to create competitive
advantage. Cost reduction or demand for increased process safety are the strongest
drivers of this movement. Automating NDT processes has always been challenging due
to compliance and liability issues.
Figure 6: High end robot system for inline X-ray inspection
6. Application on NDT : Automatized defect recognition using AI
With growing processing power, Automated Defect Recognition (ADR) is a rising
technique in the field of X-ray inspection. Automated systems need to be qualified in
order to provide reliable and reproducible inspection results and have to comply towards
industry standards.
Typical ADR applications are the detection of porosities, inclusions or cracks in casting
parts. It is possible to define certain ROIs and check defect metrics like defect density,
defect distance, defect size, defects per area and many more. Thresholds can be defined
dynamically. Training of the system does not require any programming skills and can be
done through level II or III personnel. This drives down production costs and reduces
5
the inspection bottleneck, while increasing the reliability and process safety. An
integration to Industry 4.0 factory solutions allows full traceability of the inspection
process down to single part level.
More demanding tasks like automated measurement, completeness or density checks
can be performed through the unique VAIP (VisiConsult Automated Image Processing)
module. Complex test sequences can be performed on static images or in real time.
Typical applications for defect recognition are casting parts. The ADR software should
automatically detect classify typical casting defects like porosities and inclusions.
Figure 7 shows the result of such a classification. On the left side is the evaluated part
and defect list, while on the right one can see the original X-ray image. Different parts
of the image allow different error thresholds. This can be defined by the Level III
through inspection Region of Interests (ROI). Depending on the inspection guideline
criteria for classification can be calculated defect depth (contrast), distance to surface or
defect size or defects per area.
Figure 7: Result of an automated defect recognition (ADR)
As the ADR evaluation works on the original 12-16 bit image it can operate in all
different part thicknesses and detect even slightest defects. As humans can only
perceive less than 100 grey values, only limited thicknesses can be evaluated at the
same time without window leveling. This gives automatic evaluation a big advantage
for complex parts like structured castings. Figure 8 shows the evaluation per ROI. For
this part only indication inside the green ROI (center) are relevant and lead to a rejected
part. All findings on the outside are omited.
6
Figure 8: ADR Evaluation of a ROI
AI can be useful when the reproducibility of the part is not good enough for the
traditional approach. An example is the detection of defects on welds. You can see in
figure 9 and 10 the detection provided by the software. The blue area is the ROI defined
for the inspection (around the seam)
Figure 9: Weld with ROI
7
Figure 9: Detection of defect on weld
The system can find the defects but to achieve this result, it was necessary to “teach”
how to find defects based on a high quantity of images. Previously, these images were
checked by experts to highlight defects. This learning phase is mandatory and need a lot
of work.
You will see on figure 10 that the solution is not perfect but will be improved by
accumulating more and more images and information.
Figure 10: probability of detection
8
7. Conclusion
This paper shows the principle of NDT 4.0 and the benefit that companies will get from
this new approach. Automation is considered actually as the highest potential to disrupt
or create competitive advantage.