Novus Light Home

Blog

Bonding with vision

Bonding materials with adhesives might appear to be a relatively straightforward manufacturing process. However, ensuring the reliability of the bonds that are formed between materials is of paramount importance, especially in the automotive and aerospace industries, where safety is critical. In such industries, manufacturers must be certain that the materials that are joined together will be as reliable as those joined by more conventional techniques such as welding or riveting.

To ensure that the bond between two materials will remain effective over the lifetime of a product, manufacturers must ensure that the surfaces of the materials are free from any impurities or surface defects that may adversely affect the effectiveness of the bond. To ensure that they are, the surfaces of the materials are typically pre-treated by mechanical methods such as abrasion, chemical methods such as solvent cleaning, electrolytic methods or even plasma cleaning, where an ionized gas interacts with a material substrate to remove contamination. Whatever process is chosen to clean the materials, however, the surfaces of the materials must then be inspected to ensure that the cleaning process has been effective.

For optimum adhesion, an adhesive must thoroughly “wet out” any surface to be bonded – a term that refers to how easily the adhesive can form an intimate contact with, and spread over, a given substrate. Various methods already exist for such testing, such as using test inks, employing a water-break test or by making contact angle measurements.

Test inks are usually applied to the surface of the material with a brush. If the brush stroke edges are stable for two seconds, then the surface is deemed to be easily “wettable”. In the water-break test, a part being tested is withdrawn rapidly from a bath of water and the surface of the water on the material is observed for any breaks that would indicate that the surface is contaminated. In the contact measurement technique, distilled water is sprayed over the surface and the contact angle between the water bead and the surface is measured. The more surface wets out, the smaller the contact angle is between the bead and the surface.

Vision system solution

Now, engineers at Munich-based Automation W+R – working in conjunction with a team at Puchheim-based Stemmer Imaging – have developed an alternative approach based on the use of a vision system. Based on a patent held by Fraunhofer IFAM, the company’s bonNDTinspect system enables surfaces that are to be bonded or painted to be inspected to detect any contamination that might be present.

But instead of measuring the side contact angle of a single drop, the new system analyzes the size and distribution of thousands of micro-droplets from a top view, dramatically increasing the statistical significance of the results. The result is that manufacturers will now be able to fully automate the inspection of a variety of materials prior to any subsequent automated bonding processes on their production lines.

The bonNDTinspect system itself is a contactless, non-destructive robot-assisted system for testing the wettability and bondability of surfaces. After a particular part has been cleaned (and after the appropriate pre-treatment process has been employed to activate its surface), water is sprayed onto the surface of the part from an ultrasonic nozzle which wets the surface with an extremely fine mist of ultrapure water as the nozzle is moved across the surface of the material by a robotic arm.

A Teledyne Dalsa Linea monochrome camera with a 2k pixel resolution (which is enclosed in the same housing as the ultrasonic nozzle) captures several image slices of the droplet pattern on the surface before the water evaporates. The images of the captured droplet pattern are then transferred to Industrial PC from Beckhoff (which also provides the PLC functionality for the robotic system) over a GigE interface where the images are stitched together to form a 2D image that are analyzed with image processing software.

droplet pattern

A droplet pattern on a surface is automatically captured by a camera and evaluated by means of image processing software to determine whether a surface has been contaminated. In this case, a fingerprint has been detected.

According to Célian Cherrier -- the Product Manager for the bonNDTinspect system at Automation W+R -- the development of the new vision-based test system posed several challenges. In the process of wetting the surface of the material, more than 30,000 droplets are created per second. These must be detected quickly by the Teledyne Dalsa Linea monochrome camera (which is capable of a 25kHz line rate) as the camera moves across the surface of the part at a standard speed of 100mm/sec because the use of ultra-high purity water means that the surfaces of the material under test dry without residues within a very short time.

The choice of optics and the lighting were also critical elements in the system. In the development of the bonNDTinspect system, the Automation W+R engineers recognized that they would need to capture images of the water droplets that move through the field of view of the system, and that the position of those droplets would not be repeatable. If the working distance was not identical for each droplet that the lens was imaging, the measurement of each object would vary due to a shift in magnification. To solve this issue, and to enable accurate images of the water droplets to be captured across the entire field of view of the system, Automation W+R chose to use a telecentric lens from Sill Optics Compared to entocentric lenses, only parallel principle rays are imaged through the lens which ensures a constant magnification of each of the droplets across the entire depth of field.

The developers of the system also recognized that the line scan camera they chose for the application would also require a greater level of illumination than a system that employed an area scan cameras, because of the short exposure times that were specified. To produce a light source that would enable them to do so, whilst also illuminating the surface of the part homogeneously such that the image processing system could effectively subtract the background surface image from the dark borders and  light interior images of the droplets, they developed a custom dome lighting system. The dome light radiates indirect light at various angles enabling the target objects to be uniformly lit with diffuse light from all angles and eliminating hotspots, ensuring the stability of the image inspection operation.

Image processing

Once the images of the surface of the material have been captured, it is then possible to determine whether a surface fulfils the criteria for the subsequent process steps, on the basis of the droplet distribution analyzed by the image processing system.

According to Célian Cherrier, the captured images are first processed by the system to produce an image in which all the droplets of water are highlighted in black while the background material is white. Once that has been achieved, an image subtraction algorithm removes the background from the image and a blob detector is employed to detect the blobs. After the blobs of water have been identified in the image they are then analyzed in detail to determine what percentage of the surface of the material is covered in droplets and what the size of those droplets are. In addition, the system software also filters the image to ascertain whether there are local variations in droplet size that could indicate that the surface of the material has been contaminated. Any contaminated areas are highlighted on screen together with an indication of whether a particular part has passed or failed the inspection.

Thanks to the development of the new vision-based inspection system, the wettability of entire surfaces can now by inspected automatically, reliably and in-line. This provides a key advantage over methods such as contact angle measurement or application of test inks which are primarily used in the laboratory. What is more, the system can be integrated seamlessly into a production process, enabling parts on which surfaces have been insufficiently activated (or which have some surface contamination which would weaken a joint) to be removed from the production process.

The first users of the system were the bonding specialists at Fraunhofer IFAM, who used a bonNDTinspect prototype at their Bremen site for testing the wettability of surfaces following atmospheric pressure plasma activation. Another bonNDTinspect systems has been successfully installed at the German Aerospace Center -- Deutsches Zentrum für Luft- und Raumfahrt (DLR) -- which is conducting research on bonding processes in the aerospace industry.

Written by Dave Wilson, Senior Editor, Novus Light Technologies Today

Caption for figure at top: After a part has been cleaned, water is sprayed onto the surface from an ultrasonic nozzle which wets the surface with an extremely fine mist of ultrapure water as the part is moved across the surface of the material by a robotic arm. A camera then captures several image slices of the droplet pattern on the surface before the water evaporates. The images of the captured droplet pattern are then transferred to Industrial PC from (which also provides the PLC functionality for the robotic system) where the images are stitched together to form a 2D image that are analyzed with image processing software.

Labels: vision-based inspection,wettability,adhesive bonding,bonding,vision,software,Dave Wilson

Back Back to Features
 

Illuminating Products

Copyright © 2018 Novus Media Today Group, LLC. All rights reserved. Website design and build by MM Design.