Hier finden Sie wissenschaftliche Publikationen aus den Fraunhofer-Instituten.

Advanced methods in NDE using machine learning approaches

: Wunderlich, Christian; Tschöpe, Constanze; Duckhorn, Frank

Volltext urn:nbn:de:0011-n-5038434 (1.1 MByte PDF)
MD5 Fingerprint: 0546aec20af3d1b53aad3a40ac97a462
Erstellt am: 15.05.2019

Chimenti, Dale E. (Ed.) ; American Institute of Physics -AIP-, New York:
44th Annual Review of Progress in Quantitative Nondestructive Evaluation 2017 : 16–21 July 2017, Provo, Utah, USA
Woodbury, N.Y.: AIP, 2018 (AIP Conference Proceedings 1949)
ISBN: 978-0-7354-1644-4
Art. 020022, 8 S.
Annual Review of Progress in Quantitative Nondestructive Evaluation <44, 2017, Provo/Utah>
Konferenzbeitrag, Elektronische Publikation
Fraunhofer IKTS ()
non-destructive evaluation; acoustic signal analysis; machine learning

Machine learning (ML) methods and algorithms have been applied recently with great success in quality control and predictive maintenance. Its goal to build new and/or leverage existing algorithms to learn from training data and give accurate predictions, or to find patterns, particularly with new and unseen similar data, fits perfectly to Non-Destructive Evaluation. The advantages of ML in NDE are obvious in such tasks as pattern recognition in acoustic signals or automated processing of images from X-ray, Ultrasonics or optical methods. Fraunhofer IKTS is using machine learning algorithms in acoustic signal analysis. The approach had been applied to such a variety of tasks like quality assessment of gears in automotive industry; detection of cracks and impacts in aircraft materials, quality evaluation for musical instruments, determination of softness of tissue paper or condition monitoring of train wheels. The principal approach is based on acoustic signal processing with a primary and secondary analysis step followed by a cognitive system to create model data. Already in the second analysis steps unsupervised learning algorithms as principal component analysis are used to simplify data structures. In the cognitive part of the software further unsupervised and supervised learning algorithms will be trained. Later the sensor signals from unknown samples can be recognized and classified automatically by the same algorithms trained before. Recently the IKTS team was able to transfer the software for signal processing and pattern recognition on a small printed circuit board (PCB). Still the algorithms will be trained on an ordinary PC, however trained algorithms run on the hardware comprising of a Digital Signal Processor and a FPGA chip. The identical approach will be used for pattern recognition in image analysis of OCT pictures. Optical Coherence Tomography (OCT) is used to identify failures in planar ceramic components. After the depth related grey scale compensation and image noise reduction a sliding window will scan the picture to identify various failures in the ceramic material using machine learning algorithms. Again automated classification of the components is possible. These are just two examples how machine learning can be used in quality inspection and non-destructive testing. Some key requirements have to be fulfilled however. A sufficiently large set of training data, a high signal-to-noise ratio an optimized and exact fixation of components are key requirements to get useful results. So the well trained NDT expert is still required to develop and validate the concept. The automated testing can be done subsequently by the machine. It will be of high value to collect all test data and link it to any single component. By integrating the test data of many components along the value chain and even with field use data further optimization including lifetime and durability prediction based on big data becomes possible, even if components are used in different versions or configurations. This is the promise behind German „Industrie 4.0.“