Over the course of my studies, I was able to gather a wide range of impressions from various disciplines in mechanical engineering. Ultimately, however, every thought ended with me wanting to work in research and development. The field of non-destructive testing methods quickly caught my attention. After all, quality can only be assured with suitable testing methods, which can ultimately prevent damage to people, property and the environment.
My aim is to automate the application of the digital inferometric inspection methods of speckle holography and speckle shearography. Due to their properties, these methods would be suitable for 100% inspection in the context of series testing of increasingly complex components, but are being displaced by more common methods due to their complexity and the associated costs. This is where I would like to begin and reduce the complexity with the help of machine learning methods. This will turn the testing-methods into real alternatives that can also inspect modern materials for which conventional methods have not yet been able to deliver sufficiently reliable results.
I am researching on very rarely used methods for non-destructive testing. Recent research has overcome some of I do research on very rarely used non-destructive testing methods. Recent research has overcome some of the principle limitations, making them suddenly suitable for serial testing. However, the complexity-related limitations remain.
These result in high costs, as not only is the instrumentation required, but also trained personnel to set it up, use it and, most importantly, evaluate the results. As a result, more common methods such as thermography or ultrasonic testing are used in industry. However, the potential of digital speckle interferometry is enormous as it is largely material independent and works with different types of load. Other methods with specific materials and loads are comparatively limited and therefore less sustainable.
That's why I'm working on reducing the complexity of the process. This includes, for example, automatic defect detection through image processing using artificial neural networks, or determining the setup parameters prior to measurement through simulation and hyper-parameter optimisation.
A typical working day for me therefore consists of running tests, generating data sets and developing various methods to extract suitable information from this data for my automation approach. At the same time, I am always searching for appropriate literature and maintaining contacts with industry.
Jessica Plaßmann, Ann-Kathrin Bömkles, Christopher Petry, Michel Schuth, "Anwendung der Shearografie für die schnelle Prüfung von Bauteilen aus der Massenproduktion", Mai 2022, Konferenzband DGZfP Jahrestagung 2022, Kassel
Alexej Simeth, Jessica Plaßmann, Peter Plapper, "Detection of Fluid Level in Bores for Batch Size One Assembly Automation Using Convolutional Neural Network", August 2021, Advances in Production Management Systems. Artificial Intelligence for Sustainable and Resilient Production Systems (pp.86-93), DOI: 10.1007/978-3-030-85906-0_10
Jessica Plaßmann, Christopher Petry, Ann-Kathrin Bömkes, Michael Schuth, "Anwendung der shearografischen Dehnungsmessung mit transienter Wärmeanregung zur zerstörungsfreien Prüfung für die Detektion von Rissen in faserverstärktem Kunststoff im Vergleich zur Thermografie", Juni 2021, Konferenzband DGZfP-Jahrestagung 2021
"Anwendung der Shearografie für die schnelle Prüfung von Bauteilen aus der Massenproduktion", DGZfP Jahrestagung 2022, Kassel
"Anwendung der shearografischen Dehnungsmessung mit transienter Wärmeanregung zur zerstörungsfreien Prüfung für die Detektion von Rissen in faserverstärktem Kunststoff im Vergleich zur Thermografie", DGZfP-Jahrestagung 2021
You are leaving the official website of Trier University of Applied Sciences