PhD Candidate (m/f/d) - Robot-based computed tomography for non-destructive testing of components

The University of Augsburg is looking for a motivated PhD candidate (m/f/d) for the project "Robot-based computed tomography for non-destructive testing of components". This exciting opportunity is available in our teaching and research unit "Mechanical Engineering", which focuses on innovative testing and inspection technologies for structural components in the automotive and aerospace industries.
 
In this project, we aim to further improve our existing robot-based computed tomography system. These systems use X-ray radiation for non-destructive testing of components and are used for precise component testing in the development and during series production. By combining two collaborating robots carrying the X-ray source and the detector, we enable scanning of large and complex geometries.
 
Your main tasks will include developing AI models to optimize robot movements, simulating and applying CT scans, data analysis and reporting, as well as collaboration within the team. Your qualifications may include a completed master's degree in mechanical engineering, robotics, computer science, or a related field. Solid knowledge in or interest in AI, machine learning, especially reinforcement learning. Proficiency in programming languages such as Python, C++ or similar. Strong analytical skills, with the ability to work independently and in a team. Excellent communication skills in German and English.
 
If you are interested in joining our team, please feel free to contact us proactively by phone or email.

Doctoral Researcher (m/f/d) – Robot-assisted Component Inspection Strategies for Automotive or Aircraft parts

The University of Augsburg is seeking a motivated Doctoral Researcher (m/f/d) for the Robot-assisted Component Inspection Strategies for Automotive or Aircraft Parts project. This exciting opportunity is within our research group at the research and teaching unit “Mechanical Engineering”, focusing on innovative inspection and testing technologies for structural components in the automotive and aerospace industries.

In this project, we aim to enhance existing robotic systems for the mechanical testing of structural components. By simulating real-world stresses on components, we will utilize AI techniques to enable robots to autonomously carry out precise mechanical tests. The project will leverage reinforcement learning for trajectory optimization, factoring in component positions, test paths, and force models, with real-time adjustments based on force-torque sensor data.

Your key responsibilities include development of AI models, robot-assisted testing, simulation and real-world application, data analysis and reporting, collaboration. Qualifications could include a completed Master’s degree in Mechanical Engineering, Robotics, Computer Science, or a related field. Strong experience or interest in AI, Machine Learning, particularly reinforcement learning and Sim2Real transformations. Proficiency in programming languages such as Python, C++, or similar. Strong analytical skills, with the ability to work independently and in a team. Excellent communication skills in German and English.

If you are interested in joining our team, please contact us by phone or email.

Doctoral Researcher (m/f/d) – Predictive Maintenance in Gear Monitoring

For the research group “Condition Monitoring” at the research and teaching unit "Mechanical Engineering" at the University of Augsburg, we are looking for a doctoral researcher (m/f/d).

Gear wear reduces remaining useful life and can lead to serious system failures. Therefore, it is essential to record gear wear when applying predictive maintenance. This makes maintenance more efficient and saves time, costs and resources. When monitoring gears, machine learning enables rapid and precise analysis of sensor data to detect conditions or deviations that indicate impending failures.

Your research should enable the development of new concepts for monitoring gearboxes. Our research primarily includes the creation and use of sensor data for gear monitoring. The aim is to expand the application to real application scenarios together with industrial partners. A gearbox test bench is used to demonstrate the technology. To do this, you will set up a test stand for non-destructive testing of a gearbox demonstrator. With the help of software solutions, the measurement data can be evaluated and errors or anomalies can be detected early.

If you are interested in joining our team, please contact us by phone or email.

AI-based Gear Condition Monitoring – Predictive Maintenance

The aim of the thesis is to analyze gearbox failures with a focus on monitoring anomalies resulting from various types of failures. An innovative AI-based approach will be developed to link sensor-based data with specific failure types. By combining modern data processing techniques with machine learning methods, the goal is to achieve precise identification and classification of anomalies, thereby improving the reliability and efficiency of gearboxes in a sustainable manner.

More information can be found in the hyperlinked pdf:

AI-based Gear Condition Monitoring – Predictive Maintenance

Robot Simulation for interaction with deformable bodies in an Industrial assembly

Traditional robotic assemblies rely on position-based control, which lacks the adaptability of human operators who can make fine adjustments based on tactile feedback. Integrating force and torque data into the robot control feedback loop offers significant advantages, particularly for leveraging advanced AI-based control. Extensive force and torque data are required to train AI-based robotic control efficiently. Generating this data through simulation optimises time and resources. This thesis aims to develop a generalised pipeline for generating synthetic force and torque data, focusing on an industrial assembly use case. The scenario involves interactions with deformable bodies, simulating real-world robot operations.

More information can be found in the hyperlinked pdf:

Robot Simulation for interaction with deformable bodies in an Industrial assembly

Coupling FEM simulation to robot simulations for contact-rich assembly tasks

Robot simulation software helps replicate robotic behaviour, enabling data augmentation and preempting unsafe collision paths without additional resource expenditure. However, these tools fall short in accurately modelling the mechanical behaviour of objects in the environment compared to traditional FEM software. This thesis aims to bridge this gap by integrating robot simulation with FEM-based modelling. The goal is to achieve realistic, dynamic mechanical behaviour of target objects in contact-rich assembly tasks, enabling more accurate and practical robotic simulations.

More information can be found in the hyperlinked pdf:

Coupling FEM simulation to robot simulations for contact-rich assembly tasks

Experimental Studies on Acoustic Emission Sensor Verification

This master's thesis project aims to investigate the feasibility of predicting the calibration responses of AE sensors using a simpler setup involving a Laser Doppler Vibrometer. The goal is to create a predictive model that can assess the quality of sensors based on a more accessible and cost-effective verification procedure.

More information can be found in the hyperlinked pdf:

Experimental Studies on Acoustic Emission Sensor Verification

Designing and developing a smart sensor for process and condition monitoring

For real-time monitoring, which is a central topic in the working group “condition monitoring” and is already being implemented through the use of ultrasonic sensor, a smart sensor is now to be developed. For this purpose, this work aims to form a basis by enhancing a commercially available sensor of lower frequency with a microcontroller / single-board computer in such way, that it is able to collect, process and forward the data to a central computer via a suitable communication interface. The comparison of the developed sensor with a corresponding commercial system should round off the work.

More information can be found in the hyperlinked pdf:

Designing and developing a smart sensor for process and condition monitoring

Recycling of silicate foundry sands and binders

This thesis aims at developing a process that takes advantage of the solubility of the sodium silicate binder to wash it at elevated temperatures after the casting process, using only water without additional chemicals. The washing process  produces wastewater containing dissolved foundry binder and other impurities. The second goal is to recycle this water so it can be processed into a binder and does not have to be disposed of, thereby closing the material cycle for both binder  and sand.

More information can be found in the hyperlinked pdf:

Recycling of silicate foundry sands and binders

Simulation of Friction Stir Welding

The aim of the work is to set up a finite element model in Abaqus that calculates the temperature and the formation of residual stresses during friction stir welding.

More information can be found in the hyperlinked pdf:

Simulation of Friction Stir Welding

Recycling of silicate foundry sands and binders

This thesis aims at developing a process that takes advantage of the solubility of the sodium silicate binder to wash it at elevated temperatures after the casting process, using only water without additional chemicals. The washing process  produces wastewater containing dissolved foundry binder and other impurities. The second goal is to recycle this water so it can be processed into a binder and does not have to be disposed of, thereby closing the material cycle for both binder  and sand.

More information can be found in the hyperlinked pdf:

Recycling of silicate foundry sands and binders

Development of a glass mold for observing molten metal

This thesis aims at developing a test mold that allows the mold filling to be observed through glass panels. This is recorded by  cameras and used to validate foundry simulations. Therefore, the most challenging casting geometries possible are  to be implemented, which push the existing simulation models to their limits.

More information can be found in the hyperlinked pdf:

Development of a glass mold for observing molten metal

Simulation of Friction Stir Welding

The aim of the work is to set up a finite element model in Abaqus that calculates the temperature and the formation of residual stresses during friction stir welding.

More information can be found in the hyperlinked pdf:

Simulation of Friction Stir Welding