Research Seminar on AI: AI Enhanced Estimation of Microstructure-Property-Relationships for Heavy-Section Castings

  Logo KI-Center und IoP



+49 241 80 99550



Dienstag, 13.12.2022, 16.00 Uhr

With the transition towards a climate-neutral society, the power of modern wind turbines increased resulting in size-increased cast iron components within the turbine. Such heavy-section castings present a variety in the local microstructure due to varying solidification behavior within a single component. Thus, high safety factors are considered in component design, leading to increased component weight and manufacturing cost. A more detailed understanding of local microstructure and local fatigue strength within a component is required to face these challenges. As a comprehensive understanding using classical experimental analysis is not feasible, methods from artificial intelligence are employed to enhance information acquired from a limited amount of specimens. Here, microstructure reconstruction is performed using a Continuous Conditional Generative Adversarial Network (CCGAN), which is trained using the available experimental data. Utilizing the continuous conditionality of the trained network, the microstructure of uncharacterized component regions can be estimated. It is demonstrated that a CCGAN can be trained even if the training data is subject to a strong scattering within individual classes. In order to determine the fatigue strength for a given microstructure, an ANN is trained to cut the computational cost of simulation-based micromechanical analysis. Analyzing the constructed pipeline, the chances and limitations of the AI-enhanced component design of heavy-section castings are discussed.

Felix is currently a PhD student at RWTH Aachen University. His research is focused on the determination of microstructure-property relationships for heavy-section castings. He joined the Institute for Materials Applications in Mechanical Engineering as a PhD student in 2020. There, his research is settled in the field of renewable energies, focusing on lightweight design for heavy-section castings based on local material properties. Before his PhD, he received an M.Sc. in Mechanical Engineering and a B.Sc. in Mechanical Engineering at RWTH Aachen University.


Weitere Informationen

Um aktuelle Informationen zum KI-Seminar zu erhalten, besteht die Möglichkeit, sich in einen Mailverteiler einzutragen. Dazu muss eine E-Mail mit Betreff "subscribe" an geschickt werden.