Session: 36-02 Surrogate Model-based applications

Paper Number: 151212

Lifing Prediction Under Casting and Boundary Condition Deviations in Turbine Blades With 3-D Transformer-Based Neural Operators 

Turbine design optimization is a critical aspect of the engineering process, ensuring efficient performance and longevity of these essential power generation components. One of the significant challenges in turbine design is addressing creep and thermal-mechanical fatigue (TMF) analysis to fully exploit the design space and create robust designs. This paper presents a novel approach that utilizes artificial intelligence (AI) techniques to model the relationship between deviations in geometry and boundary conditions of a turbine blade and its lifespan. By incorporating AI-driven algorithms and machine learning models, this study aims to develop an efficient, automated design optimization process that adapts to the evolving requirements of turbine technologies. The methodology encompasses data collection and analysis, feature extraction, AI model training, and application to the optimization problem. Leveraging historical performance data and relevant material properties, the AI models are trained to predict the impact of creep and TMF under varying geometry and boundary condition deviations, as well as different operating conditions. Specifically, a novel process chain is employed to generate deviations in geometry and boundary conditions for turbine blade designs. These deviations include modifications such as stacking and rotating the blade airfoil, altering the wall thickness of the airfoil's pressure and suction sides, and adjusting the location of the ribs within the blade core. Additionally, variations in cooling air supply and hot gas boundary conditions are explored. The resulting data from this process is utilized to train a point-cloud-based AI model, which serves as a 3D surrogate model evaluated for its predictive performance. The proposed AI-driven approach facilitates the identification of optimal design parameters that minimize the adverse effects of geometry and boundary condition deviations, ultimately leading to enhanced turbine performance and extended service life. By addressing these challenges, AI-driven optimization can significantly contribute to the development of more efficient and sustainable power generation solutions for the future.

Presenting Author: Jason Abdallah Siemens Energy

Presenting Author Biography: Jason Abdallah is a Senior Turbine Engineer and Mechanical Integrity Technical Lead at Siemens Energy, where he leverages over eight years of experience in mechanical design and analysis, specifically focusing on the hot gas path components of gas turbines. He holds a Diploma in Aerospace Engineering from the Technical University of Berlin.

With a strong passion for integrating artificial intelligence and machine learning into turbine design optimization, Jason Abdallah serves as the program manager for a government-funded project aimed at advancing these technologies within the industry. In addition to his technical expertise, he provides mentorship to junior engineers, participates in design reviews, and collaborates closely with manufacturing teams to ensure that product cost and quality objectives are met.

Jason Abdallah is dedicated to contributing to the innovation and excellence of Siemens Energy's products and services, as well as advancing the field of turbine engineering. His commitment to fostering a culture of continuous improvement and technological advancement positions him as a key player in the evolution of power generation solutions.

Authors:

Jason Abdallah Siemens Energy
Stefan Depeweg Siemens AG
Kai Liebelt Siemens Energy Global GmbH & Co. KG