Wednesday, June 18, 10:30 AM - 12:00 PM

Technical Session

Session Chairs:
Michael Enright
Christian Amann

Presentations

Note: Presentations may start a few minutes before the time listed in the schedule.

The use of machine learning & probabilistic methods for accelerating simulations and risk assessment of high-energy rotating components has become a standard in the gas turbine engine industry. For commercial aircraft engines, regulatory agencies in both the US and Europe recommend the use of probabilistic approaches for certification as summarized in FAA Advisory Circulars AC33.14-1, AC33.70, and AC39-8. The need for the extension of these methods to assess the risk and improve the efficiency of power generation engines is also becoming increasingly recognized in the international gas turbine engine community.

Machine learning/probabilistic methods developments, including accelerating simulations with surrogates, optimization, and robust design are applied to durability/lifing, materials development and discovery, process optimization, prognostics/predictive maintenance, and other general areas using probabilistic and machine learning methods. The Machine Learning/Probabilistic Methods and Applications will include:
-Probabilistic lifing and data analysis
-Machine learning
-Uncertainty quantification
-Probabilistic/robust optimization design
-Model calibration and Verification/Validation
-Surrogate/emulator modeling
-Design space exploration
-Sensitivity analysis
-Bayesian methods