(AerospaceBook)DETERMINATION OF THE THERMAL EXPANSION COEFFICIENT OF IN738LC WITH DUPLEX SIZE GAMMA PRIME USING NEURAL NETWORK

摘要

The objective of this work is to predict the thermal expansion coefficient of duplex gamma primeprecipitate structure of IN738LC (first stage gas turbine blade material). Levenberg-Marquardt and BayesianRegularization Back Propagation neural network is used and the thermal expansion coefficient is described as afunction of temperatures, γ’ precipitate sizes, chemical composition and elastic constants. The result ofLevenberg-Marquardt modeling is in good agreement with the experimental result available in literature.Therefore, it can be useful for design optimizations for minimizing thermo-mechanical stresses between basealloy and potential protective coatings, and consequently for increasing the service life of turbine blades.