Digital Repository, ICF12, Ottawa 2009

Font Size: 
Prediction of Fracture Energy of IN738LC Superalloy using Neural Networks
Nafisa Bano, Atef Fahim, Michel Nganbe

Last modified: 2012-10-03


IN738LC is a cast polycrystalline nickel based super alloy, primarilyused for first stage gas turbine blades which operate under severe loadingconditions and temperatures. The development of relationship among operatingtemperature, microstructure and mechanical properties of the material is veryimportant in order to understand its failure mechanisms and fracture behavior. Forthis purpose, this paper presents a three layered feed forward back propagationneural network model capable of predicting the fracture energy of IN738LC.Temperature, strain rate, gamma prime precipitate size, yield strength, ultimatetensile strength and percentage elongation to failure taken from literature are usedto train, validate and test the model. Results obtained from the neural networkmodel describe experimental values accurately for given operating conditions.Therefore, it can be used to correlate microstructural parameters to strength andtoughness properties under real operating conditions.

Full Text: PDF