Experimental data were used to develop metamodels to predict high temperature alloy chemistry trends influencing stress-to-rupture and time-to-rupture of Nickel based superalloys. Chemistry optimization utilized evolutionary neural networks, bi-objective genetic programming and pruning algorithm. Optimization results were compared with the experimental data and IOSO optimization algorithm. Response surfaces were developed through various modules available in a commercial optimization package. Pareto optimized chemistries were tested using thermodynamic database, FactSage™, by studying the phase distribution as a function of temperature of manufacture and exposure. Uniformity in the amount of critical phases over 0–1200 °C range confirmed high temperature stability for optimized alloys.

Paper published with permission.

This content is only available via PDF.