Adaptive numerical modelling of high temperature strength, creep and fatigue behaviour in Ni based superalloys

Abstract

The mechanical behaviour of high performance Ni alloys is required for many applications and where experimental data is not readily available then a suitable predictive approach would be beneficial. There are numerous routes to achieve this, however, here the data driven neural network method has been adopted to produce models for the tensile, creep and fatigue performance of nickel base alloys. These models have been successfully developed and tested against a range of criteria. The tensile and creep models have displayed excellent fidelity to known nickel alloy behaviour, while good correspondence was also achieved for the fatigue properties (both strain and stress controlled). Potential routes to further improve the performance of these models have been discussed.