Integrated modeling for the manufacture of Ni-based superalloy discs from solidification to final heat treatment

Abstract

Process models of the various stages of gas-turbine disc manufacture have been integrated to simulate the physical and microstructural transformations occurring within a nickel-based superalloy throughout the entire manufacturing route. Production of these critical rotating structural components requires several distinct processing stages: vacuum-induction melting (VIM), vacuum-arc remelting (VAR), homogenization heat treatment, cogging, forging, final heat treatment, and machining. During the course of these consecutive manufacturing stages, the various thermal and thermomechanical processes lead to significant changes in both the microstructural characteristics and internal stresses in the alloy. Although separate models have previously been developed to simulate the individual processing stages, this article describes how these models, which are explicitly expressed in terms of the initial and evolving microstructure, can be integrated to simulate the entire manufacturing process from secondary melting through to the final forging and heat treatment. The grain structure predicted for one stage is explicitly transferred as the initial conditions for the model, which simulates the grain evolution during the next step. This information also provides the basis of microstructure-explicit constitutive equations describing the material behavior. Industrial-scale manufacturing trials associated with the production of an INCONEL alloy 718 aeroengine turbine disc were used to validate the integrated model for grain size. The model also allows intrinsic or extrinsic defects entrained within the material during the initial solidification stage to be tracked through the subsequent processes. This provides a basis for calculating the areas of discs likely to be vulnerable to such defects and whether they might be removed during machining. It was shown that the microstructure of the alloy changes significantly throughout the process chain, the final microstructure and defect distribution at each stage being related to those formed in the previous stages. There was good agreement between the model predictions and experimental observations for both the intermediate and final processing stages. The implications of such integrated modeling for quality assurance through process control are discussed.