The Effects of Process‐Route Variations on the Tensile Properties of Closed‐Die Waspaloy Forgings, via Statistical Modeling Techniques

Abstract

Statistical data modelling techniques are used to predict and optimise the tensile properties of a 2-stage closed-die Waspaloy forging using real industrial data. This data includes chemical composition, inner and outer region billet grain size, γ’ solvus temperature, furnace set temperature and preform and final form furnace durations. These parameters are used to model the behaviour of the ultimate tensile strength (UTS), yield strength (YS), elongation (E) and the reduction of area (R of A), at 8 varied locations in the forging via linear and non-linear regression. This has been done to determine the sensitivity of variations in process route and alloy chemistry, in order to optimise tensile properties. Introduction A typical process route for critical section Waspaloy forgings involves a 2 stage forging operation, using 2 separate dies with perform and final heat treatments, followed by air cooling. They then undergo 3 heat treatment stages: a solution heat treatment, a stabilisation and an ageing process. Mechanical property assessment is then done on a number of sacrificial forgings within the batch at different locations in order to determine if the products reach customer specification and have a high degree of consistency. The aim of this work is to determine the effect of process route variations using various statistical data modelling analyses. This is based on the notion that the majority of the forging process route is constant and the only means of influencing tensile properties would be via the obtained parameters. These relationships may then be reverse engineered to provide optimum process routes that provide more reliable products thus eliminating time and money spent on non-conformance investigations. As such the work has two objectives: 1. Verify the significance of altering each input independently of other factors, which is not always applicable industrially. 2. Highlight the variations in modelling accuracies in the modelling of UTS, YS, E and R of A and how variations in process route to optimise one property have an influence on the other properties.