EUROPEAN CONGRESS AND EXHIBITION ON ADVANCED MATERIALS AND PROCESSES - EUROMAT 2021

In order to prevent the activation of 60Co cobalt particles in the primary circuit of nuclear reactors, it is necessary to replace cobalt-based alloys of the Stellite type, used as abrasion-resistant coatings, with cobalt-free alloys. Several materials, including commercially iron- or nickel-based alloys, have been tested in the past, but none of them has equivalent characteristics to those of Stellites. The objective of this project is to design complex concentrated alloys resistant to abrasion using data mining, thermodynamic models and genetic algorithm.

First, models could be established, using a built database and data mining tools, such as pairwise comparison algorithms or the Gaussian process, to relate alloy composition to properties such as hardness and abrasion resistance. With the CALPHAD method (Calculation of Phase Diagrams), the microstructure of an alloy could be predicted for a given composition and preparation conditions.

Furthermore, the comparison of real and predicted microstructures shows that the welding-type deposition processes match to a structure calculated according to the Scheil-Gulliver solidification model. The average composition of the austenite provides information on the content of free chromium - to ensure corrosion resistance - and enables the calculation of empirical quantities such as martensite-start temperature.

Then, the optimisation algorithm used in this study is a non-sorting genetic algorithm type with multiple objectives and constraints, calculated from the models and thermodynamic variables. Different combinations of objectives and constraints have been tested to obtain various types of composition, among which two have been elaborated and tested (microstructure, hardness, wear and mechanical properties...).