Author(s) |
Rangasayee Kanna, Gerry L. Knapp, Benjamin Stump, Ying Yang, Peeyush Nandwana, Alex Plotkowski, Ryan Dehoff, Vincent Paquit |
Abstract Scope |
As Integrated Computational Materials Engineering (ICME) tools are developed and applied to materials challenges, the processing of multi-variate datasets is becoming increasingly important. High throughput alloy design strategies can use CALPHAD methods, process models, and more, which generate highly dimensional datasets where insights can be difficult to obtain via conventional two-variable process maps. To address this problem, we apply advanced visualization techniques for the visualization of high-throughput thermodynamic calculations for multicomponent alloy systems. Here, we present two applications of this method: 1) impact of powder composition variations on delta-ferrite formation during binder jet additive manufacturing (AM) of stainless steel 316L, and 2) impact of wire composition and process parameters on hot cracking during laser hot-wire directed energy deposition (DED) of dissimilar aluminum alloys. Using high throughput CALPHAD methods, we simulate simultaneous composition variations for each alloying element within alloy specifications, similar to real-world variations between batches.
In the first case study, we consider the formation of the delta-ferrite phase during binder jet AM of stainless steel 316L as a function of alloy composition variation within the ASTM specified allowable ranges. Equilibrium phase fraction of delta-ferrite was calculated at 1000 composition variations using the TC-Python interface to the ThermoCalc software. Advanced data visualization tools were able to show the acceptable ranges for alloying element compositions to reduce delta-ferrite formation.
In the second case study, we analyze the hot cracking susceptibility of a dissimilar aluminum interface during laser hot-wire DED. To assess the composition-dependent aspects of hot cracking, a criterion was used as proposed by S. Kou by taking the maximum slope of the solid fraction curve, |d(fs^0.5)/dT|, near fs = 1. The composition of the interface was estimated using a simple analytical process model to calculate the dilution of the wire with the substrate as a function of laser power, wire temperature, and wire feed rate. Variation of the composition of the wire and substrate alloys was considered for each set process parameters, resulting in a total dataset of 25,000 Scheil solidification calculations. Visualization enabled analysis of the relationships between the process parameters, concentrations levels of alloying elements in both the wire and substrate, Scheil solidification data, and composition-based hot cracking susceptibility.
Overall, high-throughput thermodynamic calculations combined with advanced data visualization techniques provide a method to analyze the interactions between different alloying elements, process variables, and material properties for multivariate datasets. For more complex ICME pipelines, such a visualization and data mining approach will be vital to timely and insightful data analysis. |