Recycling-oriented alloy design is a crucial part of material sustainability, as it reduces the need for raw material extraction and minimises environmental impact. This requires that scraps be reused or repurposed effectively, even when the scraps are co-mingled and have higher costs for further sorting and separation. In this work, we explore an alloy design concept by creating a compositionally flexible domain that can recycle multiple alloy grades and yet maintain relatively consistent properties across chemical variations. This is demonstrated through the Fe-Cr-Ni-Mn system to identify compositionally flexible austenitic stainless steels (CF-ASS) and accommodate the recycling of mixed austenitic stainless steel scraps. Alloys within the nominal composition spaces exhibit relatively consistent mechanical properties and corrosion resistance despite significant variations in different alloy compositions. We illustrate how we can utilise the compositionally flexible austenitic stainless steels to recycle mixed 200 and 300-series stainless steel and ferronickel scraps, demonstrating its practical viability. While this demonstration focuses on the stainless steel system, the underlying principles can be extended to other systems related to mixed scrap recycling.

In this research a machine learning model for predicting the rotating bending fatigue strength and the high-throughput design of fatigue resistant steels is proposed. In this transfer prediction framework, machine learning models are first trained to estimate tensile properties (yield strength, tensile strength and elongation) on the basis of composition and critical process conditions. Then, based on the predicted tensile properties, transfer models are trained to estimate fatigue strength. The results are compared with those of a similar model not having such a transfer layer. The transfer prediction framework shows high accuracy for fatigue strength prediction with a remarkably high tolerance to limitations in the amount of calibration data available for training. By combining the transfer prediction framework with evolutionary algorithms, a robust high-throughput alloy design model is achieved requiring only tens of fatigue data points to get a decent reliability. The newly designed steel showed the predicted high fatigue strength. The method as presented here might also be applicable to other alloy design challenges in which only a limited database for the property to be optimized is available.

With the development of the materials genome philosophy and data mining methodologies, machine learning (ML) has been widely applied for discovering new materials in various systems including high-end steels with improved performance. Although recently, some attempts have been made to incorporate physical features in the ML process, its effects have not been demonstrated and systematically analysed nor experimentally validated with prototype alloys. To address this issue, a physical metallurgy (PM) -guided ML model was developed, wherein intermediate parameters were generated based on original inputs and PM principles, e.g., equilibrium volume fraction (V_f) and driving force (D_f) for precipitation, and these were added to the original dataset vectors as extra dimensions to participate in and guide the ML process. As a result, the ML process becomes more robust when dealing with small datasets by improving the data quality and enriching data information. Therefore, a new material design method is proposed combining PM-guided ML regression, ML classifier and a genetic algorithm (GA). The model was successfully applied to the design of advanced ultrahigh-strength stainless steels using only a small database extracted from the literature. The proposed prototype alloy with a leaner chemistry but better mechanical properties has been produced experimentally and an excellent agreement was obtained for the predicted optimal parameter settings and the final properties. In addition, the present work also clearly demonstrated that implementation of PM parameters can improve the design accuracy and efficiency by eliminating intermediate solutions not obeying PM principles in the ML process. Furthermore, various important factors influencing the generalizability of the ML model are discussed in detail.