GOALI: Data-driven design of recycling tolerant aluminum alloys incorporating future material flows

Metals production contributes to 8% of greenhouse gas emissions worldwide. Improving materials efficiency must play a role in decarbonizing metals production because such strategies are available now and achievable along a shorter time horizon than novel production methods. One strategy to improve material consumption is through improved materials recovery using recycling. Recycling is particularly beneficial for aluminum, where energy benefits from use of recycled materials are much improved relative to primary consumption. Significantly improved recycling, both in terms of quantity and quality, can play a role in achieving decarbonization targets over the necessary greenhouse gas reduction timeline. To this end, metals production industries have set goals to increase the use of recycled content. In the face of an evolving scrap stream and shifts in product demand, these targets will not be achievable without coupling alloy design with considerations of future end-of-life scrap streams. However, alloy design has traditionally focused on improving performance without much regard for environmental impact or the ability to recycle the materials. This project will design new alloys by including recyclability in addition to performance metrics into the design process accounting for how scrap streams are expected to evolve in the future. The goal of this project is to create a design pipeline that can be used alongside traditional alloy design to create recycling-friendly alloys. Recycling-friendly alloys are defined as alloys that can incorporate a lot of scrap in their production (sink) but that can also be used in the production of a lot of other alloys (source). This research will address this trade-off between sink and source by (i) creating a materials distribution model to inform recycling and future scrap streams, (ii) developing a Bayesian optimization algorithm to effectively explore the alloy space, and (iii) identifying the design space using constraints on alloys compositions and properties. The material distribution model will be a combination of material flow analysis and blending model that will inform the future material streams (compositions, quantities, prices of scrap streams and future demand) and how those can be used to produce new alloys. This model will then be maximized over the alloy space using the efficient optimization method. Finally, the optimization to find the best alloy will be subject to constraints on compositions, thermodynamic quantities, and properties. Overall, this work focuses on alloy design that optimizes scrap use in projected material flows by integrating a materials distribution model, efficient optimization, and constrained design space. Since these elements have never been integrated together, not only is the approach unique, but the integration involves advances in each of these areas. Thus, the proposed work represents a step forward in sustainable alloy design and integrates and innovates on previously developed approaches in this area. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.