Julius Schoop, assistant professor in the Department of Mechanical Engineering, has received a $626,000 grant from the Department of Energy and additional partners for the project titled "AI-Enabled Discovery and Physics-Based Optimization of Energy-Efficient Processing Strategies For Advanced Turbine Alloys." Mechanical engineering assistant professor Hasan Poonawala and Dale Lombardo from GE Research are Co-PIs. A technical summary of the project is given below.
Novel turbine materials, such as advanced nickel-based superalloys and gamma titanium aluminide (γ-TiAl), are capable of operating at increasingly higher temperatures, which allows for more efficient turbine operation. However, while there have been tremendous advances in the Materials Science of turbine materials that operate at elevated temperatures and extreme loading conditions, the Manufacturing Science necessary to process them efficiently under manufacturing-specific thermomechanical regimes has been lacking.
In this project, the multi-organizational team of academic and industrial researchers from the University of Kentucky and GE Research will leverage Schoop's recently developed and preliminarily validated, real-time physics-based Digital Twin models of process/structure interactions (i.e., process-induced surface integrity) to advance a paradigm of fully-integrated computational materials engineering (ICME). Using our efficient process models as the core of a digital process simulator for a reinforcement learning algorithm, we will integrate industrial data and metrics of structure/performance/energy relationships and manufacturing-related energy metrics to optimize dynamic processing parameters for significantly improved life-cycle energy efficiency of advanced γ-TiAl low-pressure turbine (LPT) alloys, as indicated by a set of design relevant parameters (e.g., residual stresses and scrap rate).
While current raw material shortages and insufficient digital manufacturing tools are forcing turbine manufacturers to empirically adopt conservative (i.e. low productivity) and inefficient (i.e., high energy consumption) processing strategies to limit manufacturing scrap and expensive rework, our work will help identify more aggressive, yet reliable, strategies for processing scarce high-temperature alloys. However, when merely optimizing a single set of "static" process parameters, evolving process states, such as progressive tool-wear and workpiece material thermomechanical response, lead to rapid degradation of design-relevant parameters, such as surface finish and residual stresses. Therefore, we propose to adopt an AI-enabled approach, using a learning-based algorithm coupled to physics-based process models to not only optimize initial conditions but also predict dynamic (time-varying) process parameters for maximum manufacturing-related energy efficiency and performance.
The key objective and anticipated outcome of the project will be to demonstrate at least a 10% reduction in life-cycle embodied energy for a recently developed, γ-TiAl low-pressure turbine (LPT) alloy (GE proprietary material), through the adoption of the proposed AI-enabled process optimization approach. Rather than following the inefficient empirical paradigm, the proposed study will demonstrate the feasibility of adopting a digital, physics-based process design and optimization paradigm. The recurring need for manual intervention, rework, re-inspection causes significant WIP and unnecessary expense associated with delivering the requisite component quality. GE expects to reap significant cost and resource savings if an AI-optimized set of parameters can be applied to specific internal and proprietary machining operations.