The selected projects are:
ArcelorMittal USA Research LLC and Lawrence Livermore National Laboratory; DEFECT PREDICTION OF STEEL SLABS WITH THE APPLICATION OF HPC AND AI; Federal Funding Amount: $400,000. The iron and steel industry is the fourth largest energy-consuming industry in the US. The iron and steel industry consumed an estimated 6% (~1470 PJ) of the total energy consumed in the whole US manufacturing sector. About 80% of the total energy consumed in the steel industry is used to produce steel slabs via the continuous casting route (96% of steel in the US today is produced via this route). Therefore, being able to produce defect free slabs (by making it right, first time, every time) would lead to a huge benefit in terms of energy savings and reduction of CO2 emissions during the steelmaking process by minimizing wastes and increasing quality and yield. Therefore, as the most recycled material on earth (more than all other materials combined), a slight improvement to steel production would have a lasting positive impact on the environment, the current decarbonization efforts and on the US energy landscape.
ClearSign Technologies Corporation and National Renewable Energy Laboratory; DEVELOPMENT OF AN ULTRA LOW NOx INDUSTRIAL HYDROGEN BURNER MODEL FOR DECARBONIZATION; Federal Funding Amount: $400,000. ClearSign Technologies Corporation, with the National Renewable Energy Laboratory (NREL) proposes a Virtual Test Furnace (VTF) using High-Performance Computing (HPC) to expand the application of 100% fueled hydrogen process burners. This project addresses retrofitting and optimizing high temperature furnaces with 100% hydrogen capable ultra low NOx burners. The focus will be on managing the high flame and furnace temperatures while controlling flame chemistry to minimize NOx emissions and ensure optimal heat flux control and operational flexibility. The VTF will simulate and optimize hydrogen burner performance, reduce development costs and time-to-market, aiding decarbonization of energy-intensive industries by nearly 276 million metric tons annually.
DNV and National Energy Technology Laboratory; ACCELERATED DESIGN AND DEVELOPMENT OF HIGH STRENGTH NICKEL ALLOYS RESISTANT TO HYDROGEN EMBRITTLEMENT; Federal Funding Amount: $400,000. High-strength precipitation hardened (PH) nickel-based superalloys are widely used for key components in extreme environments that require excellent mechanical properties and corrosion resistance. However, they are susceptible to hydrogen embrittlement (HE), as evidenced by significant decrease in fracture toughness and increase in crack growth rate (CGR). This project aims to design new cost-effective PH nickel-based alloys with improvement in fracture toughness by 25% and reduction in CGR by 5x compared to commercial IN725 in hydrogen. The team adopts an Integrated Computational Materials Engineering (ICME) approach by integrating high throughput (HT) CALPHAD and density functional theory (DFT) calculations with multi-objective machine learning to accelerate high-performance alloy design. Guided by high-performance computing (HPC), the best candidate alloy will be processed and it’s HE performance characterized. This research will greatly reduce the time for new alloy design with enhanced HE performance, enabling breakthroughs in materials for the hydrogen economy.
EarthEn Energy Inc. and Oak Ridge National Laboratory; MODELING OF A SUPERCRITICAL CO2 COMPRESSOR-AS-A-TURBINE (CaT) FOR ENERGY STORAGE SYSTEM; Federal Funding Amount: $400,000. Current energy storage solutions rely on separate turbines and compressors, thereby entailing significant efficiency losses and high capital costs. To address these challenges, EarthEn proposes the development of a Compressor-as-a-Turbine (CaT) technology. This innovative system integrates the functionality of both a turbine and a compressor into a single machine with variable blade geometry, capable of operating efficiently with supercritical CO2 (sCO2) during disparate charge-discharge cycles. This technology promises to enhance the operational flexibility and cost-effectiveness of energy storage systems by combining the use of sCO2 as an efficient thermomechanical fluid, as well as enhancing the feasibility of long-duration energy storage systems. The CaT system will be designed computationally at Oak Ridge National Laboratory (ORNL), in collaboration with Concepts NREC, leveraging their experiences with advanced turbomachinery, modeling compressible flow, GPU-accelerated fluid dynamics simulations, and advanced manufacturing techniques. Successful CaT designs could promote rapid adoption in the sCO2-turbomachinery sector.
First Light Solutions and Lawrence Livermore National Laboratory; IN SILICO SIMULATION OF CO2 UPTAKE IN CATION-EXCHANGED ZEOLITES IN THE PRESENCE OF POINT-SOURCE CAPTURE HUMIDITY LEVELS; Federal Funding Amount: $400,000. Energy generation accounts for approximately one-third of all greenhouse gas emissions, making point-source carbon capture essential to achieving the United States’ net-zero goals. However, existing carbon-capturing materials are characterized by high energy requirements and costs well above the Department of Energy’s targets. Zeolites are a class of carbon-capturing material characterized by low regeneration energies and costs, but which typically struggle to achieve significant CO2 capture levels in humid post-combustion streams. In this project, we aim to identify novel cation-exchanged zeolites that exhibit enhanced CO2 uptake under humid flue gas conditions via a phase transition. To do so, we will partner with LLNL and UC Davis to leverage high-performance computing to screen candidate structures using in silico methods and validate our simulations through industry-led laboratory synthesis and experimentation. Through this project, we will discover and manufacture carbon-capturing zeolites that deliver cost reductions and performance improvements for point-source capture.
GE Vernova Advanced Research and Oak Ridge National Laboratory; ENHANCING THE DURABILITY OF GAS TURBINE COMPONENTS AGAINST HYDROGEN EMBRITTLEMENT USING HIGH-PERFORMANCE COMPUTING (HPC); Federal Funding Amount: $400,000. Identifying materials susceptible to hydrogen embrittlement (HE) is critical for efficiently delivering hydrogen-friendly gas turbines (GTs) to market. GE Vernova Advanced Research (GEVAR) and Oak Ridge National Laboratory (ORNL) will execute a high-performance computer project to develop a framework using high-throughput crystal plasticity (CP) coupled with machine learning (ML) models to address HE in GT components. A CP-generated large dataset will be analyzed to identify key features affecting HE and used to train ML models. GEVAR and GE Vernova Gas Power’s (GEV-GP) expertise and world-class materials dataset will be employed to validate and calibrate simulation results. From identified ML models, the team will succinctly explore the high-dimensional space of various alloys and conditions pertinent to the GT operating envelope.
Helix Earth Technologies, Inc and Argonne National Laboratory; ENHANCING CO2 CAPTURE RATES USING MICRON-SCALE DROPLET SPRAY REACTORS; Federal Funding Amount: $400,000. Absorber technologies are crucial for reducing CO2 emissions thereby mitigating climate change and are a pivotal element of the energy future of the US. Traditional CO2 capture methods using liquid sorbents are hindered by slow reaction kinetics resulting in large system footprints. This project proposes a novel CO2 absorber using micron-scale sprays of liquid amines, which significantly increase the liquid-gas surface area, resulting in order-of-magnitude increases in CO2 capture rates and increased system efficiency. Helix Earth Technologies has developed a patented, high-efficiency method for droplet capture that enables cost-effective development of these droplet absorbers. We aim to optimize spray dynamics and reaction kinetics to improve CO2 capture efficiency by leveraging advanced CFD modeling at Argonne National Laboratory, which will help accelerate the development of scalable, cost-effective CO2 capture systems that substantially reduce emissions, decreases system size, and enhance operational flexibility, helping transform CO2 capture processes and other liquid-gas chemical processes.
Nucor and Lawrence Livermore National Laboratory; REUSE OF WASTE HEAT IN INDUSTRIAL PROCESSES TO REDUCE CARBON EMISSIONS AND ENERGY USE; Federal Funding Amount: $400,000. Nucor is seeking to increase energy efficiency of steel production by recycling the heat in exhaust gases from Electric Arc Furnaces (EAF). Waste heat from steel manufacturing accounts for 28-30 percent of energy inputs. The heat from the gases can be used for other energy intensive processes. However, the exhaust gases contain a large amount of particulates that must be cleaned before the gas can be used for other applications. The analysis and design of these processes is not well understood for high temperature applications. Nucor proposes an HPC4EI collaboration with Lawrence Livermore National Laboratory (LLNL) to conduct High Performance Computing analysis of these processes to better understand how their design and operation leads to efficient reductions in the particle load. Based on this, the project will develop engineering approaches to design cleaning systems to meet specific requirement of the processes that use the hot exhaust gas.
RTX Technology Research Center and Lawrence Livermore National Laboratory; FACTOR MANUFACTURING VARIABILITY IN THE DESIGN OF EFFICIENT FILM COOLING TECHNOLOGY (PHASE 2 PROJECT); Federal Funding Amount: $200,000. Reductions in cooling air flow can improve the thermal efficiency of gas turbine engines. Predictive models for near-wall cooling flow physics are, however, needed to optimize turbine cooling designs. Moreover, such models need to capture the effects of deviations from nominal design introduced by manufacturing processes. The primary objective of this proposal is to leverage advanced machine learning (ML) methods along with high-fidelity computational fluid dynamics (CFD) simulations and supercomputing to develop predictive but affordable data-driven models for near-wall mixing and heat transfer. Phase 1 developed a validated wall-resolved CFD simulation framework, an ensemble ML-augmented wall model, and proof-of-concept demonstration of a hybrid CFD-ML workflow for modeling near-wall mixing behavior in a nominal film cooling configuration. Phase 2 seeks to further advance the generalizability and broaden the applicability of the data-driven wall modeling approach by employing geometric deep learning techniques and extending it to practical scenarios incorporating manufacturing-induced variations.
Shell and Argonne National Laboratory; OPTIMAL DESIGN OF CARBONCAPTURE UNITS UNDER UNCERTAINTIES IN FEED COMPOSITIONS AND OPERATIONAL CONDITIONS; Federal Funding Amount: $400,000. Carbon capture, utilization, and storage (CCUS) technology offers an effective way to reduce carbon emissions for various industries lacking scalable decarbonization options, such as power generation, refining, and chemical production. Despite the successful application of optimization methods in carbon capture unit (CCU) design current deterministic, steady-state optimization approaches incur suboptimality when faced with uncertainties in feed conditions and/or dynamic operation conditions. This proposal aims to develop an HPC-enabled CCU design method that can consistently achieve a 95% carbon capture rate by performing design optimization with explicit considerations of uncertainties and dynamics. These problems exhibit high computational complexity, as they embed nonlinear physics, uncertainty scenarios, and discretized differential and algebraic equations (DAEs). We propose to address this challenge by leveraging (i) HPC-enabled nonlinear programming software, developed by the ANL and MIT team during the Exascale Computing Project, and (ii) HPC resources provided by the national laboratories.
