University Coalition for Fossil Energy Research Commits Funding for Round 04 Projects

Wednesday, August 7, 2019

UNIVERSITY PARK, PA.  –  On December 12, 2018, UCFER released its fourth request for proposals (RFP) to the member Coalition. The RFP closed on February 6, 2019. A total of 41 proposals were received and reviewed by the Technical Advisory Council, Core Competency Board, external reviewers, and the Executive Council. A total of $1.9 million was approved to support the following six projects:

Upscaling Experimental Measurements to the Field Scale Using a Machine-Learning-Based, Scale-Bridging Data Assimilation Approach
Virginia Tech, 24-month project, $480,000

The proposed project aims to combine physics with data analytics to address this fundamental challenge by developing a novel machine-learning-based, scale-bridging data assimilation framework.

The specific challenges in the upscaling process are: 1) extreme sparsity of available data, and 2) disparate scales at which the data are located. To tackle these challenges, we use a scale bridging data assimilation framework to upscale core-plug-scale relative permeability measurements to the reservoir model cell scale, which allows the inference of smaller-scale parameters by assimilating the field data.

To address the problem of data sparsity, we use physics-informed machine learning to augment the amount of core-scale, multiphase flow property data by learning from both on-site and off-site data in order to predict multiphase flow properties based on rock/fluid features (e.g., contact angle, viscosity ratio, Capillary number, and rock’s mineralogy and petrophysical properties).

Here, on-site data refer to laboratory measurements on the core samples directly extracted from the CO2 storage reservoir of interest, whereas off-site data refer to the enormous amounts of core analysis data and X-ray CT image data collected and stored in NETL’s Geoimaging Lab and DOE’s databases. The four spatial scales involved in the upscaling workflow are the core scale (inch), lithofacies scale (1-10 ft), geologic model cell scale (10-100 ft), and reservoir model cell scale (100-1000 ft).

Specific research aims are as follows: Develop a statistics/simulation-coupled model to upscale core-scale relative permeability curves to the lithofacies scale (Aim R1). Use machine learning to predict relative permeability curves for different lithofacies and then use analytical methods to upscale the lithofacies-scale curves to the geologic model cell scale and reservoir model cell scale (Aim R2). Develop a scale-bridging data assimilation framework to calibrate the upscaled relative permeability curves using field-scale observation data of CO2 plume migration (Aim R3). The research will be conducted under the direction of Dr. Cheng Chen.

Porous Silicon/Lignite-Derived Graphene Composite Anodes for Lithium-Ion Batteries
University of North Dakota, 18-month project, $295,311

Our approach leverages the unique properties of the low-rank coal lignite-derived humic acid (HA) and the pSi. This project is a continuing collaboration between the University of North Dakota’s (UND)’s Institute of Energy Studies (IES) and a LIB-producing company, Clean Republic, LLC, based on their success in the development of cathode materials. In the previous project, the high purity of HA extraction procedure, a key step to this project, was already established.
The ultimate outcome of this project is a market-ready Si/G anode material for LIBs with a well-balanced performance and price.

This 18-month project will be divided into two phases. In Phase 1, efforts will focus on development and optimization of the synthesis method for pSi/G anodes and evaluation of its electrochemical performance at laboratory scale. In Phase 2, the optimal procedure will be scaled up.

This proposed project will have numerous technological and economic impacts and benefits: With a projected price of $150,000/ton, if our pSi/G could replace only 10% of global market demand for LIB anodes by 2020, the total value will be $3 billion. Our competitive technology will stimulate other competitors to accelerate the commercialization of their products, and that will trigger a chain reaction in the entire LIB industry, improving the competitiveness of existing LIB products and promote innovative products by domestic businesses.

This project can utilize the HA synergistically produced by the U.S. Department of Energy (DOE) project (DE-FE0027006) at UND aiming to extract rare-earth elements from lignite. The project will promote the local energy-based economy by  For example, the high-temperature treatment process (e.g., graphitization) can consume 15,000 kWh/ton, a quarter to a third of the total cost. That is very beneficial to North Dakota, a leading energy-exporting state in the United States. The research will be conducted under the direction of Dr. Xiaodong Hou.

Developing a Novel Ultrafine Coal Dewatering Process
Virginia Tech, 12-month project, $267,471

The U.S. coal production has been declining since 2014 due to competition from shale gas as an alternate fuel source for electricity generation. One bright spot for the coal industry is, however, the export market particularly for the metallurgical coals, for which coal quality plays an important role.

Coal fines are usually dewatered by filtration, for which a ten-fold decrease in particle size would require a ten thousand times higher pressure drop across a filter cake to achieve the same level of dewatering. In effect, mechanical dewatering has reached its limit.  It is, therefore, the objective of the proposed work is to develop a novel ultrafine coal dewatering technology.

In the proposed work, the water on the surface of coal is displaced by an organic solvent, and the spent solvent is recovered and recycled. If the solvent has a higher affinity for a coal than water does, the dewatering-by-displacement (DbD) process is spontaneous. Laboratory and pilot-scale test work showed that the process can reduce the moisture to less than 5 percent by weight of coal regardless of particle size.

If this process is further developed for commercial deployment, U.S. coal can be cleaned to any desired levels, so that the U.S. coal industry can produce high quality coals that are more competitive at the export market but also can be used to produce high-value carbon products such as carbon fibers, graphene, activated carbons, etc.

A specific objective of the proposed work is to develop an efficient solvent recovery system that can recover the bulk of a spent solvent by solid-liquid separation without a phase change and only a small amount of the residual solvent is recovered by evaporation in situ. The new dewatering system will be designed by developing a rigorous heat and mass transfer model and the related computer codes, followed by a laboratory test work. The results will be used to develop a scale-up criteria via engineering and economic analysis and modeling work. It is anticipated that sufficient information will be generated within one year to the extent that a larger prototype unit can be designed in cooperation with NETL, an equipment manufacturer, and the technology provider. The research will be conducted under the direction of Dr. Rui Qiao.

Computer vision and machine learning making the processing-microstructure-property connection in heat resistant alloys
Carnegie Mellon University, 24-month project, $240,000

Microstructure denotes the substructures that form from the interaction between composition and processing. A fundamental tenet of materials science is that Processing generates the microstructure that mediates material Properties – the PSP connection. In that sense, microstructure is the key link between what we control (processing parameters) and what we achieve (material performance). Given its ability to find relationships in large, complex data sets, machine learning (ML) seems tailor-made for exploring PSP connections. 

In this project, we develop and apply computer vision (CV) tools to create quantitative representations of microstructural images and apply machine learning (ML) methods to answer the question: Can we predict material properties from images of the material microstructure? We develop these tools to be relevant to the performance of heat resistant alloys used in power plants, initially 347 stainless steel and subsequently nickel-based superalloys. Our objectives are: Collect microstructural image data and property metadata for heat resistant alloy systems; Develop material-agnostic CV techniques to extract knowledge from microstructural images; Create ML systems to find relationships between microstructures and property metadata; and, Analyze and interpret the results to discover new PSP connections.

Developing a CV/ML system to discover PSP connections will involve three stages. In the first stage, we will assemble a data set of microstructural images and their associated property metadata. We will assess three approaches (identifying and extracting existing data from archives, collecting new data, and generating synthetic data) and select the best candidate based on data set attributes including size, cost, and quality. In the next stage, we will compare two CV image representation models (constituent segmentation and measurement and CNN-based hypercolumn pixels) in order to develop a CV approach to quantify the visual information contained in the microstructural images. Finally, we will choose an ML method suitable for learning from the selected image representation.

The advantages of the integrative CV/ML system are that it is autonomous and unbiased, so can potentially discover trends that humans can’t perceive; however, it is a black box. As the first application of these methods to heat resistant alloy design, this project is certain to provide critical experience and insight to inform the path forward, with the potential to revolutionize microstructural design for performance. The research will be conducted under the direction of Dr. Elizabeth A. Holm.

Development of a Novel Supersonic Hybrid Non-Equilibrium Plasma Reactor for Efficient and Tunable Co-Production of Hydrogen and Value-Added Solid Carbons
Princeton University, 12-24-month project, $193,000

The objective of this proposal is to develop and optimize an innovative supersonic hybrid nonequilibrium plasma reactor for efficient and tunable co-production of hydrogen and value-added solid carbons with negative CO2 footprint.

The research consists of the following four major thrusts: (1) development of a supersonic hybrid non-equilibrium plasma reactor, (2) characterize and optimize the hybrid plasma properties and the residence time of supersonic reactor to control non-equilibrium species excitation and energy relaxation to increase the yield of hydrogen and the quality of valued solid carbon production, (3) understanding the effect of composition and impurity variations in natural gas on carbon and hydrogen production, and (4) analysis of the energy, mass balance, and carbon footprints for the plasma synthesis in comparison with the commercial thermal cracking method.

Specifically, in the proposed supersonic hybrid non-equilibrium plasma reactor technique, we introduce two novel design concepts to maximize the yield of hydrogen and the value of solid carbon. Firstly, the use of hybrid gliding arc with a nanosecond repetitive discharge or a microwave discharge will realize both high electron density and high electron energy for efficient non-equilibrium plasma enhanced natural gas dissociation to hydrogen and carbon.

Secondly, the use of supersonic flow not only reduces the pressure to shift the chemical reaction equilibrium towards hydrogen and carbon formation but also increases the plasma uniformity and reduces flow residence time so that the vibrational excitation mode of methane remains in non-equilibrium for fast dissociation into hydrogen and carbon. By using this novel plasma reactor, the plasma properties and the yield of hydrogen and solid carbon, and the morphology of carbon will be quantitatively examined by using Thomson/Raman scattering, micro-gas chromatography, and scanning electron microscopes.

Finally, the yield of hydrogen and carbon, mass balances, the quality of the energy cost, and CO2 footprint will be analyzed using the US electricity production and compared with the commercial thermal cracking method. Collaborations will be established with NETL scientists making use of their existing facilities such as microwave reactor and materials characterization equipment to develop a supersonic hybrid gliding arc and microwave reactor. The research will be conducted under the direction of Dr. Yiguang Ju.

Metal-free Catalyzed Synthesis of Novel Carbon by Carbon Allotrope Seeds
Penn State, 12-month project, $189,451

This proposal seeks to overcome catalyst deactivation and address catalyst regeneration by negating the need for separation and recovery of the catalyzed carbon from catalyst while simultaneously optimizing the amount/type of high-value carbon produced in order to improve process economics in thermo-catalytic decomposition of natural gas (TCD). The project will demonstrate carbon allotrope fragments as catalyst “seeds” by which to synthesize novel solid carbon forms under TCD conditions while also eliminating the need for catalyst separation and regeneration.

Specific objectives include: 1. Identify the different solid carbon forms “catalyzed” by carbon allotrope segments. 2. Optimize the carbon catalyst to obtain the desired solid carbon types. 3. Optimize the experimental parameters to obtain the best solid carbon yields. 4. Optimizing the amount of hydrogen produced by these catalyst systems.

Carbon allotrope fragments will be used as “catalysts” to seed directed growth of novel carbon forms. Allotropes will include CNTs, graphene and fullerene-like carbons. These fragments will be activated prior to TCD, by partial oxidation, mechanical milling and plasma exposure, respectively. Acting as consumable catalysts, the carbons will seed directed growth of similar morphologies under TCD conditions using natural gas as feedstock. Reaction conditions will range between 700 – 1,100 ℃ using varied synthetic natural gas mixtures, to test by design of experiments natural gas component contributions and effects of including impurities such as CO2 and H2S. Toward scalability and economic assessment, catalyst development and testing will proceed from fixed bed to entrained flow. Process intensification will be sought using microwave energy to increase radical concentrations driving the carbon growth.

Natural gas can be decomposed to manufacture hydrogen and solid carbon products without producing CO2. The co-production of hydrogen and high-value solid carbon materials from natural gas offers opportunities to 1. Reduce the costs associated with large-scale hydrogen energy product; 2. Utilize domestic natural gas for manufacturing energy and synthetic carbon products; 3. Enable scalable and economic production of novel carbons for advanced construction and structural materials. The research will be conducted under the direction of Dr. Randy Vander Wal.