Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Lead optimization”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 919 records · Page 51

Automated property optimization via ab initio O(N) elongation method: Application to (hyper-)polarizability in DNA

An automated property optimization method was developed based on the ab initio O(N) elongation (ELG) method and applied to the optimization of nonlinear optical (NLO) properties in DNA as a first test. The ELG method mimics a polymerization reaction on a computer, and the reaction terminal of a starting cluster is attacked by monomers sequentially to elongate the electronic structure of the system by solving in each step a limited space including the terminal (localized molecular orbitals at the terminal) and monomer. The ELG-finite field (ELG-FF) method for calculating (hyper-)polarizabilities was used as the engine program of the optimization method, and it was found to show linear scaling efficiency while maintaining high computational accuracy for a random sequenced DNA model. Furthermore, the self-consistent field convergence was significantly improved by using the ELG-FF method compared with a conventional method, and it can lead to more feasible NLO property values in the FF treatment. The automated optimization method successfully chose an appropriate base pair from four base pairs (A, T, G, and C) for each elongation step according to an evaluation function. From test optimizations for the first order hyper-polarizability (β) in DNA, a substantial difference was observed depending on optimization conditions between “choose-maximum” (choose a base pair giving the maximum β for each step) and “choose-minimum” (choose a base pair giving the minimum β). In contrast, there was an ambiguous difference between these conditions for optimizing the second order hyper-polarizability (γ) because of the small absolute value of γ and the limitation of numerical differential calculations in the FF method. It can be concluded that the ab initio level property optimization method introduced here can be an effective step towards an advanced computer aided material design method as long as the numerical limitation of the FF method is taken into account.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Additive‐Free, Low‐Temperature Crystallization of Stable α‐FAPbI 3 Perovskite

Abstract Formamidinium lead triiodide (FAPbI 3 ) is attractive for photovoltaic devices due to its optimal bandgap at around 1.45 eV and improved thermal stability compared with methylammonium‐based perovskites. Crystallization of phase‐pure α‐FAPbI 3 conventionally requires high‐temperature thermal annealing at 150 °C whilst the obtained α‐FAPbI 3 is metastable at room temperature. Here, aerosol‐assisted crystallization (AAC) is reported, which converts yellow δ‐FAPbI 3 into black α‐FAPbI 3 at only 100 °C using precursor solutions containing only lead iodide and formamidinium iodide with no chemical additives. The obtained α‐FAPbI 3 exhibits remarkably enhanced stability compared to the 150 °C annealed counterparts, in combination with improvements in film crystallinity and photoluminescence yield. Using X‐ray diffraction, X‐ray scattering, and density functional theory simulation, it is identified that relaxation of residual tensile strains, achieved through the lower annealing temperature and post‐crystallization crystal growth during AAC, is the key factor that facilitates the formation of phase‐stable α‐FAPbI 3 . This overcomes the strain‐induced lattice expansion that is known to cause the metastability of α‐FAPbI 3 . Accordingly, pure FAPbI 3 p–i–n solar cells are reported, facilitated by the low‐temperature (≤100 °C) AAC processing, which demonstrates increases of both power conversion efficiency and operational stability compared to devices fabricated using 150 °C annealed films.

36 MATERIALS SCIENCE↗

Function-Biased Choice of Additives for Optimization of Protein Crystallization: The Case of the Putative Thioesterase PA5185 from Pseudomonas aeruginosa PAO1

The crystal structure of PA5185, a putative thioesterase from Pseudomonas aeruginosa strain PAO1, was solved using multi-wavelength anomalous diffraction to 2.4 {angstrom}. Analysis of the structure and information about the putative function of the protein were used to optimize crystallization conditions. The crystal growth was optimized by applying additives with chemical similarity to a fragment of a putative PA5185 substrate (CoA or its derivative). Using new crystallization conditions containing this function-biased set of additives, several new crystal forms were produced, and structures of three of them (in three different space groups) were determined. One of the new crystal forms had an improved resolution limit of 1.9 {angstrom}, and another displayed an alternative conformation of the highly conserved loop containing Asn26, which could play a physiological role. Surprisingly, none of the additives were ordered in the crystal structures. Application of function-biased additives could be used as a standard optimization protocol for producing improved diffraction, or new crystal forms, which may lead to better understanding of the biological functions of proteins.

36 MATERIALS SCIENCE↗

Optimizing the optimizer for physics-informed neural networks and Kolmogorov-Arnold networks

Physics-Informed Neural Networks (PINNs) have revolutionized the computation of PDE solutions by integrating partial differential equations (PDEs) into the neural network’s training process as soft constraints, becoming an important component of the scientific machine learning (SciML) ecosystem. More recently, physics-informed Kolmogorv-Arnold networks (PIKANs) have also shown to be effective and comparable in accuracy with PINNs. In their current implementation, both PINNs and PIKANs are mainly optimized using first-order methods like Adam, as well as quasi-Newton methods such as BFGS and its low-memory variant, L-BFGS. However, these optimizers often struggle with highly nonlinear and non-convex loss landscapes, leading to challenges such as slow convergence, local minima entrapment, and (non)degenerate saddle points. In this study, we investigate the performance of Self- Scaled BFGS (SSBFGS), Self-Scaled Broyden (SSBroyden) methods and other advanced quasi-Newton schemes, including BFGS and L-BFGS with different line search strategies. These methods dynamically rescale updates based on historical gradient information, thus enhancing training efficiency and accuracy. We systematically compare these optimizers – using both PINNs and PIKANs – on key challenging PDEs, including the Burgers, Allen-Cahn, Kuramoto-Sivashinsky, Ginzburg-Landau, and Stokes equations. Additionally, we evaluate the performance of SSBFGS and SSBroyden for Deep Operator Network (DeepONet) architectures, demonstrating their effectiveness for data-driven operator learning. Our findings provide state-of-the-art results with orders-of-magnitude accuracy improvements without the use of adaptive weights or any other enhancements typically employed in PINNs. More broadly, our work reveal insights into the effectiveness of quasi-Newton optimization strategies in significantly improving the convergence and accurate generalization of PINNs and PIKANs.

97 MATHEMATICS AND COMPUTING↗

Distributed Control and Optimization of Virtual Inertia in Power Systems

The integration of renewable energy sources into power grids through inverter-based resources (IBRs) reduces system inertia, leading to faster frequency dynamics and potential grid instability. This paper proposes a distributed, consensus-based approach for the real-time control and optimization of inertia sources during system disturbances, enhancing both system stability and economic performance. Furthermore, we integrate this approach with the traditional inertia placement problem, demonstrating its suitability for existing frameworks. We validate its effectiveness in a 4-bus test system and a two-area 8-bus test system, showcasing its potential for scalable, dynamic inertia management in low-inertia grids.

Yadav, Ajay [ORNL] (ORCID:000000016111881X)↗

Effect of frequency detuning on Brillouin lasing in microcavities

We have theoretically investigated Brillouin lasing in microcavities in the case when the Brillouin shift and the intermode spacing of the microcavity do not coincide. It is shown that, despite the rise of the lasing threshold, a significant increase in the Brillouin signal intensity can be achieved in this case compared to the resonance one. A necessary condition for this effect is to select an optimal value of detuning of the pump radiation frequency from the frequency of the corresponding microcavity mode. The result of increasing the lasing threshold is also a narrowing of the Brillouin signal lasing range in the nonresonant case, which leads to a decrease in the signal noise level provided the detuning value is optimally selected. Analytical calculations are confirmed by the results of numerical modelling. (paper)

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Renormalization group improved optimized perturbation theory: Revisiting the mass gap of the O(2N) Gross-Neveu model

We introduce an extension of a variationally optimized perturbation method, by combining it with renormalization group properties in a straightforward (perturbative) form. This leads to a very transparent and efficient procedure, with a clear improvement of the nonperturbative results with respect to previous similar variational approaches. This is illustrated here by deriving optimized results for the mass gap of the O(2N) Gross-Neveu model, compared with the exactly known results for arbitrary N. At large N, the exact result is reproduced already at the very first order of the modified perturbation using this procedure. For arbitrary values of N, using the original perturbative information only known at two-loop order, we obtain a controllable percent accuracy or less, for any N value, as compared with the exactly known result for the mass gap from the thermodynamical Bethe Ansatz. The procedure is very general and can be extended straightforwardly to any renormalizable Lagrangian model, being systematically improvable provided that a knowledge of enough perturbative orders of the relevant quantities is available.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Utilization of Nitrophenylphosphates and Oxime-Based Ligation for the Development of Nanomolar Affinity Inhibitors of the Yersinia pestis Outer Protein H (YopH) Phosphatase

Our current study reports the first K{sub M} optimization of a library of nitrophenylphosphate-containing substrates for generating an inhibitor lead against the Yersinia pestis outer protein phosphatase (YopH). A high activity substrate identified by this method (K{sub M} = 80 {micro}M) was converted from a substrate into an inhibitor by replacement of its phosphate group with difluoromethylphosphonic acid and by attachment of an aminooxy handle for further structural optimization by oxime ligation. A cocrystal structure of this aminooxy-containing platform in complex with YopH allowed the identification of a conserved water molecule proximal to the aminooxy group that was subsequently employed for the design of furanyl-based oxime derivatives. By this process, a potent (IC{sub 50} = 190 nM) and nonpromiscuous inhibitor was developed with good YopH selectivity relative to a panel of phosphatases. The inhibitor showed significant inhibition of intracellular Y. pestis replication at a noncytotoxic concentration. The current work presents general approaches to PTP inhibitor development that may be useful beyond YopH.

59 BASIC BIOLOGICAL SCIENCES↗

Predictive Tools for Customizing Heat Treatment of Additively Manufactured Aerospace Components

Laser-bed powder fusion (LBPF) additive manufacturing is increasingly being used to produce components of complex geometries using the Ni-base superalloy Inconel 718. The composition and the microstructure of the alloy are currently well optimized for wrought components made using conventional manufacturing processes such as rolling, forging, extrusion, etc. The attractive mechanical properties of the alloy result from the underlying austenitic matrix with fine equiaxed grains, and a high density and uniform distribution of the precipitation hardening phase, γ". Heat treatment steps such as homogenization, solutioning and aging are well documented for the wrought alloy. However, when the same wrought alloy compositions are used for the additive manufacturing (AM) processes, the asprocessed microstructure is significantly different, because of the different thermal history associated with LBPF, including rapid solidification and multiple temperature excursions that lead to multiple re-melting and reheating in the solid state. Rapid solidification introduces potential non-equilibrium effects at the moving solid-liquid interfaces that impact the extent of solute segregation, as well as the morphology of the dendritic grains that form. In order to recover the target mechanical properties, AM components have to undergo post-process heat treatments. However, such heat treatments have to be custom designed for the AM process and the component geometry because of the expected vast differences in the microstructure at various locations of a component with complex geometry. The homogenization and precipitation steps should be optimized for the component so that target mechanical properties can be obtained throughout the part. The objective of this research is to utilize High Performance Computing in phase field simulations of microstructure evolution during post-processing of AM components. The physics-based modeling will be beneficial in reducing the experimental effort required for heat treatment process selection, optimization, and certification, thus leading to a significant reduction in energy consumption for AM and post-processing heat treatment. The optimization study will help identify heat treatments steps that are critical for development of a final desired microstructure with the minimum energy input. This combined with shortening of the production cycle (time-to-market) by reducing the number of failed parts (property targets), and reduction in the number of iterations for process optimization, will enable 30-40% savings in the energy costs. Phase field simulations of the degree of homogenization and the effect of local matrix composition on the nucleation and growth of competing precipitating phases were performed using the Microstructure Evolution Using Massively Parallel Phase Field Simulations code developed in-house at the Oak Ridge National Laboratory. The simulations were able to successfully capture the kinetics of nucleation and growth, and morphologies of various precipitating phases as a function of local matrix compositions and composition gradients characteristic of local microstructures arising from location-dependent variations in the thermal conditions. Future work will involve extending the simulations to a length scale consisting of multiple dendrites, so that the effect of homogenization on the coarsening of the dendrites can be simulated and used as an additional input to the optimization of the heat treatment process.

36 MATERIALS SCIENCE↗

SYNTHESIS AND CHARACTERIZATION OF PYRIDINE END-FUNCTIONALIZED POLY (3-HEXYLTHIOPHENE)s: POTENTIAL LIGANDS FOR SQD/P3HT BHJs

Extensive research has been done to optimize donor-acceptor properties of bulk heterojunction organic photovoltaics (BHJ OPVs) leading to devices with relatively high power conversion efficiencies (>5%). Polymer-nanoparticle hybrid solar cells are one class of BHJ OPVs that incorporate nanoparticles (electron carrier) in a polymer matrix (hole carrier). Blending inorganic nanoparticles with an organic polymer matrix is challenging due to unfavorable interactions between the nanoparticle and polymer leading to aggregation and poor charge separation. One method to improve ordering of the BHJ is by use of ligands on the inorganic nanoparticles that facilitate homogenous dispersion within the polymer matrix. This paper will discuss the preparation of pyridine end-functionalized P3HTs by in situ quenching of the Grignard Metathesis Polymerization with functional Grignard reagents. The degree of functionality of the polymers is determined by MALDI-TOF MS. In addition, methods to optimize the functionality of the resulting polymer will also be discussed.

14 SOLAR ENERGY↗

Treatment of Magnesium-Thorium Alloy Used in Military Defence Equipment by Separation of Thorium and/or Thermal Oxidation - 17009

Magnesium-alloys doped with thorium (Th, 3% by weight and with a total activity content in the alloy of 120 Bq/g) to improve mechanical properties has been used in, among other applications, military equipment such as in the jet-engines in military aircraft. The material cannot, due to the presence of the radioactive Th, be handled as an ordinary scrap metal. It needs to be treated to render a more stable waste form prior to final disposal in which the amount of disposed metallic magnesium is limited due to the pyrophoric properties of such type of materials. By thermally treating the material it is possible to transform the alloy into oxide form (MgO and ThO{sub 2}) which is a much more stable form than metallic. This itself enables a form suitable for final disposal of the material. The oxidation step does however lead to an increase in volume and is therefore not an optimal treatment method. It is possible to treat the metal oxide powder in order to get a better volume reduction such as by super compaction but a better result can be obtained by separating the Th from the alloy, leaving a much smaller volume for final disposal. One way of separating the Th from the metal is by dissolution of metallic material in acid followed by directed precipitation of Th as Th-oxalate. Several parameters such as the choice of acid, metal concentration in the acid solution, pH prior to precipitation and amount of added oxalic acid have been investigated. Another method to separate Th from Mg is to perform a selective dissolution of MgO from an MgO/ThO{sub 2} mixture with HNO{sub 3} leaving only ThO{sub 2} as a residue. These tests had to be preceded with an oxidation of the alloy. The tests show that it is possible to separate Th from the metal alloy by either dissolution followed by precipitation or selective dissolution of MgO from an MgO/ThO{sub 2} mixture. This leads to very effective volume reduction of the waste. For the direct dissolution and precipitation tests typically > 95% of the magnesium can be found in the acid solution while 99.5 % of the Th is found as a solid in the filters. For the oxide dissolution tests typically 98% of the Th can be separated from the bulk material, leaving the acid solution possible to free release and dispose of in the same way as for the previous method. (authors)

36 MATERIALS SCIENCE↗

Efficient energy transfer in light-harvesting systems: Quantum-classical comparison, flux network, and robustness analysis

Following the calculation of optimal energy transfer in thermal environment in our first paper [J. L. Wu, F. Liu, Y. Shen, J. S. Cao, and R. J. Silbey, New J. Phys. 12, 105012 (2010)], full quantum dynamics and leading-order 'classical' hopping kinetics are compared in the seven-site Fenna-Matthews-Olson (FMO) protein complex. The difference between these two dynamic descriptions is due to higher-order quantum corrections. Two thermal bath models, classical white noise (the Haken-Strobl-Reineker (HSR) model) and quantum Debye model, are considered. In the seven-site FMO model, we observe that higher-order corrections lead to negligible changes in the trapping time or in energy transfer efficiency around the optimal and physiological conditions (2% in the HSR model and 0.1% in the quantum Debye model for the initial site at BChl 1). However, using the concept of integrated flux, we can identify significant differences in branching probabilities of the energy transfer network between hopping kinetics and quantum dynamics (26% in the HSR model and 32% in the quantum Debye model for the initial site at BChl 1). This observation indicates that the quantum coherence can significantly change the distribution of energy transfer pathways in the flux network with the efficiency nearly the same. The quantum-classical comparison of the average trapping time with the removal of the bottleneck site, BChl 4, demonstrates the robustness of the efficient energy transfer by the mechanism of multi-site quantum coherence. To reconcile with the latest eight-site FMO model which is also investigated in the third paper [J. Moix, J. L. Wu, P. F. Huo, D. F. Coker, and J. S. Cao, J. Phys. Chem. Lett. 2, 3045 (2011)], the quantum-classical comparison with the flux network analysis is summarized in Appendix C. The eight-site FMO model yields similar trapping time and network structure as the seven-site FMO model but leads to a more disperse distribution of energy transfer pathways.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Energy Saving Melting and Revert Reduction Technology: Improved Die Casting Process to Preserve the Life of the Inserts

The goal of this project was to study the combined effects of die design, proper internal cooling and efficient die lubricants on die life. The project targeted improvements in die casting insert life by: Optimized Die Design for Reduced Surface Temperature: The life of die casting dies is significantly shorter when the die is exposed to elevated temperature for significant periods of time. Any die operated under conditions leading to surface temperature in excess of 1050oF undergoes structural changes that reduce its strength. Optimized die design can improve die life significantly. This improvement can be accomplished by means of cooling lines, baffles and bubblers in the die. A key objective of the project was to establish criteria for the minimal distance of the cooling lines from the surface. This effort was supported with alloys and machining by BohlerUddeholm, Dunn Steel, HH Stark and Rex Buckeye. In plant testing and evaluation was conducted as in-kind cost share at St. Clair Die Casting. The Uddeholm Dievar steel evaluated in this program showed superior resistance to thermal fatigue resistance. Based on the experimental evidence, cooling lines could be placed as close as 0.5" from the surface. Die Life Extension by Optimized Die Lubrication: The life of die casting dies is affected by additions made to its surface with the proper lubricants. These lubricants will protect the surface from the considerable temperature peaks that occur when the molten melt enters the die. Dies will reach a significantly higher temperature without this lubricant being applied. The amount and type of the lubricant are critical variables in the die casting process. However, these lubricants must not corrode the die surface. This effort was supported with alloys and machining by BohlerUddeholm, Dunn Steel, HH Stark and Rex Buckeye. In plant testing and evaluation was conducted as in-kind cost share at St. Clair Die Casting. Chem- Trend participated in the program with die lubricants and technical support. Experiments conducted with these lubricants demonstrated good protection of the substrate steel. Graphite and boron nitride used as benchmarks are capable of completely eliminating soldering and washout. However, because of cost and environmental considerations these materials are not widely used in industry. The best water-based die lubricants evaluated in this program were capable of providing similar protection from soldering and washout. In addition to improved part quality and higher production rates, improving die casting processes to preserve the life of the inserts will result in energy savings and a reduction in environmental wastes. Improving die life by means of optimized cooling line placement, baffles and bubblers in the die will allow for reduced die temperatures during processing, saving energy associated with production. The utilization of optimized die lubricants will also reduce heat requirements in addition to reducing waste associated with soldering and washout. This new technology was predicted to result in an average energy savings of 1.1 trillion BTU's/year over a 10 year period. Current (2012) annual energy saving estimates, based on commercial introduction in 2010, a market penetration of 70% by 2020 is 1.26 trillion BTU's/year. Along with these energy savings, reduction of scrap and improvement in casting yield will result in a reduction of the environmental emissions associated with the melting and pouring of the metal which will be saved as a result of this technology. The average annual estimate of CO2 reduction per year through 2020 is 0.025 Million Metric Tons of Carbon Equivalent (MM TCE).

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

Computational Research of Electronic and Magnetic Properties of Nonmetal Doping of Graphene-Like ZnO Monolayer

In this paper, the geometrical, electronic, and magnetic properties of nonmetal (B, C, N, F, Si, P, or S) atom doping of g-ZnO monolayer supercell have been comprehensively calculated using the first-principle method. The structural optimization implies that nonmetal atom doping of g-ZnO monolayer leads to the structural distortion around the doping atoms compared with pure g-ZnO sheet. We find that g-ZnO monolayer with one O atom per supercell substituted by one B or C atom is ferromagnetic half metal, and that substituted by one N, Si, or P atom exhibits a ferromagnetic semiconductor feature, while that substituted by one F or S atom appears to be nonmagnetic semiconductor. The total magnetic moments mainly originates from the spin polarization of the doping atom (B, C, N, Si, and P), and the rest comes from the nearest Zn and O atoms.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

Long-Term Stewardship Optimization for a Decommissioned Nuclear Reactor - 19579

The decommissioned Hallam Nuclear Power Facility (Hallam), located in Lancaster County, Nebraska, approximately 19 miles south of Lincoln, was developed as a prototype to investigate the technical and economic feasibility of a sodium-cooled, graphite-moderated nuclear reactor. As part of its Power Demonstration Program, the US Atomic Energy Commission (US AEC), a predecessor agency of the US Department of Energy (US DOE), operated the Hallam reactor from 1962 to 1964. The radiological surveillance of the Hallam site was established when the operating agreement between AEC and the Nebraska Public Power District was terminated. The US DOE Chicago Operations Office and the Nebraska Department of Health (NDH) agreed to install a shallow groundwater monitoring system as part of the environmental surveillance program. Surveillance at the site was scheduled to end in fiscal year 2005, with no further activities planned after that. At that time, the plan was to transition the site to the Nebraska Public Power District with future use of the facility to remain restricted. US DOE has title to and responsibility for the entombed radioactive materials. A surveillance and monitoring program was initiated in 1970 by NDH and was funded by US AEC. It included the analysis of samples from deep production wells (groundwater from the regional aquifer at depths greater than 55 meters [180 feet]) at the Sheldon Station. NDH was concerned in 1990 about the possibility of shallow groundwater coming in contact with the buried radiological materials along the buried walls of the reactor. Subsequently, DOE and NDH agreed to further characterize hydrologic conditions and establish a monitoring program in the shallow, perched groundwater zones. The current monitoring program focuses on the shallow, perched groundwater. Today, decades since facility decommissioning, site inspections and groundwater monitoring continue at this Cold War-era legacy site. The US DOE Office of Legacy Management (LM) monitors groundwater as a best management practice in response to a request from the Nebraska Department of Health and Human Services (DHHS). In 2006, LM recommended discontinuing groundwater monitoring because analytical results since 1970 demonstrated there had been no impact to shallow, perched groundwater and no current or anticipated unacceptable risk to human health and the environment. The State of Nebraska did not concur with the recommendation to stop monitoring but did agree to a reduction in sampling and analysis from once a year to once every 2 years--a schedule that began in 2008. An updated assessment of the groundwater monitoring effort was issued in 2016. The entire body of water-quality data continues to demonstrate that there have been no negative impacts to the shallow, perched groundwater from radioactive materials entombed at the site 46 years ago. LM, with concurrence from Nebraska DHHS, changed the sampling frequency, effective in 2018, as follows: once every 5 years from 2021 to 2041 and once every 10 years from 2041 to 2071. In summary, LM remains committed to its primary mission to protect human health and the environment at each of its 92 legacy sites. With results from groundwater sampling events demonstrating compliance for 48 years, the Hallam site was able to set an LM benchmark to optimize long-term stewardship and define an end-state strategy for other decommissioned reactor facilities across the DOE complex. Continued collaboration with local communities, stakeholder, and regulators often leads to outcomes of sound protectiveness without compromise to safety or compliance, as demonstrated by the optimization of the long-term stewardship at the Hallam site. (authors)

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

Balancing aggregation and smoothing errors in inverse models

Inverse models use observations of a system (observation vector) to quantify the variables driving that system (state vector) by statistical optimization. When the observation vector is large, such as with satellite data, selecting a suitable dimension for the state vector is a challenge. A state vector that is too large cannot be effectively constrained by the observations, leading to smoothing error. However, reducing the dimension of the state vector leads to aggregation error as prior relationships between state vector elements are imposed rather than optimized. Here we present a method for quantifying aggregation and smoothing errors as a function of state vector dimension, so that a suitable dimension can be selected by minimizing the combined error. Reducing the state vector within the aggregation error constraints can have the added advantage of enabling analytical solution to the inverse problem with full error characterization. We compare three methods for reducing the dimension of the state vector from its native resolution: (1) merging adjacent elements (grid coarsening), (2) clustering with principal component analysis (PCA), and (3) applying a Gaussian mixture model (GMM) with Gaussian pdfs as state vector elements on which the native-resolution state vector elements are projected using radial basis functions (RBFs). The GMM method leads to somewhat lower aggregation error than the other methods, but more importantly it retains resolution of major local features in the state vector while smoothing weak and broad features.

Turner, A. J.↗

Improvement of glass-forming ability and phase separation in Cu Ti-rich

Present study reports improvement of glass-forming ability (GFA) and phase separation in Cu Ti-rich Cu Ti Zr Ni Si bulk metallic glasses (BMGs) by tailoring the constituent elements. The MA of metalloid element, Sn having relatively large negative enthalpy of mixing can lead to improve GFA (up to 8mm in diameter) as well as thermal stability (up toTx = 48K) by optimizing the substitution element. And the addition of elements having relatively large positive enthalpy of mixing (partial substitution of Zr or Ti with Y) can lead to the liquid state phase separation in Cu Ti Sn Zr Ni Si BMG, although the addition lead to drastic deterioration of the GFA.

Park, E S↗

A convex pseudolikelihood framework for high dimensional partial correlation estimation with convergence guarantees

Sparse high dimensional graphical model selection is a topic of much interest in modern day statistics. A popular approach is to apply l1-penalties to either parametric likelihoods, or regularized regression/pseudolikelihoods, with the latter having the distinct advantage that they do not explicitly assume Gaussianity. As none of the popular methods proposed for solving pseudolikelihood-based objective functions have provable convergence guarantees, it is not clear whether corresponding estimators exist or are even computable, or if they actually yield correct partial correlation graphs. We propose a new pseudolikelihood-based graphical model selection method that aims to overcome some of the shortcomings of current methods, but at the same time retain all their respective strengths. In particular, we introduce a novel framework that leads to a convex formulation of the partial covariance regression graph problem, resulting in an objective function comprised of quadratic forms. The objective is then optimized via a coordinate wise approach. The specific functional form of the objective function facilitates rigorous convergence analysis leading to convergence guarantees; an important property that cannot be established by using standard results, when the dimension is larger than the sample size, as is often the case in high dimensional applications. These convergence guarantees ensure that estimators are well defined under very general conditions and are always computable. In addition, the approach yields estimators that have good large sample properties and also respect symmetry. Furthermore, application to simulated and real data, timing comparisons and numerical convergence is demonstrated. We also present a novel unifying framework that places all graphical pseudolikelihood methods as special cases of a more general formulation, leading to important insights.

97 MATHEMATICS AND COMPUTING↗