Search NASA⌕ Search

SEARCH · Search NASA

Results for “Match Machining”

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 217 records · Page 12

Instrumentation and Beam Dynamics Study of Advanced Electron-Photon Facility in Indiana University

The Advanced eLectron-PHoton fAcility (ALPHA) is a compact electron accelerator under construction and being commissioned at the Indiana University Center for Exploration of Energy and Matter (CEEM). In this thesis, we have studied the refurbished Cooler Injector Synchrotron (CIS) RF cavity using both the transmission line model and SUPERFISH simulation. Both low power and high power RF measurements have been carried out to characterize the cavity. Considering the performance limit of ferrite, we have designed a new ferrite loaded, co-axial quarter wave like cavity with similar structure but a more suitable ferrite material. We have also designed a traveling wave stripline kicker for fast extraction by POISSON and Microwave Studio. The strips geometry is trimmed to maximize the uniformity of the kicking field and match the impedance of the power cables. The time response simulation shows the kicker is fast enough for machine operation. The pulsed power supply requirement has also been specified. For the beam diagnosis in the longitudinal direction, we use a wideband Wall Gap Monitor (WGM) served in CIS. With proper shielding and amplification to get good WGM signal, we have characterized the injected and extracted beam signal in single pass commissioning, and also verified the debunching effect of the ALPHA storage ring. A modulation-demodulation signal processing method is developed to measure the current and longitudinal profile of injected beam. By scanning the dipole strength in the injection line, we have reconstructed the tomography of the longitudinal phase space of the LINAC beam. In the accumulation mode, ALPHA will be operated under a low energy and high current condition, where intra beam scattering (IBS) becomes a dominant effect on the beam emittance. A self consistent simulation, including IBS effect, gas scattering and linear coupling, has been carried out to calculate the emittance of the stored beam.

43 PARTICLE ACCELERATORS↗

Ka-Band Waveguide Three-Way Serial Combiner for MMIC Amplifiers

In this innovation, the three-way combiner consists internally of two branch-line hybrids that are connected in series by a short length of waveguide. Each branch-line hybrid is designed to combine input signals that are in phase with an amplitude ratio of two. The combiner is constructed in an E-plane split-block arrangement and is precision machined from blocks of aluminum with standard WR-28 waveguide ports. The port impedances of the combiner are matched to that of a standard WR-28 waveguide. The component parts include the power combiner and the MMIC (monolithic microwave integrated circuit) power amplifiers (PAs). The three-way series power combiner is a six-port device. For basic operation, power that enters ports 3, 5, and 6 is combined in phase and appears at port 1. Ports 2 and 4 are isolated ports. The application of the three-way combiner for combining three PAs with unequal output powers was demonstrated. NASA requires narrow-band solid-state power amplifiers (SSPAs) at Ka-band frequencies with output power in the range of 3 to 5 W for radio or gravity science experiments. In addition, NASA also requires wideband, high-efficiency SSPAs at Ka-band frequencies with output power in the range of 5 to 15 W for high-data-rate communications from deep space to Earth. The three-way power combiner is designed to operate over the frequency band of 31.8 to 32.3 GHz, which is NASA s deep-space frequency band.

Wintucky, Edwin G.↗

Distributed-memory Parallel Algorithms for Matching and Coloring

Graph matching and coloring constitute two fundamental classes of combinatorial problems having numerous established as well as emerging applications in computational science and engineering, high-performance computing, and informatics. We provide a snapshot of an on-going work on the design and implementation of new highly-scalable distributed-memory parallel algorithms for two prototypical problems from these classes, edge-weighted matching and distance-1 vertex coloring. Graph algorithms in general have low concurrency and poor data locality, making it challenging to achieve scalability on massively parallel machines. We overcome this challenge by employing a variety of techniques, including approximation, speculation and iteration, optimized communication, and randomization, in concert. We present preliminary results on weak and strong scalability studies conducted on an IBM Blue Gene/P machine employing up to tens of thousands of processors. The results show that the algorithms hold strong potential for computing at petascale.

Combinatorial scientific computing↗

Towards Precision Photometric Type ia Supernova Cosmology With Machine Learning

The revolutionary discovery of dark energy and accelerating cosmic expansion was made with just 42 type Ia supernovae (SNe Ia) in 1999. Since then, large synoptic surveys, e.g., Dark Energy Survey (DES), have observed thousands more SNe Ia and the upcoming Rubin Legacy Survey of Space and Time (LSST) and Roman Space Telescope promise to deliver millions in the next decade. This unprecedented data volume can produce the required precision to unambiguously test concordance cosmology which could represent a monumental shift in our understanding of dark energy and its role in cosmic history. However, extracting a pure SN Ia sample with accurate redshifts for such a large dataset will be a challenge. Specifically, spectroscopic classification will not be possible for the vast majority of discovered objects, and only ~25% will have spectroscopic redshifts. This thesis presents a series of observational and methodological studies designed to address the questions associated with this new era of photometric SN Ia cosmology. First, we present a machine learning method for photometric classification of SNe, Supernova Classification with a COnvolutional Neural Network (SCONE). Photometric classification enables SNe with no spectroscopic information to be confidently categorized, and is a critical component of current and future analysis pipelines. SCONE achieves >99% accuracy distinguishing simulated SNe Ia from non-Ia SNe, and has been integrated into DES, LSST, and Roman analysis pipelines. We also demonstrate the efficacy of SCONE on early-time photometric classification, which will be vital for optimal allocation of spectroscopic resources. We show that SCONE can distinguish between 6 SN types with 75% accuracy on the night of initial discovery, comparable to results in the literature for full-phase SNe. Next, we study current methods for estimating SN Ia redshifts and propose a machine learning alternative that uses SN photometry alone to extract redshift information. Most SNe Ia inherit redshift information from their host galaxy, but the process of matching SNe to the correct host galaxy can be challenging. We systematically analyze the impact of incorrect redshifts from host galaxy mismatch on 5 years of DES SN data, and conclude that improved host matching or redshift estimation methods can reduce our systematic errors by ~10%. In response to this finding, we present a SN photometry-only method for estimating redshifts independent of host galaxy information, Photo-zSNthesis. We show that Photo-zSNthesis redshift estimates are accurate to within 2% across the full redshift range of LSST, a first in the literature. Finally, we focus on the robustness of machine learning (ML) algorithms for real-world and scientific applications. ML models generalize poorly beyond their training set and often experience severe performance degradation when deployed on new data. We demonstrate a general method for improving robustness that achieves new state-of-the-art results on astronomical object classification, wildlife identification, and tumor detection.

79 ASTRONOMY AND ASTROPHYSICS↗

Microstructure Generation via Generative Adversarial Network for Heterogeneous, Topologically Complex 3D Materials

Using a large-scale, experimentally captured 3D microstructure dataset, we implement the generative adversarial network (GAN) framework to learn and generate 3D microstructures of solid oxide fuel cell electrodes. The generated microstructures are visually, statistically, and topologically realistic, with distributions of microstructural parameters, including volume fraction, particle size, surface area, tortuosity, and triple phase boundary density, being highly similar to those of the original microstructure. These results are compared and contrasted with those from an established, grain-based generation algorithm (DREAM.3D). Importantly, simulations of electrochemical performance, using a locally resolved finite element model, demonstrate that the GAN generated microstructures closely match the performance distribution of the original, while DREAM.3D leads to significant differences. Finally, the ability of the generative machine learning model to recreate microstructures with high fidelity suggests that the essence of complex microstructures may be captured and represented in a compact and manipulatable form.

36 MATERIALS SCIENCE↗

LHC hadronic jet generation using convolutional variational autoencoders with normalizing flows

Abstract In high energy physics, one of the most important processes for collider data analysis is the comparison of collected and simulated data. Nowadays the state-of-the-art for data generation is in the form of Monte Carlo (MC) generators. However, because of the upcoming high-luminosity upgrade of the Large Hadron Collider (LHC), there will not be enough computational power or time to match the amount of needed simulated data using MC methods. An alternative approach under study is the usage of machine learning generative methods to fulfill that task. Since the most common final-state objects of high-energy proton collisions are hadronic jets, which are collections of particles collimated in a given region of space, this work aims to develop a convolutional variational autoencoder (ConVAE) for the generation of particle-based LHC hadronic jets. Given the ConVAE’s limitations, a normalizing flow (NF) network is coupled to it in a two-step training process, which shows improvements on the results for the generated jets. The ConVAE+NF network is capable of generating a jet in 18.30 ± 0.04 μ s , making it one of the fastest methods for this task up to now.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

First 3 years of operation of RIACS (Research Institute for Advanced Computer Science) (1983-1985)

The focus of the Research Institute for Advanced Computer Science (RIACS) is to explore matches between advanced computing architectures and the processes of scientific research. An architecture evaluation of the MIT static dataflow machine, specification of a graphical language for expressing distributed computations, and specification of an expert system for aiding in grid generation for two-dimensional flow problems was initiated. Research projects for 1984 and 1985 are summarized.

Denning, P. J.↗

Modeling and Mitigation of dv/dt and Transmission Line Effect for Motor Drive System in Electric Aircraft Propulsion

Due to constraints in space and structure as well as other practical limitations in aircraft system, the motor controller or electric motor are often installed far away from each other and interconnected via long cables. This amplifies the transmission line effect excited by the dv/dt PWM voltage and induces high voltage stress on the machine winding. This paper investigates the state-of-art modeling approaches for such systems, identifies several limitations in matching the tested voltage reflection waveform in terms of peak voltage, oscillation frequency and damping speed, and offers several key considerations not well covered in literature or in engineering practice. It then proposes a practical modeling methodology which, as validated by experimental results, truly captures these critical reflection voltage specs. To mitigate the high peak voltage and strong oscillation, dv/dt filter is applied and a parametric design optimization based on widely adopted LC-R topology is illustrated, which not only alleviates the voltage stress but also minimized the power loss of the passive damping resistor.

Inductor↗

In Situ Machine Learning for Intelligent Data Capture on Exascale Platforms. Final Report

In many dynamic systems, interesting events occur locally in time and space. Examples of such systems include ignition events in combustion simulations, material fractures in mechanics simulations, and extreme weather events in climate simulations. Due to memory constraints and data I/O costs, current simulation workflows save data at regularly spaced time-steps, at a fixed rate determined before the start of the simulation. Often this mode of operation results in missed events of interest, necessitating a simulation restart from before an event occurred with more frequent data saves. This data saving workflow is grossly inefficient and is already a bottleneck in the computing process. We propose to develop machine learning algorithms that can detect when interesting dynamical events are occurring, triggering data saves. These machine learning algorithms will perform in situ anomaly detection to flag regions with different dynamical properties than those previously recorded. The adaptive data saves would be local in time and space to match the event of interest, thereby enabling a much more efficient workflow that will reduce data I/O costs and data storage memory requirements. The algorithms will be tested on two applications: auto-ignition simulations and climate simulations. A critical component of this project will be developing machine learning algorithms that can be deployed efficiently in situ on HPC platforms with out-of-the-box functionality. The development of in situ machine learning methods to detect anomalous events would enable a more efficient and effective workflow, in which all the relevant data are saved in a single simulation run, without re-starts or scientist intervention.

42 ENGINEERING↗

Dynamic Ride-Matching for Large-Scale Transportation Systems

Efficient dynamic ride-matching (DRM) in large-scale transportation systems is a key driver in transport simulations to yield answers to challenging problems. Although the DRM problem is simple to solve, it quickly becomes a computationally challenging problem in large-scale transportation system simulations. Therefore, this study thoroughly examines the DRM problem dynamics and proposes an optimization-based solution framework to solve the problem efficiently. To benefit from parallel computing and reduce computational times, the problem’s network is divided into clusters utilizing a commonly used unsupervised machine learning algorithm along with a linear programming model. Then, these sub-problems are solved using another linear program to finalize the ride-matching. At the clustering level, the framework allows users adjusting cluster sizes to balance the trade-off between the computational time savings and the solution quality deviation. A case study in the Chicago Metropolitan Area, U.S., illustrates that the framework can reduce the average computational time by 58% at the cost of increasing the average pick up time by 26% compared with a system optimum, that is, non-clustered, approach. Another case study in a relatively small city, Bloomington, Illinois, U.S., shows that the framework provides quite similar results to the system-optimum approach in approximately 62% less computational time.

33 ADVANCED PROPULSION SYSTEMS↗

Pre-trained network-based transfer learning: A small-sample machine learning approach to nuclear power plant classification problem

Some research topics belonging to classification problems in the nuclear industry, such as fault diagnosis and accident identification, can be solved by feature extraction and subsequent application of statistical machine learning classifiers. Recently, deep neural network-based methods with automatic feature extraction and high accuracy have gained wide attention. They usually require large-scale training data, however, plant fault or accident data are scarce or difficult to obtain. Here this paper proposes a convolutional network (CNN)-based transfer learning method to solve this problem. The network's shallow layer is derived from a pre-trained CNN based on the ImageNet database to automatically extract features, and the deep layer is customized to match the classification problem. Data in non-image formats are converted to image formats and subsequently used to train the network. Case studies of rotating machines fault diagnosis show that the proposed method requires only limited training data to achieve high accuracy.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗

Brazil disc Testing of LX-17-1

When considering alternate test methods for assessing the tensile performance of insensitive plastic bonded explosive (PBX) materials, Brazil disc tensile testing is attractive from several standpoints. First, compared to the traditional dog-bone shaped specimen normally used in the direct tensile test, the Brazil test specimen is considerably easier to produce. Second, the Brazil specimen is compact, with a low aspect ratio that allows it to be machined from a wide variety of pressed geometries. Assuming test results are consistent and interpretable, the method could open up possibilities for rapid assessment of ultimate tensile strengths, and might be especially attractive when a large number of samples are to be tested. The trade-off for specimen simplicity is, however, that interpretation of the results from Brazil tests is inherently far less intuitive, and more difficult, than is the case with traditional tensile testing, since the method involves complex stress loading. As a result, the derivation of what one hopes are equivalent tensile failure stresses from Brazil testing necessitates some fairly involved calculations, requiring parameters that are not always readily obtained. In addition to some significant uncertainties, the specimen geometry does not lend itself easily to the accommodation of traditional strain measurement devices such as extensometers. As a consequence, optical techniques such as digital image correlation (DIC) are preferred when stress-strain curves are needed. In this test series we performed a sequence of experiments using Brazil discs made from LX-17-1. Tests were performed on samples at two different scaled sizes: discs that were 0.5 inches in diameter and 0.25 inch thick, and discs that were 1 inch in diameter and a 0.5 inch thick. Each of the specimen sets was tested using matching, equivalently scaled test fixtures, each with anvils whose radii were 0.25% greater than their respective specimen radii. All specimens tested were machined from billet pressed material made from Holston #851009 (LLNL tag #C-329) LX-17-1 molding powder. Tests were performed at two temperatures, +23 °C and +50 °C, with three each test redundancies at each condition. Based on data from this set of experiments we hoped to provide preliminary answers to a number of questions, including: 1) Roughly, what is the level of scatter in the data? 2) Are results scale independent, at least within the limited range of specimen sizes tested? 3) Using a consistent set of parameters, can we use the equations given by Sato and Awaji1 to calculate peak tensile stresses that are nearly equivalent to those measured using traditional dog-bone tensile testing methods.

45 MILITARY TECHNOLOGY, WEAPONRY, AND NATIONAL DEF↗

Constraints on the Physical Properties of GW190814 through Simulations Based on DECam Follow-up Observations by the Dark Energy Survey

On 2019 August 14, the LIGO and Virgo Collaborations detected gravitational waves from a black hole and a 2.6 solar mass compact object, possibly the first neutron star–black hole merger. In search of an optical counterpart, the Dark Energy Survey (DES) obtained deep imaging of the entire 90% confidence level localization area with Blanco/DECam 0, 1, 2, 3, 6, and 16 nights after the merger. Objects with varying brightness were detected by the DES Pipeline, and we systematically reduced the candidate counterparts through catalog matching, light-curve properties, host-galaxy photometric redshifts, Southern Astrophysical Research spectroscopic follow-up observations, and machine-learning-based photometric classification. All candidates were rejected as counterparts to the merger. To quantify the sensitivity of our search, we applied our selection criteria to full light-curve simulations of supernovae and kilonovae as they would appear in the DECam observations. Because the source class of the merger was uncertain, we utilized an agnostic, three-component kilonova model based on tidally disrupted neutron star (NS) ejecta properties to quantify our detection efficiency of a counterpart if the merger included an NS. We find that, if a kilonova occurred during this merger, configurations where the ejected matter is greater than 0.07 solar masses, has lanthanide abundance less than 10 -8.56 , and has a velocity between 0.18c and 0.21c are disfavored at the 2σ level. Furthermore, we estimate that our background reduction methods are capable of associating gravitational wave signals with a detected electromagnetic counterpart at the 4σ level in 95% of future follow-up observations.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

Design of a Pulsed Flux Concentrator for the ILC Positron Source

The Positron Source for the International Linear Collider requires an optical matching device after the target to increase the capture efficiency for positrons. Pulsed flux concentrators have been used by previous machines to improve the capture efficiency but the ILC has a 1 ms long pulse train which is too long for a standard flux concentrator. A pulsed flux concentrator with a 40 ms flat top was created for a hyperon experiment in 1965 which used liquid nitrogen cooling to reduce the resistance of the concentrating plates and extend the lifetime of the pulse. We report on a design for a 1 ms device based on this concept.

43 PARTICLE ACCELERATORS↗

Using Computer Simulations to Optimize Biofuel Production

The DOE strives to ensure America's security and prosperity by addressing energy challenges. NREL shares this goal and tries to achieve a clean energy world. Fossil fuels are problematic for both organizations. Using them endangers American security. Their supply is finite and burning them causes environmental damage. Biofuels are a good alternative to fossil fuels. They are renewably produced on American soil and can lower greenhouse gas emissions. Also, cars and planes need no costly mechanical adjustments to use biofuels. However, the fuels themselves are expensive. For my SULI project, I reduced the cost of biofuels by optimizing the production process through computer simulations. Existing simulations were accurate but slow. One simulation takes up to eight hours, and researchers must do hundreds. My solution reduces the computing time. I treated the biomass particles in the simulation as one-dimensional. That simplified the simulation equations, making them easier for the computer to solve. Still, biomass particles are three-dimensional. The 1D assumption was wrong and produced inaccurate results. To maintain accuracy while increasing speed, I developed a method to convert 1D simulation results into usable 3D data. I adjusted the 1D simulation until the output matched the 3D results for a specific environment. I found out how much the simulation changed when the environment changed. Machine learning algorithms defined a relationship between 1D and 3D data for all environments. This lets scientists convert fast 1D simulation results into valid 3D data.

1D↗

High Tonnage Forest Biomass Production Systems from Southern Pine Energy Plantations

In this study, a high-tonnage harvesting system designed specifically to operate efficiently in the expected stand types of a bioenergy scenario was built, deployed, and evaluated in a production setting. Stands on which the system was evaluated exhibited the heavy stocking levels (> 600 stems per acre) and tree size distributions with significant volume in small stems (down to 2” DBH) that were expected in the modified energy plantation silvicultural approach. The harvest system also was designed to be functional in the traditional plantation stands dominating the commercial forestry landscape in the region. The Tigercat 845D feller buncher, which was a prototype machine designed for the high tonnage harvest system, used a boom-mounted prototype DT1802 shear felling head and incorporated a number of options intended to maximize its small-stem productivity, including: a high-speed shear severing system that was cheaper to operate than a saw; a large-pocket felling head that allowed larger accumulations of small stems to be built before expending the time to drop them for the skidder; efficient, low ground pressure, tracked carrier system to decrease the amount of maneuvering, saving time and minimizing soil disturbance; and various energy-saving devices to lower fuel costs and minimize air quality impacts. Overall, the feller buncher represented a quantum advance in small-stem harvesting technology. Extensive testing showed the machine’s production rate to be relatively insensitive to piece size, much less so than comparable traditional equipment. In plantation stands, the feller buncher was able to produce approximately 100 green tons of biomass per productive machine hour (PMH), and in natural stands, it produced nearly 120 green tons per PMH. The ability of the high tonnage feller buncher to maintain high productivity in stands with smaller diameter stems is something that has not been achieved in previous feller buncher designs. The Tigercat 845D feller buncher is now a production machine for Tigercat and is being sold in their current product line. The high-speed felling system was paired with a Tigercat 630D skidder and high-capacity grapple; one that could match the felling productivity when pulling small stems. The harvesting system minimized hourly costs using a single, high-capacity skidder (with a single operator), rather than two smaller ones, which is the traditional practice. The skidder itself can be considered a mid-range size and had an engine no larger than other machines in its class, but it incorporated a very large capacity 25 ft2 grapple. The large grapple is well suited to grabbing and hauling a large bunch of small-diameter trees, as produced by the high tonnage feller buncher. The grapple worked effectively in larger stems as well, but its ability to carry large numbers of small stems meant the average payload did not drop as stand DBH decreased. Tests with the machine indicated its travel speeds were nearly the same as, or perhaps slightly better than, conventionally equipped skidders, but grapple capacity was 75% larger. Productivity and cost per ton of the new skidder were better than conventional skidders for average skid distances of any length greater than 100 feet. Measured skidder productivity was as high as 143 gt/PMH. Its productivity exceeded that of the high-capacity feller buncher for skid distances out to nearly 700 feet, so system productivity could be expected to remain high for stands of a size typical in the southern U.S. The Tigercat 630D skidder is a production machine for Tigercat and the large grapple can now be ordered by customers using it for small diameter trees. When the feller buncher and skidder are analyzed as a two-machine system, overall productivity is fixed at the level of the least productive machine. Results from a set of side-by-side tests in the same density stand with conventional feller bunchers and skidders showed that the high tonnage system produced 97 gt/PMH versus 68 gt/PMH for a comparable conventional system. Machine rate costs for felling and skidding were $2.31/gt and $3.72/gt for the high tonnage, and conventional systems, respectively. However, the most significant result of the project is that the high tonnage system was shown to be relatively insensitive to tree size. This ability to maintain felling and skidding productivity and cost as tree size decreases is a breakthrough in harvesting systems for southern pine plantations. The concept of transpirational drying of woody biomass was tested at an industrial scale at multiple locations during this project. Felled trees were allowed to dry in two scenarios: 1) in bunches where they were felled, and 2) in roadside piles. Although the wood piled in large piles at roadside did experience drying, the wood left in bunches experienced a greater moisture reduction. Drying times of 72 days in the late summer resulted in mean wood moisture content of 26% for skidder bunches and 39% for the large pile at roadside as compared to moisture contents of 55% to 58% for freshly cut trees. An existing whole-tree chipper, Precision 2675, was modified to allow production of chips smaller than the traditional pulp size chip (i.e. “microchips”). Feed rates and knife placements were retained in the new design, while additional pockets were incorporated in the chipper disk to allow the attachment of either four knives for pulp chips or eight knives for microchips. This design facilitated switching between the energy and pulp chip product options at relatively low expense (about ½ day downtime). Chipping of whole-trees into pulp chips and microchips with the Precision 2675 disk chipper resulted in average productivities of 79.5 gt/PMH and 70.7 gt/PMH, respectively. Production rates of the chipper were lower when producing microchips by about 10% relative to producing pulp chips, but rates were similar to those achievable when making clean pulp chips. Particle size analysis for clean pine microchips revealed 26.6% retention on a 13 mm (slightly less than 3/8-inch) round hole screen and 25.9% retention for whole-tree pine microchips. For comparison, clean pine pulp chips had 52.2% retained. Ash content (% dry basis) was 0.54% for clean pine microchips and 0.62% for whole-tree pine microchips. Ash content for clean pine pulp chips was 0.39%. For transpirationally-dried material there was 38.1% retention for whole-tree microchips on a 13mm screen compared to 70.1% for dried clean pulp chips. Ash content was 0.78% and 0.44% respectively for these two chip types. Clean pine microchips stored at roadside had 25.2% retention on a 13 mm screen and 0.50% ash content. For mixed species (pine and hardwood), whole-tree microchips had 25.1% retention on a 13 mm screen compared to 50.6% for whole-tree pulp chips. Ash content was 2.12% and 2.74% respectively for these two chip types. Clean hardwood microchips stored at roadside had 35.0% retention on a 13 mm screen and an ash content of 1.24%. There are two significant advantages to using transpirational drying: reduced transportation costs, and reduced drying costs (capital and operating costs) for the biorefinery. This project evaluated the potential to reduce transportation costs through transpirational drying, and it included a component that tested higher capacity chip trailers (23% larger volume) to be able to transport dry wood with a lower bulk density. For transpirationally-dried chips at 35% MC, the high-capacity trailers achieved loads with a mean payload of 24 tons with maximum payloads of 29 tons. The typical legal payload on this trailer is 28.5 tons. Therefore, the project demonstrated that it is possible to achieve maximum legal payloads on chip trailers with transpirationally dried wood. Assuming that the truck is loaded to the legal payload limit, the transportation costs of chips can be reduced from $15.91/dry ton (dt) for 56% MC wood to $10.77/dt for transpirationally dried wood at 35% MC (for an example 50-mile haul distance at $0.14 per one-way ton-mile). For longer haul distances, these savings in trucking costs become even more significant. These results have demonstrated how significant savings in transportation costs can be achieved through transpirational drying. Also, these results show that it may be possible to increase the procurement radius for a biorefinery by using transpirational drying. Further cost reductions can be realized by the biorefinery when drying costs are reduced. The goal of this study was development of a timber harvesting system as productive in stands optimized for biomass production as it was in stands grown for roundwood markets. If that goal is achieved, a logger can invest in a single suite of equipment and operate efficiently in any future silvicultural regime that might include energy feedstocks as an output. It was the premise of the study that a future biomass market would shift the age distribution and stem size in stands grown for energy downward, and the key strategy in developing a harvest system for that scenario would be creating one with logging costs relatively insensitive to tree size. Our vision for such a system included a felling machine with a large capacity head to minimize time spent building bunches, plus a skidder capable of moving large volumes of small trees. The study proposed building the system and testing it against existing equipment in stands similar to those envisioned as resulting from biomass-optimized silviculture. As stated previously, the new feller buncher and skidder evaluated on their own merits showed their designs were clearly a step in the right direction - their productivity was indeed high and less sensitive to reductions in stem size. Cost projections based on extensive time and production studies of the high tonnage and benchmark operations showed modest advantages in FOB costs of the new system in both ‘average’ and simulated ‘energy’ stands (7.7% and 9.5%, respectively). But it was clear, when coupled into a traditional logging system, the in-woods productivity advantage of the modified equipment was easily overwhelmed by inefficiencies in chipping or trucking. Some additional savings can be achieved by spreading the cost of the feller buncher over multiple chipping operations (another 7.5%), but generally, in stands with average DBH above 6 inches, the in-woods equipment was not limiting productivity, and costs were driven by chipping and transport. Our results were a positive step in lowering delivered cost of trees grown for energy purposes, but they also argue strongly for a more comprehensive approach in solving this issue. The procurement system in its entirety has to be optimized to take full advantage of the productivity gains achieved with the machines and transpirational drying techniques developed in this project. We have to understand the true costs of all logistical options, particularly those of the choice in chipping strategy and in truck allocation, both of which seemed, in this study, to be the greatest source of variability in cost, and often the most expensive operations as well.

09 BIOMASS FUELS↗

Leveraging large language models to automate the identification of healthcare access barriers for veterans

Objective: To develop and evaluate an automated system for identifying healthcare barriers focusing on transportation issues in veterans’ clinical notes using large language models (LLMs) and to assess the impact of different prompting strategies on classification performance and explanation consistency. Methods: We developed a hybrid system combining pattern matching for templated notes with LLM analysis for free-text notes. Using 2000 manually annotated clinical notes, we compared four prompting strategies (dual-role short, dual-role long, analysis-first, analysis-only) across Mistral-7B and Llama-3.1 models. We evaluated classification performance using standard metrics and assessed explanation consistency through embedding similarity analysis. Results: The analysis-first strategy achieved superior performance, with Mistral-7B reaching an F1 score of 0.914, outperforming traditional machine learning approaches (GBM: 0.786, BERT: 0.811). LLMs demonstrated higher explanation consistency within models (mean cosine similarity 0.887–0.908) compared to cross-model similarities (0.767–0.872). Pattern matching successfully handled 6.7% of templated notes deterministically. Mistral-7B showed greater internal consistency but higher abstention rates compared to Llama-3.1. Conclusion: Requiring LLMs to analyze evidence before classification improves both accuracy and explanation consistency for identifying transportation barriers in clinical notes. This approach enables automated barrier detection at scale while providing clinically relevant explanations, supporting both population-level healthcare planning and individual patient care decisions.

Healthcare access barriers↗

A Hierarchical Framework for Demand-Side Frequency Control

With large-scale plans to integrate renewable generation, more resources will be needed to compensate for the uncertainty associated with intermittent generation resources. Under such conditions, performing frequency control using only supply-side resources become not only prohibitively expensive but also technically difficult. It is therefore important to explore how a sufficient proportion of the loads could assume a routine role in frequency control to maintain the stability of the system at an acceptable cost. In this paper, a novel hierarchical decentralized framework for frequency based load control is proposed. The framework involves two decision layers. The top decision layer determines the optimal droop gain required from the aggregated load response on each bus using a robust decentralized control approach. The second layer consists of a large number of devices, which switch probabilistically during contingencies so that the aggregated power change matches the desired droop amount according to the updated gains. The proposed framework is based on the classical nonlinear multi-machine power system model, and can deal with timevarying system operating conditions while respecting the physical constraints of individual devices. Realistic simulation results based on a 68-bus system are provided to demonstrate the effectiveness of the proposed strategy.

Moya, Christian↗