Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “data reduction methods”

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 1,765 records · Page 98

Pathway for decarbonizing residential building operations in the US and China beyond the mid-century

With global carbon budget targets looming, residential buildings in top economies must become carbon neutral as soon as possible to reserve more emission space for emerging carbon-emitting economies. This study is the first to compare the operational decarbonization process of China’s and the United States (US) residential buildings from 2000 to 2060 by combining the end-use emission model with the decomposing structural decomposition (DSD) method and Monte Carlo simulation. The results show that from 2001 to 2020 China decarbonized 1544 mega-tons of carbon dioxide (MtCO 2 ) and the US decarbonized 1848 MtCO 2 . In the business-as-usual scenario, China will hit its emission peak in 2031 (±3) with 934 (±61) MtCO 2 , while the US will maintain a lock-in level of 736 (±133) MtCO 2 since the 2030s. In the decarbonization scenario, operational carbon neutrality for residential buildings in 2060 is promoted by an increase in clean power generation proportion, building-integrated power generation level, building electrification level, and a reduction in end-use energy intensity, which will contribute 34.4 %, 21.4 %, 14.3 %, and 29.9 % in China and 32.9 %, 33.1 %, 8.2 %, and 25.8 % in the US, respectively. Especially, building-integrated power generation in China only costs about 40 % of what it costs in the US. Besides, high-decarbonization strategies for residential building operations are proposed as references for governments to formulate targeted climate policies. Overall, this study offers data benchmarks for buildings’ carbon neutrality of top economies to further promote synergistic carbon neutrality with the buildings of emerging economies in the age of Post COP27.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

Automated Clean Chemistry for Bulk Analysis of Environmental Swipe Samples – Final Report

Sample preparation methods for mass spectrometry are being automated using commercial-off-the-shelf (COTS) equipment to shorten lengthy and costly manual chemical purification procedures. This development addresses a serious need in the International Atomic Energy Agency’s (IAEA) Network of Analytical Laboratories (NWAL) to increase efficiency in the Bulk Analysis of Environmental Samples for Safeguards program with a method that allows unattended, overnight operation. In collaboration with Elemental Scientific Inc. (ESI), the prepFAST-MC2 was designed based on COTS equipment. It was modified for uranium/plutonium separations using renewable columns packed with Eichrom TEVA® and UTEVA® resins, with a chemical separation method based on the Oak Ridge National Laboratory (ORNL) NWAL chemical procedure. Initial verification of individual columns yielded small elution volumes with consistent elution profiles and good recovery. Combined column calibration demonstrated ample separation without crosscontamination of the eluent. Automated packing and unpacking of the built-in columns initially showed >15% deviation in resin loading by weight, which can lead to inconsistent separations. Optimization of the packing and unpacking methods led to a reduction in the variability of the packed resin to less than 5% daily. Based on the initial experience at ORNL with the prepFAST-MC2, a second-generation model, the prepFAST-SR, was built by ESI. The newly designed prepFAST-SR has several upgrades compared with the original prepFAST-MC2. The upgrades largely addressed issues with the reliability of unpacking the Eichrom columns and with ventilation to ensure adequate removal of acid vapors. Both systems are currently installed in the Ultra-Trace Forensics Science Center at ORNL. Separations of mixed uranium and plutonium samples containing certified reference materials were analyzed by multi-collector inductively coupled plasma mass spectrometry and yielded good results. Current efforts have demonstrated a wider applicability of the prepFAST system. Near-quantitative removal of metal interferences was achieved with the system in both uranium and plutonium separated fractions. Successful system validation was completed with several archived samples. Isotopic results from archived samples and certified reference materials were well within data quality limits for the IAEA NWAL. Additional COTS equipment has been evaluated for its potential to aid the prepFAST-SR system in reducing the time allotments and cleanroom infrastructure requirements for accurate separations. Overall, these efforts help ensure continued efficient and effective operation of the NWAL. Comparison of the amount of personnel time necessary for successful manual vs. automated chemical separations showed a significant decrease in hands-on time from 9.8 hours to 35 minutes for seven samples, respectively. This documented time savings and reduced labor translates to a significant cost savings per sample. Overall, the system will enable faster sample reporting times at reduced cost by limiting personnel hours dedicated to the chemical separation.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

The distribution of galaxies’ gravitational field stemming from their tidal interaction

We calculate the distribution function of astronomical objects’ (like galaxies and/or smooth halos of different kinds) gravitational fields due to their tidal interaction. For that we apply the statistical method of Chandrasekhar, used originally to calculate the famous Holtzmark distribution. We show that in our approach the distribution function is never Gaussian, its form being dictated by the potential of interaction between objects. This calculation permits us to perform a theoretical analysis of the relation between angular momentum and mass (richness) of the galaxy clusters. To do so, we follow the ideas of Catelan and Theuns and Heavens and Peacock. The main difference is that here we reduce the problem to a discrete many-body case, where all physical properties of the system are determined by the interaction potential V(r{sub ij}). The essence of reduction is that we use the multipole (up to quadrupole here) expansion of Newtonian potential so that all hydrodynamic, “extended” characteristics of an object, such as its density mass, are “integrated out,” leaving its “point-like” characteristics, such as mass and quadrupole moment. In that sense we do not distinguish between galaxies and smooth components such as halos. We compare our theoretical results with observational data.

79 ASTRONOMY AND ASTROPHYSICS↗

Antiscatter grids in mobile C-arm cone-beam CT: Effect on image quality and dose

Purpose: X-ray scatter is a major detriment to image quality in cone-beam CT (CBCT). Existing geometries exhibit strong differences in scatter susceptibility with more compact geometries, e.g., dental or musculoskeletal, benefiting from antiscatter grids, whereas in more extended geometries, e.g., IGRT, grid use carries tradeoffs in image quality per unit dose. This work assesses the tradeoffs in dose and image quality for grids applied in the context of low-dose CBCT on a mobile C-arm for image-guided surgery. Methods: Studies were performed on a mobile C-arm equipped with a flat-panel detector for high-quality CBCT. Antiscatter grids of grid ratio (GR) 6:1-12:1, 40 lp/cm, were tested in ''body'' surgery, i.e., spine, using protocols for bone and soft-tissue visibility in the thoracic and abdominal spine. Studies focused on grid orientation, CT number accuracy, image noise, and contrast-to-noise ratio (CNR) in quantitative phantoms at constant dose. Results: There was no effect of grid orientation on possible gridline artifacts, given accurate angle-dependent gain calibration. Incorrect calibration was found to result in gridline shadows in the projection data that imparted high-frequency artifacts in 3D reconstructions. Increasing GR reduced errors in CT number from 31%, thorax, and 37%, abdomen, for gridless operation to 2% and 10%, respectively, with a 12:1 grid, while image noise increased by up to 70%. The CNR of high-contrast objects was largely unaffected by grids, but low-contrast soft-tissues suffered reduction in CNR, 2%-65%, across the investigated GR at constant dose. Conclusions: While grids improved CT number accuracy, soft-tissue CNR was reduced due to attenuation of primary radiation. CNR could be restored by increasing dose by factors of {approx}1.6-2.5 depending on GR, e.g., increase from 4.6 mGy for the thorax and 12.5 mGy for the abdomen without antiscatter grids to approximately 12 mGy and 30 mGy, respectively, with a high-GR grid. However, increasing the dose poses a significant impediment to repeat intraoperative CBCT and can cause the cumulative intraoperative dose to exceed that of a single diagnostic CT scan. This places the mobile C-arm in the category of extended CBCT geometries with sufficient air gap for which the tradeoffs between CNR and dose typically do not favor incorporation of an antiscatter grid.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

SU-E-I-78: Improving Prostatic Delineation Using Dual-Energy CT

Purpose: Visual prostatic definition is difficult using conventional CT. This is because the prostate is surrounded closely with tissue of similar electron density. Definition is further hindered when the region contains high-Z material (such as fiducial markers). Dual-energy CT (DECT) is a technique where images are rendered using two tube voltages during a single scan session. This study evaluates DECT as a means of improving prostatic volume delineation for radiation oncology. Methods: The patients were scanned using a Definition AS20 (Siemens Healthcare, Malvern, PA). This device uses a single-tube configuration, where two scans of differing energies are performed in serial. The scans are acquired with tube voltage of 80kVp and 140kVp. Following acquisition, these scan data were used to generate effective monoenergetic scans ranging from 40keV to 190keV. In the current study, the data were presented to observers using a novel program, which allows real-time adjustment of window, level, and effective keV; all while scrolling through volumetric slices. Three patients were scanned, each with a different high-contrast material in or around the prostate: I-125 seeds, gold fiducial markers, and prostatic calcifications. These images are compared to a weighted average of the 80kVp and 140kVP scans, which yield a scan similar to that of a 120 kVp scan, which is a common tube voltage in radiation oncology. Results: Prostatic definition improved in each case. Differentiation of soft tissue from surrounding adipose improved with lower keV, while higher keV provided a reduction of high-z artifacts. Furthermore, the dynamic adjustment of the keV allowed observers to better recognize regions of differing tissue composition within this relatively homogeneous area. Conclusion: By simultaneously providing the observer with the benefits of high-energy images and low-energy images, and allowing adjustment in real-time, improved imaging in highly homogeneous regions such as the male pelvis is achievable.

60 APPLIED LIFE SCIENCES↗

Effect of reconstruction methods and x-ray tube current–time product on nodule detection in an anthropomorphic thorax phantom: A crossed-modality JAFROC observer study

Purpose: To evaluate nodule detection in an anthropomorphic chest phantom in computed tomography (CT) images reconstructed with adaptive iterative dose reduction 3D (AIDR{sup 3D}) and filtered back projection (FBP) over a range of tube current–time product (mAs). Methods: Two phantoms were used in this study: (i) an anthropomorphic chest phantom was loaded with spherical simulated nodules of 5, 8, 10, and 12 mm in diameter and +100, −630, and −800 Hounsfield units electron density; this would generate CT images for the observer study; (ii) a whole-body dosimetry verification phantom was used to ultimately estimate effective dose and risk according to the model of the BEIR VII committee. Both phantoms were scanned over a mAs range (10, 20, 30, and 40), while all other acquisition parameters remained constant. Images were reconstructed with both AIDR{sup 3D} and FBP. For the observer study, 34 normal cases (no nodules) and 34 abnormal cases (containing 1–3 nodules, mean 1.35 ± 0.54) were chosen. Eleven observers evaluated images from all mAs and reconstruction methods under the free-response paradigm. A crossed-modality jackknife alternative free-response operating characteristic (JAFROC) analysis method was developed for data analysis, averaging data over the two factors influencing nodule detection in this study: mAs and image reconstruction (AIDR{sup 3D} or FBP). A Bonferroni correction was applied and the threshold for declaring significance was set at 0.025 to maintain the overall probability of Type I error at α = 0.05. Contrast-to-noise (CNR) was also measured for all nodules and evaluated by a linear least squares analysis. Results: For random-reader fixed-case crossed-modality JAFROC analysis, there was no significant difference in nodule detection between AIDR{sup 3D} and FBP when data were averaged over mAs [F(1, 10) = 0.08, p = 0.789]. However, when data were averaged over reconstruction methods, a significant difference was seen between multiple pairs of mAs settings [F(3, 30) = 15.96, p < 0.001]. Measurements of effective dose and effective risk showed the expected linear dependence on mAs. Nodule CNR was statistically higher for simulated nodules on images reconstructed with AIDR{sup 3D} (p < 0.001). Conclusions: No significant difference in nodule detection performance was demonstrated between images reconstructed with FBP and AIDR{sup 3D}. mAs was found to influence nodule detection, though further work is required for dose optimization.

60 APPLIED LIFE SCIENCES↗

Extending Parsimonious Bayesian Inference

Parsimonious Bayesian inference is a theoretical framework for efficient data assimilation that seeks to balance increased consistency between predictions and training data against corresponding increases in model complexity. Within this framework, over-training is understood as optimization that encodes excessive information within model parameters while only achieving small improvements between predictions and training data. This project aims to develop practical methods of limiting excess model information during optimization. One key observation is that practical heuristics for parsimonious learning in high-dimensions must balance expressivity, i.e. the ability of the model to capture diverse predictions with only a few non-zero parameters, against discoverability, i.e. the ability to train the model with gradient-based optimization and drive parameters to low information states. As such, we developed logical activation functions that are able to adaptively approximate arbitrary truth tables that define Boolean logic operations within a probabilistic framework. These functions have demonstrated the ability to learn exclusive disjunction (XOR) and conditioned disjunction (if [condition] then [result_if_true] else [result_if_false]) within a single layer of a neural network. To efficiently exploit these activation functions to drive parsimonious learning required several other advances within the domain of variational inference. The most efficient form of complexity suppression is structured sparsification, driving most model parameters to zero while achieving the structural coherence among nonzeros needed for bandwidth reduction. Such models are not only far more efficient at suppressing information-theoretic complexity, they also reduce the other forms of complexity (computations, communication, storage, and the number of dependencies needed to evaluate predictions). Aiming to support enhanced sparsification, this project examined new approaches to high-dimensional variational inference that allow us to calibrate and control parameter uncertainty during optimization. By identifying which parameters can sustain sparsifying perturbations with little impact on prediction quality, we can develop better pruning strategies by framing them as approximate Bayesian inference. These advances also open paths to mitigate concerns with deploying advanced learning methods in resource-constrained environments, such as running models on power-limited or communication-limited devices.

97 MATHEMATICS AND COMPUTING↗

Electrochemical hydrogen Storage Systems

As the global need for energy increases, scientists and engineers have found a possible solution by using hydrogen to power our world. Although hydrogen can be combusted as a fuel, it is considered an energy carrier for use in fuel cells wherein it is consumed (oxidized) without the production of greenhouse gases and produces electrical energy with high efficiency. Chemical storage of hydrogen involves release of hydrogen in a controlled manner from materials in which the hydrogen is covalently bound. Sodium borohydride and aminoborane are two materials given consideration as chemical hydrogen storage materials by the US Department of Energy. A very significant barrier to adoption of these materials as hydrogen carriers is their regeneration from 'spent fuel,' i.e., the material remaining after discharge of hydrogen. The U.S. Department of Energy (DOE) formed a Center of Excellence for Chemical Hydrogen Storage, and this work stems from that project. The DOE has identified boron hydrides as being the main compounds of interest as hydrogen storage materials. The various boron hydrides are then oxidized to release their hydrogen, thereby forming a 'spent fuel' in the form of a lower boron hydride or even a boron oxide. The ultimate goal of this project is to take the oxidized boron hydrides as the spent fuel and hydrogenate them back to their original form so they can be used again as a fuel. Thus this research is essentially a boron hydride recycling project. In this report, research directed at regeneration of sodium borohydride and aminoborane is described. For sodium borohydride, electrochemical reduction of boric acid and sodium metaborate (representing spent fuel) in alkaline, aqueous solution has been investigated. Similarly to literature reports (primarily patents), a variety of cathode materials were tried in these experiments. Additionally, approaches directed at overcoming electrostatic repulsion of borate anion from the cathode, not described in the previous literature for electrochemical reduction of spent fuels, have been attempted. A quantitative analytical method for measuring the concentration of sodium borohydride in alkaline aqueous solution has been developed as part of this work and is described herein. Finally, findings from stability tests for sodium borohydride in aqueous solutions of several different compositions are reported. For aminoborane, other research institutes have developed regeneration schemes involving tributyltin hydride. In this report, electrochemical reduction experiments attempting to regenerate tributyltin hydride from tributyltin chloride (a representative by-product of the regeneration scheme) are described. These experiments were performed in the non-aqueous solvents acetonitrile and 1,2-dimethoxyethane. A non-aqueous reference electrode for electrolysis experiments in acetonitrile was developed and is described. One class of boron hydrides, called polyhedral boranes, became of interest to the DOE due to their ability to contain a sufficient amount of hydrogen to meet program goals and because of their physical and chemical safety attributes. Unfortunately, the research performed here has shown that polyhedral boranes do not react in such a way as to allow enough hydrogen to be released, nor do they appear to undergo hydrogenation from the spent fuel form back to the original hydride. After the polyhedral boranes were investigated, the project goals remained the same but the hydrogen storage material was switched by the DOE to ammonia borane. Ammonia borane was found to undergo an irreversible hydrogen release process, so a direct hydrogenation was not able to occur. To achieve the hydrogenation of the spent ammonia borane fuel, an indirect hydrogenation reaction is possible by using compounds called organotin hydrides. In this process, the organotin hydrides will hydrogenate the spent ammonia borane fuel at the cost of their own oxidation, which forms organotin halides. To enable a closed-loop cycle, our task was then to be able to hydrogenate the organotin halides back to their hydride form. In addition to this experimental work, a parallel project was carried out to develop a new model of electrochemical impedance spectroscopy (EIS) that could be used to define the mechanisms of the electrochemical hydrogenation reactions. The EIS technique is capable of probing complex chemical and electrochemical reactions, and our model was written into a computer code that allowed the input of experimental EIS data and the extraction of kinetic parameters based on a best-fit analysis of theoretical reaction schemes. Finally, electrochemical methods for hydrogenating organic and metallo-organic materials have been explored.

08 HYDROGEN↗

Atmospheric Moisture Budget and Spatial Resolution Dependence of Precipitation Extremes in Aquaplanet Simulations

This study investigates the resolution dependency of precipitation extremes in an aqua-planet framework. Strong resolution dependency of precipitation extremes is seen over both tropics and extra-tropics, and the magnitude of this dependency also varies with dynamical cores. Moisture budget analyses based on aqua-planet simulations with the Community Atmosphere Model (CAM) using the Model for Prediction Across Scales (MPAS) and High Order Method Modeling Environment (HOMME) dynamical cores but the same physics parameterizations suggest that during precipitation extremes moisture supply for surface precipitation is mainly derived from advective moisture convergence. The resolution dependency of precipitation extremes mainly originates from advective moisture transport in the vertical direction. At most vertical levels over the tropics and in the lower atmosphere over the subtropics, the vertical eddy transport of mean moisture field dominates the contribution to precipitation extremes and its resolution dependency. Over the subtropics, the source of moisture, its associated energy, and the resolution dependency during extremes are dominated by eddy transport of eddies moisture at the mid- and upper-troposphere. With both MPAS and HOMME dynamical cores, the resolution dependency of the vertical advective moisture convergence is mainly explained by dynamical changes (related to vertical velocity or omega), although the vertical gradients of moisture act like averaging kernels to determine the sensitivity of the overall resolution dependency to the changes in omega at different vertical levels. The natural reduction of variability with coarser resolution, represented by areal data averaging (aggregation) effect, largely explains the resolution dependency in omega. The thermodynamic changes, which likely result from non-linear feedback in response to the large dynamical changes, are small compared to the overall changes in dynamics (omega). However, after excluding the data aggregation effect in omega, thermodynamic changes become relatively significant in offsetting the effect of dynamics leading to reduce differences between the simulated and aggregated results. Compared to MPAS, the simulated stronger vertical motion with HOMME also results in larger resolution dependency. Compared to the simulation at fine resolution, the vertical motion during extremes is insufficiently resolved/parameterized at the coarser resolution even after accounting for the natural reduction in variability with coarser resolution, and this is more distinct in the simulation with HOMME. To reduce uncertainties in simulated precipitation extremes, future development in cloud parameterizations must address their sensitivity to spatial resolution as well as dynamical cores.

aqua-planet↗

Forward sensitivity analysis and mode dependent control for closure modeling of Galerkin systems

Model reduction by projection-based approaches is often associated with losing some of the important features that contribute towards the dynamics of the retained scales. As a result, a mismatch occurs between the predicted trajectories of the original system and the truncated one. We put forth a framework to apply a continuous time control signal in the latent space of the reduced order model (ROM) to account for the effect of truncation. We set the control input using parameterized models by following energy transfer principles. Our methodology relies on observing the system behavior in the physical space and using the projection operator to restrict the feedback signal into the latent space. Then, we leverage the forward sensitivity method (FSM) to derive relationships between the feedback and the desired mode-dependent control. We test the performance of the proposed approach using two test cases, corresponding to viscous Burgers and vortex merger problems at high Reynolds number. Results show that the ROM trajectory with the applied FSM control closely matches its target values in both the data-dense and data-sparse regimes.

97 MATHEMATICS AND COMPUTING↗

WE-AB-202-01: Evaluating the Toxicity Reduction with CT-Ventilation Functional Avoidance Radiation Therapy

Purpose: CT-ventilation is an exciting new imaging modality that uses 4DCTs to calculate lung ventilation. Studies have proposed to use 4DCT-ventilation imaging for functional avoidance radiotherapy which implies designing treatment plans to spare functional portions of the lung. Although retrospective studies have been performed to evaluate the dosimetric gains to functional lung; no work has been done to translate the dosimetric gains to an improvement in pulmonary toxicity. The purpose of our work was to evaluate the potential reduction in toxicity for 4DCT-ventilation based functional avoidance. Methods: 70 lung cancer patients with 4DCT imaging were used for the study. CT-ventilation maps were calculated using the patient’s 4DCT, deformable image registrations, and a density-change-based algorithm. Radiation pneumonitis was graded using imaging and clinical information. Log-likelihood methods were used to fit a normal-tissue-complication-probability (NTCP) model predicting grade 2+ radiation pneumonitis as a function of doses (mean and V20) to functional lung (>15% ventilation). For 20 patients a functional plan was generated that reduced dose to functional lung while meeting RTOG 0617-based constraints. The NTCP model was applied to the functional plan to determine the reduction in toxicity with functional planning Results: The mean dose to functional lung was 16.8 and 17.7 Gy with the functional and clinical plans respectively. The corresponding grade 2+ pneumonitis probability was 26.9% with the clinically-used plan and 24.6% with the functional plan (8.5% reduction). The V20-based grade 2+ pneumonitis probability was 23.7% with the clinically-used plan and reduced to 19.6% with the functional plan (20.9% reduction). Conclusion: Our results revealed a reduction of 9–20% in complication probability with functional planning. To our knowledge this is the first study to apply complication probability to convert dosimetric results to toxicity improvement. The results presented in the current work provide seminal data for prospective clinical trials in functional avoidance. YV discloses funding from State of Colorado. TY discloses National Lung Cancer Partnership; Young Investigator Research grant.

60 APPLIED LIFE SCIENCES↗

Determining Coal Ash Impacts on Radium and Thorium Background Levels in Soil Using Rare Earth Element Analyses - 18604

One challenge at large remediation sites contaminated with radiological materials is differentiating between low-level site-related contamination and NORM. This is because sites contaminated with radiological materials, such as thorium processing and mining and milling facilities, often generate wastes that contain contaminants that are similar to NORM deposits. For sites in urban areas, this issue is further compounded due to the presence of historic industrial fill (historic fill) containing materials such as fly ash, coal ash, and red brick, that exhibit elevated radioactivity levels. These materials often have activity levels near, or even well above, a site's cleanup goal. At a large Superfund Site in the eastern US (Site), many of the suspected Site-related fill areas are commingled with coal and fly ash. As the cleanup criteria for the radionuclides of concern (ROC), radium (Ra) and thorium (Th), is 5 pico-Curies per gram above background, a four-pronged approach was developed to determine if the coal and fly ash are impacting the background levels of the ROCs. This evaluation included a review of boring logs to visually confirm the presence of historic fill, as well as coal fragments and coal ash; a literature review to establish the typical concentrations of Ra and Th found in coal and fly ash from US sources; the use of rare earth analyses to establish chemical signatures for non-Site related historic fill containing coal ash versus Site-related fill; and a quantitative analysis of Site-related data to determine if the radioactivity levels in historic fill containing coal were elevated. Soil samples were collected from two locations at the Site; fill areas containing coal ash that were not suspected to be impacted by Site activities, as well as fill areas containing coal ash that were associated with the former processing facility. These samples were analyzed for rare earth elements (REE), as well as Ra-228 and Th-232 and the ratios of lanthanum, cerium, and neodymium to Ra-228 and Th-232 were established for each sample. Based on a comparison of these data, a strong correlation was established between the rare earth/thorium decay chain ratios and the classification of material as either Site or non-Site related. These data were used to develop a higher background activity for fill areas containing non-Site related NORM. While the cleanup level defined in the Record of Decision remained unchanged, the higher background derived from the fly ash/coal ash material, as identified by REE analyses, resulted in higher project screening levels, which ensured that excavation activities did not 'chase' material that was not Site-related. To date, use of the coal ash background level has saved millions of dollars in cleanup costs since non-Site related contamination is not being remediated. Use of this approach to determine the appropriate background level for a site is relatively straightforward since it relies on visual observation of the soil for coal and fly ash, along with REE analyses, which is an inexpensive, quick, and established analytical method for soil samples. As long as a good rare earth profile of both contaminated and natural (in this case fly ash and coal ash) material can be established, the results can be quite definitive, which could result in significant volume reduction by minimizing the amount of material excavated and disposed of. This will lead to cost and time savings for what could be an otherwise costly and potentially complex excavation. Anyone performing environmental remediation of NORM contaminated sites may benefit from this paper and presentation. (authors)

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

A method for deriving a 4D-interpolated balanced planning target for mobile tumor radiotherapy

Purpose: Tumor control and normal tissue toxicity are strongly correlated to the tumor and normal tissue volumes receiving high prescribed dose levels in the course of radiotherapy. Planning target definition is, therefore, crucial to ensure favorable clinical outcomes. This is especially important for stereotactic body radiation therapy of lung cancers, characterized by high fractional doses and steep dose gradients. The shift in recent years from population-based to patient-specific treatment margins, as facilitated by the emergence of 4D medical imaging capabilities, is a major improvement. The commonly used motion-encompassing, or internal-target volume (ITV), target definition approach provides a high likelihood of coverage for the mobile tumor but inevitably exposes healthy tissue to high prescribed dose levels. The goal of this work was to generate an interpolated balanced planning target that takes into account both tumor coverage and normal tissue sparing from high prescribed dose levels, thereby improving on the ITV approach. Methods: For each 4DCT dataset, 4D deformable image registration was used to derive two bounding targets, namely, a 4D-intersection and a 4D-composite target which minimized normal tissue exposure to high prescribed dose levels and maximized tumor coverage, respectively. Through definition of an ''effective overlap volume histogram'' the authors derived an ''interpolated balanced planning target'' intended to balance normal tissue sparing from prescribed doses with tumor coverage. To demonstrate the dosimetric efficacy of the interpolated balanced planning target, the authors performed 4D treatment planning based on deformable image registration of 4D-CT data for five previously treated lung cancer patients. Two 4D plans were generated per patient, one based on the interpolated balanced planning target and the other based on the conventional ITV target. Plans were compared for tumor coverage and the degree of normal tissue sparing resulting from the new approach was quantified. Results: Analysis of the 4D dose distributions from all five patients showed that while achieving tumor coverage comparable to the ITV approach, the new planning target definition resulted in reductions of lung V{sub 10}, V{sub 20}, and V{sub 30} of 6.3% {+-} 1.7%, 10.6% {+-} 3.9%, and 12.9% {+-} 5.5%, respectively, as well as reductions in mean lung dose, mean dose to the GTV-ring and mean heart dose of 8.8% {+-} 2.5%, 7.2% {+-} 2.5%, and 10.6% {+-} 3.6%, respectively. Conclusions: The authors have developed a simple and systematic approach to generate a 4D-interpolated balanced planning target volume that implicitly incorporates the dynamics of respiratory-organ motion without requiring 4D-dose computation or optimization. Preliminary results based on 4D-CT data of five previously treated lung patients showed that this new planning target approach may improve normal tissue sparing without sacrificing tumor coverage.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

Reactant Carrier Microfoam Technology for In-Situ Remediation of Radionuclide and Metallic Contaminants in Deep Vadose Zone

The U.S. Department of Energy (DOE) is currently developing advanced remedial technologies for addressing metal and radionuclide (Cr, Tc, and U) contamination in deep vadose zone environments. One of the transformational technology alternatives being considered by the DOE Office of Environmental Management, is the use of Reactant Carrier Microfoams (RCM) as a minimally invasive method for delivery and emplacement of reagents for in-situ immobilization of contaminants. Penetration of low permeability zones deep within the subsurface for Enhance Oil Recovery (EOR) has been well-established. Use of surfactant foams have also been explored for mobilizing DNAPL from sediments. So far, the concept of using RCM for immobilizing labile metal and long-lived radionuclide contaminants in the deep vadose zone has not been explored. We, at the Pacific Northwest National Laboratory (PNNL), conducted studies to develop stable foams as a means to deliver reductive and/or precipitating reactants to the deep subsurface. To test the feasibility of this approach, we developed a preliminary foam formulation consisting of a mixture of an anionic and a nonionic surfactant with a reactant consisting of a 9:1 blend of tripoly- and orthophosphate. The MSE Technology Applications, Inc (MSE) in collaboration with PNNL, conducted a scale-up test to evaluate the efficacy of this reactant carrier foam for in-situ immobilization of U containing sediment zones in a heterogenous sediment matrix. The data indicated that successful immobilization of U contamination is feasible using specifically tailored reactant carrier foam injection technology. Studies are continuing for developing more robust optimized RCM for highly mobile contaminants such as Cr (VI), Tc (VII) in the deep vadose zone.

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

Projected Growth in Small-Scale, Fossil-Fueled Distributed Generation: Potential Implications for the U.S. Greenhouse Gas Inventory

The generation capacity of small-scale (less than one megawatt) fossil-fueled electricity in the United States is anticipated to grow by threefold to twenty-fold from 2015 to 2040. However, in adherence with internationally agreed upon carbon accounting methods, the Environmental Protection Agency's (EPA's) U.S. Greenhouse Inventory (GHGI) does not currently attribute greenhouse gases (GHGs) from these small-scale distributed generation sources to the electric power sector and instead accounts for these emissions in the sector that uses the distributed generation (e.g., the commercial sector). In addition, no other federal electric-sector GHG emission data product produced by the EPA or the U.S. Energy Information Administration (EIA) can attribute these emissions to electricity. We reviewed the technical documentation for eight federal electric-sector GHG emission data products, interviewed the data product owners, collected their GHG emission estimates, and analyzed projections for growth in fossil-fueled distributed generation. We show that, by 2040, these small-scale generators could account for at least about 1%- 5% of total CO2 emissions from the U.S. electric power sector. If these emissions fall outside the electric power sector, the United States may not be able to completely and accurately track changes in electricity-related CO2 emissions, which could impact how the country sets GHG reduction targets and allocates mitigation resources. Because small-scale, fossil-fueled distributed generation is expected to grow in other countries as well, the results of this work also have implications for global carbon accounting.

54 ENVIRONMENTAL SCIENCES↗

Jensen–Shannon divergence based novel loss functions for Bayesian neural networks

Bayesian neural networks (BNNs) are state-of-the-art machine learning methods that can naturally regularize and systematically quantify uncertainties using their stochastic parameters. Kullback–Leibler (KL) divergence-based variational inference used in BNNs suffer from unstable optimization and challenges in approximating light-tailed posteriors due to the unbounded nature of the KL divergence. To resolve these issues, we formulate a novel loss function for BNNs based on a new modification to the generalized Jensen–Shannon (JS) divergence, which is bounded. In addition, we propose a Geometric JS divergence-based loss, which is computationally efficient since it can be evaluated analytically. We found that the JS divergence-based variational inference is intractable, and hence employed a constrained optimization framework to formulate these losses. Our theoretical analysis and empirical experiments on multiple regression and classification data sets suggest that the proposed losses perform better than the KL divergence-based loss, especially when the data sets are noisy or biased. Specifically, there are approximately 5% and 8% improvements in accuracy for a noise-added CIFAR-10 dataset and a regression dataset, respectively. There is about 13% reduction in false negative predictions of a biased histopathology dataset. Additionally, we quantify and compare the uncertainty metrics for the regression and classification tasks.

97 MATHEMATICS AND COMPUTING↗