Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “data representations”

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 1,693 records · Page 94
All sourcesNASA NTRSDOE OSTI
DOE OSTI2020

Ice Nucleating Particles, Aerosols and Clouds over the Higher Latitude Southern Ocean

Improved understanding of aerosol-cloud-precipitation interactions is a key goal of the DOE-ASR program. Ice Nucleating Particles (INPs) are rare aerosol particles that trigger ice formation in clouds. By doing so they may initiate precipitation in clouds, and this impacts cloud lifetime. Nowhere may the role of INPs be manifested so dramatically as in the Southern Ocean (SO). This vast region was the focus of the DOE-ARM MARCUS (Measurement of Aerosols, Radiation and CloUds over the Southern Oceans) and MICRE (Macquarie Island Cloud and Radiation Experiment) campaigns. In the SO, a deficit of INPs has been a leading hypothesis for why liquid clouds persist, which underlays the bias of global climate models in predicting excess shortwave radiation reaching the ocean surface between 55°S and Antarctica. In this study, we used the 6-month acquisition of INP data on four MARCUS cruises from Hobart, Tasmania, to Antarctica, and the single-location full annual INP cycle observed in MICRE, to gain a holistic, quantitative picture of the INP number concentrations as a function of temperature and their variability over SO latitudes from 43°S to Antarctica. Augmenting initial processing of filter collections of aerosol particles during MARCUS and MICRE, additional freezing studies following thermal treatment (to removes INPs associated with microbes) and peroxide treatment (to remove all organics), as well as ionic chemistry and total organic carbon analysis, were applied to samples to improve time and space resolution, and differentiate marine from terrestrial contributions. We also performed Next Generation Sequencing to determine that bacterial composition can tag periods of enhanced and degraded marine organic INPs and the relative lack of occurrence of land-sourced bio-particles. Through these and other analyses, INP data were categorized into representative source types (e.g., marine versus terrestrially-influenced). The products of our analyses are being used to parameterize ice formation for use in numerical modeling studies to determine if, and when, ocean-derived INPs control the microphysical composition and radiative balance of SO clouds in a manner not presently captured by global models. The specific objectives and approaches proposed were largely accomplished, including: 1) Additional and value-added chemical and biological analyses of archived samples of aerosols were completed to categorize and quantify INP types and concentrations over the SO. 2) Analyses were completed to place INP data in meteorological and aerosol context. 3) Sea spray aerosol INPs were demonstrated to dominate the SO region in all but episodic events. 4) New parameterizations for INP sources relevant to the SO were constructed. 5) Collaborative modeling studies demonstrating the crucial role of INPs in determining cloud radiative and precipitation properties were conducted with MARCUS and MICRE partners. Collaborative publication submissions in this regard ensued within the two year study. This work will advance the science of interactions of aerosols, clouds and precipitation, to improve their representation in regional and global climate models. Results have informed representation of primary ice crystal nucleation, give inference to the role of secondary processes, and improve investigations of aerosol influences on clouds via ice nucleation over SO high latitudes and similar vast ocean regions. Application will improve representation of such interactions for clouds in both regional and global climate models. The data base and analyses methods applied will continue to serve research studies in the future.

58 GEOSCIENCES↗
DOE OSTI2025

Towards provision of regularly updated climate data from the Coupled Model Intercomparison Project

The Coupled Model Intercomparison Project (CMIP) is a flagship of the World Climate Research Programme (WCRP). CMIP has become a recognised ‘brand’ in climate circles evolving over the last thirty years from a targeted research activity by a small number of climate modelling centres intercomparing their Earth System Model (ESM) simulations to a broad international coordinated research effort (Durack et al, 2025). CMIP is organized as a research activity leveraging funded and in-kind contributions from experts within modelling centres and the broader scientific community supported more recently by a fully-funded International Project Office. Within CMIP, Model Intercomparison Projects (MIPs) are community-designed to understand past, present and future climate. CMIP data provides a valuable resource for climate research and is routinely used to assess model representation of climate processes and test scientific hypotheses in the context of model uncertainty and (forced and internal) variability as evident from its prolific use in scientific publications1 . The impact relies on enabling infrastructure (most prominently via the Earth System Grid Federation (ESGF)), which allows sharing of simulation output, provision of the boundary conditions used in each simulation, and definition of the data standards that are essential to facilitating wide use of the data. The impact is supplemented by the wide-ranging scrutiny to which model simulations are subjected. Beyond its use in research, CMIP data is a key resource for communities producing derived climate information from downscaling and impact studies, such as the Coordinated Regional Downscaling Experiment (CORDEX; Gutowski et al., 2016) and the Intersectoral Impacts MIP (ISIMIP; Frieler et al., 2024). Government, academic and commercial entities also increasingly rely on CMIP and its downstream data for climate risk assessments and climate services (for example, Copernicus Climate Change Service and World Bank portal). This means that, although CMIP is a research activity, it increasingly serves a secondary and very relevant role as a provider of climate data – a long-recognised dichotomy (Stevens, 2024). Research and applications have distinct needs, with the former requiring flexibility and generality and the latter consistency. Here we explain how the design of the research activity has been adapted to reduce the burdens imposed by applications and how the research infrastructure might evolve to further enable scientific inquiry. We propose one possible approach to consistently providing model information and projections for applications in the future.

Environmental sciences↗
DOE OSTI2025

Poster Abstract: Leveraging Large Language Models to Reveal Interpretable Cooling Behaviors from Smart Thermostat Data

Frequent heatwaves and hot summers increasingly challenge occupant comfort, health, and energy grid stability. Addressing these challenges requires a detailed understanding of household cooling behaviors, such as thermostat adjustments and adaptive responses to extreme conditions. Traditional analyses often rely on aggregated numerical metrics that overlook subtle but important household-specific variations. In this study, we introduce a generalizable methodology that integrates large language models (LLMs) with vision capabilities to enable scalable and detailed analysis of residential thermostat data. Using Ecobee's Donate Your Data (DYD) dataset—which provides five-minute records of indoor temperatures, thermostat setpoints, and HVAC runtimes—we focus on two U.S. cities with contrasting summer climates : Austin (TX) and Phoenix (AZ). Because raw time-series data are not well suited for direct LLM analysis, we transform them into visual representations, such as daily indoor temperature trajectories and weekly runtime histograms, to better capture behavioral variations. Leveraging LLMs' visual interpretation, we extract descriptive behavioral features, including temperature preferences, time-of-day cooling orientation, anticipatory versus reactive heatwave responses, and behavioral consistency. These semantic features support unsupervised clustering to identify distinct occupant archetypes at scale, revealing differences—such as morning-centric anticipatory coolers versus households that shift toward warmer setpoints during heatwaves—that can inform demand response, resilience planning, and health-aware interventions. By converting raw numerical data into interpretable behavioral patterns, this methodology enables scalable and practical analysis of occupant behavior, supporting actionable insights for comfort, resilience, and energy management.

Nihar, Kopal↗
NASA NTRS2019

Modeling Phenological Controls on Carbon Dynamics in Dryland Sagebrush Ecosystems

Dryland ecosystems play an important role in determining how precipitation anomalies affect terrestrial carbonfluxes at regional to global scales. Thus, to understand how climate change may affect the global carbon cycle,we must also be able to understand and model its effects on dryland vegetation. Dynamic Global VegetationModels (DGVMs) are an important tool for modeling ecosystem dynamics, but they often struggle to reproduceseasonal patterns of plant productivity. Because the phenological niche of many plant species is linked to bothtotal productivity and competitive interactions with other plants, errors in how process-based models representphenology hinder our ability to predict climate change impacts. This may be particularly problematic in drylandecosystems where many species have developed a complex phenology in response to seasonal variability in bothmoisture and temperature. Here, we examine how uncertainty in key parameters as well as the structure ofexisting phenology routines affect the ability of a DGVM to match seasonal patterns of leaf area index (LAI) andgross primary productivity (GPP) across a temperature and precipitation gradient. First, we optimized modelparameters using a combination of site-level eddy covariance data and remotely-sensed LAI data. Second, wemodified the model to include a semi-deciduous phenology type and added flexibility to the representation ofgrass phenology. While optimizing parameters reduced model bias, the largest gains in model performance wereassociated with the development of our new representation of phenology. This modified model was able to bettercapture seasonal patterns of both leaf area index (R2=0.75) and gross primary productivity (R2=0.84), thoughits ability to estimate total annual GPP depended on using eddy covariance data for optimization. The new modelalso resulted in a more realistic outcome of modeled competition between grass and shrubs. These findingsdemonstrate the importance of improving how DGVMs represent phenology in order to accurately forecastclimate change impacts in dryland ecosystems.

Ecosystem model Phenology Parameter optimization E↗
DOE OSTI2015

Quantifying and Reducing Curve-Fitting Uncertainty in Isc: Preprint

Current-voltage (I-V) curve measurements of photovoltaic (PV) devices are used to determine performance parameters and to establish traceable calibration chains. Measurement standards specify localized curve fitting methods, e.g., straight-line interpolation/extrapolation of the I-V curve points near short-circuit current, Isc. By considering such fits as statistical linear regressions, uncertainties in the performance parameters are readily quantified. However, the legitimacy of such a computed uncertainty requires that the model be a valid (local) representation of the I-V curve and that the noise be sufficiently well characterized. Using more data points often has the advantage of lowering the uncertainty. However, more data points can make the uncertainty in the fit arbitrarily small, and this fit uncertainty misses the dominant residual uncertainty due to so-called model discrepancy. Using objective Bayesian linear regression for straight-line fits for Isc, we investigate an evidence-based method to automatically choose data windows of I-V points with reduced model discrepancy. We also investigate noise effects. Uncertainties, aligned with the Guide to the Expression of Uncertainty in Measurement (GUM), are quantified throughout.

14 SOLAR ENERGY↗
DOE OSTI2017

Fast parametric relationships for the large-scale reservoir simulation of mixed CH 4 -CO 2 gas hydrate systems

A recent Department of Energy field test on the Alaska North Slope has increased interest in the ability to simulate systems of mixed CO 2 -CH 4 hydrates. However, the physically realistic simulation of mixed-hydrate simulation is not yet a fully solved problem. Limited quantitative laboratory data leads to the use of various ab initio, statistical mechanical, or other mathematic representations of mixed-hydrate phase behavior. Few of these methods are suitable for inclusion in reservoir simulations, particularly for systems with large number of grid elements, 3D systems, or systems with complex geometric configurations. In this paper, we present a set of fast parametric relationships describing the thermodynamic properties and phase behavior of a mixed methane-carbon dioxide hydrate system. We use well-known, off-the-shelf hydrate physical properties packages to generate a sufficiently large dataset, select the most convenient and efficient mathematical forms, and fit the data to those forms to create a physical properties package suitable for inclusion in the TOUGH+ family of codes. Finally, the mapping of the phase and thermodynamic space reveals the complexity of the mixed-hydrate system and allows understanding of the thermodynamics at a level beyond what much of the existing laboratory data and literature currently offer.

58 GEOSCIENCES↗
DOE OSTI2024

Learning together: Towards foundation models for machine learning interatomic potentials with meta-learning

Abstract The development of machine learning models has led to an abundance of datasets containing quantum mechanical (QM) calculations for molecular and material systems. However, traditional training methods for machine learning models are unable to leverage the plethora of data available as they require that each dataset be generated using the same QM method. Taking machine learning interatomic potentials (MLIPs) as an example, we show that meta-learning techniques, a recent advancement from the machine learning community, can be used to fit multiple levels of QM theory in the same training process. Meta-learning changes the training procedure to learn a representation that can be easily re-trained to new tasks with small amounts of data. We then demonstrate that meta-learning enables simultaneously training to multiple large organic molecule datasets. As a proof of concept, we examine the performance of a MLIP refit to a small drug-like molecule and show that pre-training potentials to multiple levels of theory with meta-learning improves performance. This difference in performance can be seen both in the reduced error and in the improved smoothness of the potential energy surface produced. We therefore show that meta-learning can utilize existing datasets with inconsistent QM levels of theory to produce models that are better at specializing to new datasets. This opens new routes for creating pre-trained, foundation models for interatomic potentials.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2023

Collaborative Proposal: Improving Understanding of the Internal Structure and Dynamics of Deep Convection Using ARM Observations and Large Eddy Simulations (Final progress report)

This is the final technical report for the DOE Atmospheric System Research funded project entitled "Collaborative Proposal: Improving Understanding of the Internal Structure and Dynamics of Deep Convection Using ARM Observations and Large Eddy Simulations". This project addressed several science questions related to the structure and behavior of deep convective clouds in the atmosphere. It applied knowledge gained from this work to improve the deep cumulus convection parameterization in the Community Earth System Model (CESM). Specific outcomes related to the project include an improved understanding of: (1) how convective cloud environments regulate the size of thermals reaching the upper troposphere, and hence the shallow-to-deep convective transition; (2) the role of dynamic pressure forcing in thermal and convective cloud evolution; and (3) the role of vertical wind shear on the behavior of moist thermals and the transition from shallow to deep convection. Although the primary motivation of this project was to improve basic understanding of thermal dynamics in atmospheric convection, an important component of this work was using this improved understanding to modify the representation of entrainment and vertical velocity profiles in the Zhang-McFarlane cumulus parameterization used in CESM. This work also had a strong observational component, with analyses of Atmospheric Radiation Measurement (ARM) observations to characterize thermal characteristics and behavior. Observations collected during the ARM-supported CACTI field project were used with high-resolution simulations to explore environmental controls on deep convection initiation. Data from the ARM-supported MC3E field project were also used to evaluate the representation of updraft vertical velocities in the Zhang-McFarlane cumulus parameterization. This project directly supported the training of two post-doctoral researchers and led to 16 peer-reviewed published papers.

54 ENVIRONMENTAL SCIENCES↗
NASA NTRS1990

System identification in the repetition domain

Procedures for system identification using realization theory in conjunction with learning control ideas are developed. The Markov parameters of the system are identified by combining data from repeated experiments. Three approaches are discussed for identification of as many Markov parameters as sample points in the experiment. Making use of all the parameters, realization theory is then employed to determine the system order and to obtain a minimal order representation. The first two approaches are non-recursive, which in the case of noise-free data yields a one step solution. The third approach uses a recursive formulation rendered from adaptive control but modified for successive experiments. A simple example shows the numerical convergence of the identified parameters as a function of the number of experiments. The procedure presented herein is an extension of the existing Eigensystem Realization Algorithm (ERA), which has been successfully applied for system identification of large structures.

Juang, Jer-Nan↗
NASA NTRS1998

Functional and structural mapping of human cerebral cortex: solutions are in the surfaces

The human cerebral cortex is notorious for the depth and irregularity of its convolutions and for its variability from one individual to the next. These complexities of cortical geography have been a chronic impediment to studies of functional specialization in the cortex. In this report, we discuss ways to compensate for the convolutions by using a combination of strategies whose common denominator involves explicit reconstructions of the cortical surface. Surface-based visualization involves reconstructing cortical surfaces and displaying them, along with associated experimental data, in various complementary formats (including three-dimensional native configurations, two-dimensional slices, extensively smoothed surfaces, ellipsoidal representations, and cortical flat maps). Generating these representations for the cortex of the Visible Man leads to a surface-based atlas that has important advantages over conventional stereotaxic atlases as a substrate for displaying and analyzing large amounts of experimental data. We illustrate this by showing the relationship between functionally specialized regions and topographically organized areas in human visual cortex. Surface-based warping allows data to be mapped from individual hemispheres to a surface-based atlas while respecting surface topology, improving registration of identifiable landmarks, and minimizing unwanted distortions. Surface-based warping also can aid in comparisons between species, which we illustrate by warping a macaque flat map to match the shape of a human flat map. Collectively, these approaches will allow more refined analyses of commonalities as well as individual differences in the functional organization of primate cerebral cortex.

Review↗
DOE OSTI2020

Efficient I/O for Neural Network Training with Compressed Data

FanStore is a shared object store that enables efficient and scalable neural network training on supercomputers. By providing a global cache layer on node-local burst buffers using a compressed representation, it significantly enhances the processing capability of deep learning (DL) applications on existing hardware. In addition, FanStore allows POSIX-compliant file access to the compressed data in user space. We investigate the tradeoff between runtime overhead and data compression ratio using real-world datasets and applications, and propose a compressor selection algorithm to maximize storage capacity given performance constraints. We consider both asynchronous (i.e., with prefetching) and synchronous I/O strategies, and propose mechanisms for selecting compressors for both approaches. Using FanStore, the same storage hardware can host 2-13x more data for example applications without significant runtime overhead. Empirically, our experiments show that FanStore scales to 512 compute nodes with near linear performance scalability.

Zhang, Zhao↗
NASA NTRS2014

Near-Real Time Satellite-Retrieved Cloud and Surface Properties for Weather and Aviation Safety Applications

Cloud properties determined from satellite imager radiances provide a valuable source of information for nowcasting and weather forecasting. In recent years, it has been shown that assimilation of cloud top temperature, optical depth, and total water path can increase the accuracies of weather analyses and forecasts. Aircraft icing conditions can be accurately diagnosed in near-­‐real time (NRT) retrievals of cloud effective particle size, phase, and water path, providing valuable data for pilots. NRT retrievals of surface skin temperature can also be assimilated in numerical weather prediction models to provide more accurate representations of solar heating and longwave cooling at the surface, where convective initiation. These and other applications are being exploited more frequently as the value of NRT cloud data become recognized. At NASA Langley, cloud properties and surface skin temperature are being retrieved in near-­‐real time globally from both geostationary (GEO) and low-­‐earth orbiting (LEO) satellite imagers for weather model assimilation and nowcasting for hazards such as aircraft icing. Cloud data from GEO satellites over North America are disseminated through NCEP, while those data and global LEO and GEO retrievals are disseminated from a Langley website. This paper presents an overview of the various available datasets, provides examples of their application, and discusses the use of the various datasets downstream. Future challenges and areas of improvement are also presented.

Minnis, Patrick↗
NASA NTRS2006

Diurnal Variability of Vertical Structure from a TRMM Passive Microwave "Virtual Radar" Retrieval

Robust description of the diurnal cycle from TRMM observations is complicated by the limitations of Low Earth Orbit (LEO) sampling; from a 'climatological' perspective, sufficient sampling must exist to control for both spatial and seasonal variability, before tackling an additional diurnal component (e.g., with 8 additional 3-hourly or 24 1-hourly bins). For documentation of vertical structure, the narrow sample swath of the TRMM Precipitation Radar limits the resolution of any of these components. A neural-network based 'virtual radar" retrieval has been trained and internally validated, using multifrequency / multipolarization passive microwave(TM1) brightness temperatures and textures parameters and lightning (LIS) observations, as inputs, and PR volumetric reflectivity as targets (outputs). By training the algorithms (essentially highly multivariate, nonlinear regressions) on a very large sample of high-quality co-located data from the center of the TRMM swath, 3D radar reflectivity and derived parameters (VIL, IWC, Echo Tops, etc.) can be retrieved across the entire TMI swath, good to 8-9% over the dynamic range of parameters. As a step in the retrieval (and as an output of the process), each TMI multifrequency pixel (at 85 GHz resolution) is classified into one of the 25 archetypal radar profile vertical structure "types", previously identified using cluster analysis. The dynamic range of retrieved vertical structure appears to have higher fidelity than the current (Version 6) experimental GPROF hydrometeor vertical structure retrievals. This is attributable to correct representation of the prior probabilities of vertical structure variability in the neural network training data, unlike the GPROF cloud-resolving model training dataset used in the V6 algorithms. The LIS lightning inputs are supplementary inputs, and a separate offline neural network has been trained to impute (predict) LIS lightning from passive-microwave-only data. The virtual radar retrieval is thus, in principle, extensible to Aqua/AMSR-E and NPOESS/CMIS passive microwave instruments. The virtual radar approach yields a threefold increase in effective sampling from the mission, albeit of lower-quality "retrieved" data, reducing the variance of local estimates by one third (or the standard deviation by-0.57). In this talk, the variance reduction is leveraged to more finely resolve global diurnal variability in both space and time (local hour).

Boccippio, Dennis J.↗
DOE OSTI2018

Using the JCC-71 neutron coincidence collar as a benchmark for detector characterization with PTR-32 list mode data acquisition

Augmenting or interchanging current international safeguards shift registerbased data acquisition systems with list mode data acquisition systems gives nondestructive assay (NDA) systems greater versatility. Neutron list mode data analysis offers comparable analytical results to the more widely used shift register analysis in nuclear material quantification applications and offers several diagnostic tools that are specifically beneficial to in-laboratory characterization and calibration measurements. These benefits include identification of non-ideal behavior, optimization of operational parameters from a single measurement, and an improved understanding of the physics-based behavior of NDA systems for a more precise system representation and more confident assay results. In this work the advantages of using list mode data acquisition for detector characterization are demonstrated experimentally. Two commercial-off-the-shelf International Atomic Energy Agency-supported technologies are used for a typical detector characterization procedure and their performance is compared. Specifically, a 3 He-based Canberra Industries JCC-71 Neutron Coincidence Collar is characterized using the Hungarian Institute of Isotopes’ Pulse Train Recorder-32 (PTR-32) list mode data acquisition system, and the results are compared to those obtained using the standard Canberra Industries JSR-15 model shift register. The quantitative results from the two systems are in agreement, which demonstrates that the PTR-32 is a technically viable alternative to conventional shift register electronics for this task. A suitable procedure for full instrument characterization is described, and the added benefits of list mode for characterization and data collection are discussed. Furthermore, this is an important step toward establishing a procedure for the routine use of list mode data acquisition and analysis for neutron NDA system characterization in safeguards field applications.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗
NASA NTRS1994

Experiments with the Mesoscale Atmospheric Simulation System (MASS) using the synthetic relative humidity

This study is intended to examine the impact of the synthetic relative humidity on the model simulation of mesoscale convective storm environment. The synthetic relative humidity is derived from the National Weather Services surface observations, and non-conventional sources including aircraft, radar, and satellite observations. The latter sources provide the mesoscale data of very high spatial and temporal resolution. The synthetic humidity data is used to complement the National Weather Services rawinsonde observations. It is believed that a realistic representation of initial moisture field in a mesoscale model is critical for the model simulation of thunderstorm development, and the formation of non-convective clouds as well as their effects on the surface energy budget. The impact will be investigated based on a real-data case study using the mesoscale atmospheric simulation system developed by Mesoscale Environmental Simulations Operations, Inc. The mesoscale atmospheric simulation system consists of objective analysis and initialization codes, and the coarse-mesh and fine-mesh dynamic prediction models. Both models are a three dimensional, primitive equation model containing the essential moist physics for simulating and forecasting mesoscale convective processes in the atmosphere. The modeling system is currently implemented at the Applied Meteorology Unit, Kennedy Space Center. Two procedures involving the synthetic relative humidity to define the model initial moisture fields are considered. It is proposed to perform several short-range (approximately 6 hours) comparative coarse-mesh simulation experiments with and without the synthetic data. They are aimed at revealing the model sensitivities should allow us both to refine the specification of the observational requirements, and to develop more accurate and efficient objective analysis schemes. The goal is to advance the MASS (Mesoscal Atmospheric Simulation System) modeling expertise so that the model output can provide reliable guidance for thunderstorm forecasting.

Chang, Chia-Bo↗
DOE OSTI2020

Optical Wavelength Guided Self-Supervised Feature Learning For Galaxy Cluster Richness Estimate

Most galaxies in the nearby Universe are gravitationally bound to a cluster or group of galaxies. Their optical contents, such as optical richness, are crucial for understanding the co-evolution of galaxies and large-scale structures in modern astronomy and cosmology. The determination of optical richness can be challenging. We propose a self-supervised approach for estimating optical richness from multi-band optical images. The method uses the data properties of the multi-band optical images for pre-training, which enables learning feature representations from a large but unlabeled dataset. We apply the proposed method to the Sloan Digital Sky Survey. The result shows our estimate of optical richness lowers the mean absolute error and intrinsic scatter by 11.84% and 20.78%, respectively, while reducing the need for labeled training data by up to 60%. We believe the proposed method will benefit astronomy and cosmology, where a large number of unlabeled multi-band images are available, but acquiring image labels is costly.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2026

A Physical Model Enhanced Data Driven Method for High-Resolution Residential Load Profile Generation

Residential buildings account for significant energy consumption, creating opportunities to offer grid services. As electric utilities seek to implement effective system operation strategies, understanding residential energy consumption patterns becomes essential; However, the time intervals of load profiles measured by utilities' smart meters are typically from 15 minutes to 60 minutes. The low-resolution data make it hard to extract appliance-level load information, which is critical for providing grid services. This paper presents a load profile generator designed to produce synthetic load profiles for residential buildings that emphasizes the importance of accurate representations of realistic energy consumption patterns. The generator takes realistic low-resolution residential load measurements and weather data as inputs, producing 1-minute interval profiles that match the characteristics of the original profiles. Further, this generator can be used to populate load profiles in areas where actual measurements are limited to improve the ability of utilities to analyze their distribution systems. By providing more high-resolution residential building load profiles, this tool supports electric utilities to enhance their residential building load control strategies and improve overall grid stability.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗
DOE OSTI2020

Can machine learning improve the model representation of turbulent kinetic energy dissipation rate in the boundary layer for complex terrain?

Current turbulence parameterizations in numerical weather prediction models at the mesoscale assume a local equilibrium between production and dissipation of turbulence. As this assumption does not hold at fine horizontal resolutions, improved ways to represent turbulent kinetic energy (TKE) dissipation rate (ϵ) are needed. Here, we use a 6-week data set of turbulence measurements from 184 sonic anemometers in complex terrain at the Perdigão field campaign to suggest improved representations of dissipation rate. First, we demonstrate that the widely used Mellor, Yamada, Nakanishi, and Niino (MYNN) parameterization of TKE dissipation rate leads to a large inaccuracy and bias in the representation of ϵ. Next, we assess the potential of machine-learning techniques to predict TKE dissipation rate from a set of atmospheric and terrain-related features. We train and test several machine-learning algorithms using the data at Perdigão, and we find that the models eliminate the bias MYNN currently shows in representing ϵ, while also reducing the average error by up to almost 40 %. Of all the variables included in the algorithms, TKE is the variable responsible for most of the variability of ϵ, and a strong positive correlation exists between the two. These results suggest further consideration of machine-learning techniques to enhance parameterizations of turbulence in numerical weather prediction models.

58 GEOSCIENCES↗