SEARCH · Engineering Papers

Results for “synthetic forecast”

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.

SourceResearch connectionFrom yearThrough year

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

At least 37 records · Page 2

All sources NASA NTRS DOE OSTI

DOE OSTI2022

Tsunami Early Warning From Global Navigation Satellite System Data Using Convolutional Neural Networks

Abstract We investigate the potential of using Global Navigation Satellite System (GNSS) observations to directly forecast full tsunami waveforms in real time. We train convolutional neural networks to use less than 9 min of GNSS data to forecast the full tsunami waveforms over 6 hr at select locations, and obtain accurate forecasts on a test data set. Our training and test data consists of synthetic earthquakes and associated GNSS data generated for the Cascadia Subduction Zone using the MudPy software, and corresponding tsunami waveforms in Puget Sound computed using GeoClaw. We use the same suite of synthetic earthquakes and waveforms as in earlier work where tsunami waveforms were used for forecasting, and provide a comparison. We also explore varying the number of GNSS stations, their locations, and their observation durations.

Rim, Donsub↗

DOE OSTI2026

Integrated Approach to Ancillary PV Component Reliability Assessment (Final Report)

In this project, we have established a nondestructive, generalized methodology that (1) fuses rich field data with advanced ML for proactive reliability forecasting, (2) dramatically reduces experimental iterations via synthetic dataset generation, and (3) achieves unprecedented regression precision in both anomaly detection and component-level degradation assessment—paving the way for truly predictive maintenance of grid-tied PV inverters under diverse outdoor conditions.

14 SOLAR ENERGY↗

DOE OSTI2023

Stochastic Price Generation for Evaluating Wholesale Electricity Market Bidding Strategies

This work presents a novel method for generating electricity price scenarios from statistical properties of past electricity prices using a hybrid statistical and reduced-form stochastic model. Previous work in applying stochastic differential equations (SDE) to model electricity prices has focused on daily average prices. To extend stochastic price generation methods to hourly or sub-hourly pricing, we address several weaknesses in the state-of-the-art: (1) we replace the mean-reversion component of the SDE with an ARIMA process that is better able to characterize the daily and weekly trends; (2) we extend the price-spike, or jump process to account for conditional probabilities of price spikes occurring in consecutive time steps by replacing the traditional Poisson process for modeling jumps with a generalized point process model inspired by brain neuron models; and (3) we replace the traditional method of estimating spike intensity with empirical variance with a Markov process based on observed price spike intensity transitions. The method is demonstrated with electricity prices from the US ERCOT market and a use-case example is provided for bidding an energy storage unit into the day-ahead and real-time energy markets of ERCOT using stochastic optimization methods. Results show that the the synthetic price model out performs a (naive) persistence forecast model by resulting in 24% to 47% more in profits over 168 simulated days.

29 ENERGY PLANNING, POLICY, AND ECONOMY↗

NASA NTRS2013

The Influence of Observation Errors on Analysis Error and Forecast Skill Investigated with an Observing System Simulation Experiment

The Global Modeling and Assimilation Office (GMAO) observing system simulation experiment (OSSE) framework is used to explore the response of analysis error and forecast skill to observation quality. In an OSSE, synthetic observations may be created that have much smaller error than real observations, and precisely quantified error may be applied to these synthetic observations. Three experiments are performed in which synthetic observations with magnitudes of applied observation error that vary from zero to twice the estimated realistic error are ingested into the Goddard Earth Observing System Model (GEOS-5) with Gridpoint Statistical Interpolation (GSI) data assimilation for a one-month period representing July. The analysis increment and observation innovation are strongly impacted by observation error, with much larger variances for increased observation error. The analysis quality is degraded by increased observation error, but the change in root-mean-square error of the analysis state is small relative to the total analysis error. Surprisingly, in the 120 hour forecast increased observation error only yields a slight decline in forecast skill in the extratropics, and no discernable degradation of forecast skill in the tropics.

Observation Error↗

NASA NTRS2000

A Method for Assessing the Quality of Model-Based Estimates of Ground Temperature and Atmospheric Moisture Using Satellite Data

A method is developed for validating model-based estimates of atmospheric moisture and ground temperature using satellite data. The approach relates errors in estimates of clear-sky longwave fluxes at the top of the Earth-atmosphere system to errors in geophysical parameters. The fluxes include clear-sky outgoing longwave radiation (CLR) and radiative flux in the window region between 8 and 12 microns (RadWn). The approach capitalizes on the availability of satellite estimates of CLR and RadWn and other auxiliary satellite data, and multiple global four-dimensional data assimilation (4-DDA) products. The basic methodology employs off-line forward radiative transfer calculations to generate synthetic clear-sky longwave fluxes from two different 4-DDA data sets. Simple linear regression is used to relate the clear-sky longwave flux discrepancies to discrepancies in ground temperature ((delta)T(sub g)) and broad-layer integrated atmospheric precipitable water ((delta)pw). The slopes of the regression lines define sensitivity parameters which can be exploited to help interpret mismatches between satellite observations and model-based estimates of clear-sky longwave fluxes. For illustration we analyze the discrepancies in the clear-sky longwave fluxes between an early implementation of the Goddard Earth Observing System Data Assimilation System (GEOS2) and a recent operational version of the European Centre for Medium-Range Weather Forecasts data assimilation system. The analysis of the synthetic clear-sky flux data shows that simple linear regression employing (delta)T(sub g)) and broad layer (delta)pw provides a good approximation to the full radiative transfer calculations, typically explaining more thin 90% of the 6 hourly variance in the flux differences. These simple regression relations can be inverted to "retrieve" the errors in the geophysical parameters, Uncertainties (normalized by standard deviation) in the monthly mean retrieved parameters range from 7% for (delta)T(sub g) to approx. 20% for the lower tropospheric moisture between 500 hPa and surface. The regression relationships developed from the synthetic flux data, together with CLR and RadWn observed with the Clouds and Earth Radiant Energy System instrument, ire used to assess the quality of the GEOS2 T(sub g) and pw. Results showed that the GEOS2 T(sub g) is too cold over land, and pw in upper layers is too high over the tropical oceans and too low in the lower atmosphere.

Wu, Man Li C.↗

NASA NTRSDate not supplied

High resolution assimilation of multiple satellite retrievals with emissions adjustment to improve air quality forecasting with WRF-Chem/DART

We will present results from medium (15km, 6hr cycling) and high (4 km, 6 hr cycling) spatiotemporal resolution applications of the WRF-Chem/DART ensemble, regional, air quality (AQ) forecast/assimilation system.The medium-resolution setup is applied to the Discover AQ/Front Range Air Pollution and Photochemistry Experiment (FRAPPE) domain from July 14 to July 29, 2014. The high-resolution setup is applied to a Colorado domain from July 14 to July 29, 2020. For the FRAPPE application, we assimilate MOPITT CO; IASI CO;MODIS AOD; OMI O3, NO2; and AirNow CO, O3, NO2, SO2, PM10, and PM2.5. For the Colorado application, we assimilate the same MOPITT, MODIS, and AirNow constituents as in the FRAPPE application and TROPOMI CO, O3, NO2; and synthetic TEMPO O3and NO2. WRF-Chem/DART integrates the Weather Research and Forecast (WRF) model with on-line chemistry (WRF-Chem) into the Data Assimilation Research Testbed (DART). It assimilates AirNow CO, O3, NO2, SO2, PM10, and PM2.5 measurements, MOPITT CO; IASI CO, O3; OMI O3, NO2, SO2; TROPOMI CO, O3, NO2, SO2; MODIS AOD; and synthetic TEMPO O3 and NO2 total/partial column and/or profile retrievals.WRF-Chem/DART uses: (i) the state augmentation method for adjusting emissions; (ii) state-space localization; and (iii) a near-real time scripting system. We use the medium-resolution FRAPPE application to demonstrate the incremental benefits from assimilating OMI observations with emissions adjustment and the high-resolution Colorado application to demonstrate the incremental benefits from assimilating syntheticTEMPO observations with emissions adjustment. For both applications, we expect that: (i) assimilating chemical observations will increaseAQ forecast skill; (ii) including emissions adjustment will increase forecast skill/predictability; and (iii) including assimilation of synthetic TEMPO observations will further increase forecast skill/predictability.

High resolution↗

NASA NTRS2014

Status of the NASA GMAO Observing System Simulation Experiment

An Observing System Simulation Experiment (OSSE) is a pure modeling study used when actual observations are too expensive or difficult to obtain. OSSEs are valuable tools for determining the potential impact of new observing systems on numerical weather forecasts and for evaluation of data assimilation systems (DAS). An OSSE has been developed at the NASA Global Modeling and Assimilation Office (GMAO, Errico et al 2013). The GMAO OSSE uses a 13-month integration of the European Centre for Medium- Range Weather Forecasts 2005 operational model at T511/L91 resolution for the Nature Run (NR). Synthetic observations have been updated so that they are based on real observations during the summer of 2013. The emulated observation types include AMSU-A, MHS, IASI, AIRS, and HIRS4 radiance data, GPS-RO, and conventional types including aircraft, rawinsonde, profiler, surface, and satellite winds. The synthetic satellite wind observations are colocated with the NR cloud fields, and the rawinsondes are advected during ascent using the NR wind fields. Data counts for the synthetic observations are matched as closely as possible to real data counts, as shown in Figure 2. Errors are added to the synthetic observations to emulate representativeness and instrument errors. The synthetic errors are calibrated so that the statistics of observation innovation and analysis increments in the OSSE are similar to the same statistics for assimilation of real observations, in an iterative method described by Errico et al (2013). The standard deviations of observation minus forecast (xo-H(xb)) are compared for the OSSE and real data in Figure 3. The synthetic errors include both random, uncorrelated errors, and an additional correlated error component for some observational types. Vertically correlated errors are included for conventional sounding data and GPS-RO, and channel correlated errors are introduced to AIRS and IASI (Figure 4). HIRS, AMSU-A, and MHS have a component of horizontally correlated error. The forecast model used by the GMAO OSSE is the Goddard Earth Observing System Model, Version 5 (GEOS-5) with Gridpoint Statistical Interpolation (GSI) DAS. The model version has been updated to v. 5.13.3, corresponding to the current operational model. Forecasts are run on a cube-sphere grid with 180 points along each edge of the cube (approximately 0.5 degree horizontal resolution) with 72 vertical levels. The DAS is cycled at 6-hour intervals, with 240 hour forecasts launched daily at 0000 UTC. Evaluation of the forecasting skill for July and August is currently underway. Prior versions of the GMAO OSSE have been found to have greater forecasting skill than real world forecasts. It is anticipated that similar forecast skill will be found in the updated OSSE.

OSSE↗

DOE OSTI2023

BuildingsBench: A Benchmark for Universal Building Load Forecasting [SWR-23-51]

The residential and commercial building stock in the United States is responsible for a significant percentage of energy consumption and greenhouse gas emissions. Electrification of end-uses, as well as decarbonizing the electrical grid through renewable energy sources such as solar and wind, constitutes the pathway to zero-emission buildings. Forecasting day-ahead building energy consumption is an integral part of this solution. Currently, specialized forecasting models are hand-made for each individual building, which is time-consuming, expensive, and leads to duplicated efforts. BuildingsBench is a Python software framework for training and comparing generalized machine learning models for universal building load forecasting. This challenge tasks a single foundational model to generalize its forecasts for a wide variety of buildings, across geographic regions, building types, weather patterns, and more. This software provide code for pre-training such models and subsequently evaluating their performance on a suite of hundreds of diverse real and synthetic buildings. BuildingsBench is a platform for: - Large-scale pretraining with the synthetic Buildings-900K dataset for short-term load forecasting (STLF). Buildings-900K is statistically representative of the entire U.S. building stock and is extracted from the NREL End-Use Load Profiles database. - Benchmarking on two tasks evaluating generalization: zero-shot STLF and transfer learning for STLF. We provide an index-based PyTorch Dataset for large-scale pretraining, easy data loading for multiple real building energy consumption datasets as PyTorch Tensors or Pandas DataFrames, simple (persistence) to advanced (transformer) baselines, metrics management, and more.

Emami, Patrick↗

DOE OSTI2021

State-of-the-Art Techniques for Large-Scale Stochastic Unit Commitment

Recent advances in deterministic unit commitment, both formulaic and algorithmic, along with modern algorithmic approaches for stochastic programming, have enabled the solution of stochastic unit commitment problems with hundreds of scenarios on large-scale transmission networks. In this presentation, we will give an overview of these methods, including lazy transmission constraint generation, lower-bounding techniques, and heuristics, all of which can be executed in concert with customized decomposition approaches for optimization under uncertainty. We demonstrate the effectiveness of these techniques on the TAMU Texas7K synthetic transmission network, leveraging realistic high-resolution forecasts based on NREL renewable resource availability data. The software leveraged for these demonstrations is available via the open-source software packages EGRET (for electrical grid optimization) and mpi-sppy (for optimization under uncertainty).

27 ARPA - Advanced Research Projects Agency-Energy↗

NASA NTRS2022

Quantifying Errors in 3D CME Parameters Derived from Synthetic Data Using White-Light Reconstruction Techniques

Current efforts in space weather forecasting of CMEs have been focused on predicting their arrival time and magnetic structure. To make these predictions, methods have been developed to derive the true CME speed, size, position, and mass, among others. Difficulties in determining the input parameters for CME forecasting models arise from the lack of direct measurements of the coronal magnetic fields and uncertainties in estimating the CME 3D geometric and kinematic parameters after eruption. White-light coronagraph images are usually employed by a variety of CME reconstruction techniques that assume more or less complex geometries. This is the first study from our International Space Science Institute (ISSI) team “Understanding Our Capabilities in Observing and Modeling Coronal Mass Ejections”, in which we explore how subjectivity affects the 3D CME parameters that are obtained from the Graduated Cylindrical Shell (GCS) reconstruction technique, which is widely used in CME research. To be able to quantify such uncertainties, the “true” values that are being fitted should be known, which are impossible to derive from observational data. We have designed two different synthetic scenarios where the “true” geometric parameters are known in order to quantify such uncertainties for the first time. We explore this by using two sets of synthetic data: 1) Using the ray-tracing option from the GCS model software itself, and 2) Using 3D magnetohydrodynamic (MHD) simulation data from the Magnetohydrodynamic Algorithm outside a Sphere code. Our experiment includes different viewing configurations using single and multiple viewpoints. CME reconstructions using a single viewpoint had the largest errors and error ranges overall for both synthetic GCS and simulated MHD white-light data. As the number of viewpoints increased from one to two, the errors decreased by approximately 4° in latitude, 22° in longitude, 14° in tilt, and 10° in half-angle. Our results quantitatively show the critical need for at least two viewpoints to be able to reduce the uncertainty in deriving CME parameters. We did not find a significant decrease in errors when going from two to three viewpoints for our specific hypothetical three spacecraft scenario using synthetic GCS white-light data. As we expected, considering all configurations and numbers of viewpoints, the mean absolute errors in the measured CME parameters are generally significantly higher in the case of the simulated MHD white-light data compared to those from the synthetic white-light images generated by the GCS model. We found the following CME parameter error bars as a starting point for quantifying the minimum error in CME parameters from white-light reconstructions: Δθ (latitude)=6° +2° -3° , Δϕ (longitude)=11° +18° -6° , Δγ (tilt)=25° +8° -7° , Δx (half-angle)=10° +12° -6° , Δh (height)=0.6 +1.2 -0.4 R ⨀ , and Δκ (ratio)=0.1 +0.03 -0.02 .

Coronal mass ejections↗

NASA NTRS2022

Quantifying Errors in 3D CME Parameters Derived From Synthetic Data Using White-Light Reconstruction Techniques

Coronal mass ejections↗

DOE OSTI2022

A multi-scale time-series dataset with benchmark for machine learning in decarbonized energy grids

The electric grid is a key enabling infrastructure for the ambitious transition towards carbon neutrality as we grapple with climate change. With deepening penetration of renewable resources, the reliable operation of the electric grid becomes increasingly challenging. In this paper, we present PSML, a first-of-its-kind open-access multi-scale time-series dataset, to aid in the development of data-driven machine learning (ML)-based approaches towards reliable operation of future electric grids. The dataset is synthesized from a joint transmission and distribution electric grid to capture the increasingly important interactions and uncertainties of the grid dynamics, containing power, voltage and current measurements over multiple spatio-temporal scales. Using PSML, we provide state-of-the-art ML benchmarks on three challenging use cases of critical importance to achieve: (i) early detection, accurate classification and localization of dynamic disturbances; (ii) robust hierarchical forecasting of load and renewable energy; and (iii) realistic synthetic generation of physical-law-constrained measurements. We envision that this dataset will provide use-inspired ML research in safety-critical systems, while simultaneously enabling ML researchers to contribute towards decarbonization of energy sectors.

54 ENVIRONMENTAL SCIENCES↗

NASA NTRS1992

A snow wetness retrieval algorithm for SAR

The objectives of this study are: (1) to evaluate the backscattering signals response to snow wetness; and (2) to develop an algorithm for snow wetness measurement using C-band polarimetric synthetic aperture radar (SAR). In hydrological investigations, modeling and forecasting of snowmelt runoff requires information about snowpack properties and their spatial variability. In particular, timely measurement of snow parameters is needed for operational hydrology. The liquid water content of snowpack is one of the important parameters. Active microwave sensors are highly sensitive to liquid water in the snowpack because of the large dielectric contrast between ice and water in the microwave spectrum. They are not affected by weather and have a spatial resolution compatible with the topographic variation in alpine regions. However, a quantitative algorithm for retrieval snow wetness has not yet been developed.

Shi, Jian-Cheng↗

DOE OSTI2022

Device Classification for Industrial Control Systems Using Predicted Traffic Features

To achieve a secure interconnected Industrial Control System (ICS) architecture, security practitioners depend on accurate identification of network host behavior. However, accurate machine learning based host identification methods depends on the availability of significant quantities of network traffic data, which can be difficult to obtain due to system constraints such as network security, data confidentiality, and physical location. In this work, we propose a network traffic feature prediction method based on a generative model, which achieves high host identification accuracy. Furthermore, we develop a joint training algorithm to improve host identification performance compared to separate training of the generative model and the classifier responsible for host identification.

97 MATHEMATICS AND COMPUTING↗

NASA NTRS2013

Multi-sensor Cloud Retrieval Simulator and Remote Sensing from Model Parameters : Synthetic Sensor Radiance Formulation - Pt. 1

In this paper we describe a general procedure for calculating synthetic sensor radiances from variable output from a global atmospheric forecast model. In order to take proper account of the discrepancies between model resolution and sensor footprint, the algorithm takes explicit account of the model subgrid variability, in particular its description of the probability density function of total water (vapor and cloud condensate.) The simulated sensor radiances are then substituted into an operational remote sensing algorithm processing chain to produce a variety of remote sensing products that would normally be produced from actual sensor output. This output can then be used for a wide variety of purposes such as model parameter verification, remote sensing algorithm validation, testing of new retrieval methods and future sensor studies.We show a specific implementation using the GEOS-5 model, the MODIS instrument and the MODIS Adaptive Processing System (MODAPS) Data Collection 5.1 operational remote sensing cloud algorithm processing chain (including the cloud mask, cloud top properties and cloud optical and microphysical properties products). We focus on clouds because they are very important to model development and improvement.

Simulations↗

NASA NTRS2011

Prediction of Hydrological Drought: What Can We Learn From Continental-Scale Offline Simulations?

Land surface model experiments are used to quantify, across the coterminous United States, the contributions (isolated and combined) of soil moisture and snowpack initialization to the skill of seasonal streamflow forecasts at multiple leads and for different start dates. Forecasted streamflows are compared to naturalized streamflow observations where available and to synthetic (model-generated) streamflow data elsewhere. We find that snow initialization has a major impact on skill in the mountainous western U.S. and in a portion of the northern Great Plains; a mid-winter (January 1) initialization of snow in these areas leads to significant skill in the spring melting season. Soil moisture initialization also contributes to skill, and although the maximum contributions are not as large as those seen for snow initialization, the soil moisture contributions extend across a much broader geographical area. Soil moisture initialization can contribute to skill at long leads (up to 5 or 6 months), particularly for forecasts issued during winter.

Koster, Randal↗

DOE OSTI2022

Multi-Fidelity Stochastic Economic Dispatch for Operating Low Carbon Power Grids

Grid operators can address the inherently stochastic nature of renewables by solving a two-stage stochastic programming model that minimizes the cost of dispatch decisions while accounting for the complex grid dynamics. It is common to use a sample average approximation to estimate the expectation of the second stage costs in this model. However, the large sample count needed for numerical accuracy makes effective modeling large-scale electric grids computationally intractable. We introduce a control variate multi-fidelity estimator for the second-stage recourse that enables high quality dispatch decisions in real-time with a reduced computational burden. We obtain a hierarchy of model fidelities by linearizing the AC power flow system representation to DC power flow, and by relaxing transmission and voltage network constraints. We evaluate the performance of our proposed method on a synthetic grid with 73 buses against a deterministic baseline with persistence forecast and a high-fidelity reference. Our analysis shows a computational speed-up of 7.62x with a minimal loss in accuracy. The multi-fidelity method is well suited to fidelity combinations that use a simplified network topology in their lower fidelity model and is an attractive option for applications where accurate grid modeling needed on a limited computational budget.

24 POWER TRANSMISSION AND DISTRIBUTION↗

DOE OSTI2024

From RNNs to Foundation Models: An Empirical Study on Commercial Building Energy Consumption

Accurate short-term energy consumption forecasting for commercial buildings is crucial for smart grid operations. While smart meters and deep learning models enable forecasting using past data from multiple buildings, data heterogeneity from diverse buildings can reduce model performance. The impact of increasing dataset heterogeneity in time series forecasting, while keeping size and model constant, is understudied. We tackle this issue using the ComStock dataset, which provides synthetic energy consumption data for U.S. commercial buildings. Two curated subsets, identical in size and region but differing in building type diversity, are used to assess the performance of various time series forecasting models, including finetuned open-source foundation models (FMs). The results show that dataset heterogeneity and model architecture have a greater impact on post-training forecasting performance than the parameter count. Moreover, despite the higher computational cost, finetuned FMs demonstrate competitive performance compared to base models trained from scratch.

commercial buildings↗