Search NASA⌕ Search

SEARCH · Search NASA

Results for “RMSE”

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 127 records · Page 7
All sourcesNASA NTRSDOE OSTI
DOE OSTI2024

Artificial ground reflector size and position effects on energy yield and economics of single‐axis‐tracked bifacial photovoltaics

Abstract Artificial ground reflectors improve bifacial energy yield by increasing both front and rear‐incident irradiance. Studies have demonstrated an increase in energy yield due to the addition of artificial reflectors; however, they have not addressed the effect of varying reflector dimensions and placement on system performance and the impact of these parameters on the reflectors' financial viability. We studied the effect of high albedo (70% reflective) artificial reflectors on single‐axis‐tracked bifacial photovoltaic systems through ray‐trace modeling and field measurements. In the field, we tested a range of reflector configurations by varying reflector size and placement and demonstrated that reflectors increased daily energy yield up to 6.2% relative to natural albedo for PERC modules. To confirm the accuracy of our model, we compared modeled and measured power and found a root mean square error (RMSE) of 5.4% on an hourly basis. We modeled a typical meteorological year in Golden, Colorado, to demonstrate the effects of artificial reflectors under a wide range of operating conditions. Seventy percent reflective material can increase total incident irradiance by 1.9%–8.6% and total energy yield by 0.9%–4.5% annually after clipping is considered with a DC–AC ratio of 1.2. Clipping has a significant effect on reflector impact and must be included when assessing reflector viability because it reduces reflector energy gain. We calculated a maximum viable cost for these improvements of up to $2.50–4.60/m 2 , including both material and installation, in Golden. We expanded our analysis to cover a latitude range of 32–48°N and demonstrated that higher‐latitude installations with lower energy yield and higher diffuse irradiance content can support higher reflector costs. In both modeling and field tests, and for all locations, the ideal placement of the reflectors was found to be directly underneath the module due to the optimized rear irradiance increase.

14 SOLAR ENERGY↗
DOE OSTI2024

A Novel Non‐Destructive Rapid Tool for Estimating Amino Acid Composition and Secondary Structures of Proteins in Solution

Abstract Amino‐acid protein composition plays an important role in biology, medicine, and nutrition. Here, a groundbreaking protein analysis technique that quickly estimates amino acid composition and secondary structure across various protein sizes, while maintaining their natural states is introduced and validated. This method combines multivariate statistics and the thermostable Raman interaction profiling (TRIP) technique, eliminating the need for complex preparations. In order to validate the approach, the Raman spectra are constructed of seven proteins of varying sizes by utilizing their amino acid frequencies and the Raman spectra of individual amino acids. These constructed spectra exhibit a close resemblance to the actual measured Raman spectra. Specific vibrational modes tied to free amino and carboxyl termini of the amino acids disappear as signals linked to secondary structures emerged under TRIP conditions. Furthermore, the technique is used inversely to successfully estimate amino acid compositions and secondary structures of unknown proteins across a range of sizes, achieving impressive accuracy ranging between 1.47% and 5.77% of root mean square errors (RMSE). These results extend the uses for TRIP beyond interaction profiling, to probe amino acid composition and structure.

Chemistry↗
DOE OSTI2022

Estimation of the Surface Fluxes for Heat and Momentum in Unstable Conditions with Machine Learning and Similarity Approaches for the LAFE Data Set

Abstract Measurements of three flux towers operated during the land atmosphere feedback experiment (LAFE) are used to investigate relationships between surface fluxes and variables of the land–atmosphere system. We study these relations by means of two machine learning (ML) techniques: multilayer perceptrons (MLP) and extreme gradient boosting (XGB). We compare their flux derivation performance with Monin–Obukhov similarity theory (MOST) and a similarity relationship using the bulk Richardson number (BRN). The ML approaches outperform MOST and BRN. Best agreement with the observations is achieved for the friction velocity. For the sensible heat flux and even more so for the latent heat flux, MOST and BRN deviate from the observations while MLP and XGB yield more accurate predictions. Using MOST and BRN for latent heat flux, the root mean square errors (RMSE) are 107 Wm $$^{-2}$$ - 2 and 121 Wm $$^{-2}$$ - 2 , respectively, as well as the intercepts of the regression lines are $$\approx 110$$ ≈ 110 Wm $$^{-2}$$ - 2 . For the ML methods, the RMSEs reduce to 31 Wm $$^{-2}$$ - 2 for MLP and 33 Wm $$^{-2}$$ - 2 for XGB as well as the intercepts to just 4 Wm $$^{-2}$$ - 2 for MLP and $$-1$$ - 1 Wm $$^{-2}$$ - 2 for XGB with slopes of the regression lines close to 1, respectively. These results indicate significant deficiencies of MOST and BRN, particularly for the derivation of the latent heat flux. In fact, in contrast to the established theories, feature importance weighting demonstrates that the ML methods base their improved derivations on net radiation, the incoming and outgoing shortwave radiations, the air temperature gradient, and the available water contents, but not on the water vapor gradient. The results imply that further studies of surface fluxes and other turbulent variables with ML techniques provide great promise for deriving advanced flux parameterizations and their implementation in land–atmosphere system models.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2023

Robust deep learning framework for constitutive relations modeling

Modeling the full-range deformation behaviors of materials under complex loading and materials conditions is a significant challenge for constitutive relations (CRs) modeling. Here, we propose a general encoder-decoder deep learning framework that can model high-dimensional stress-strain data and complex loading histories with robustness and universal capability. The framework employs an encoder to project high-dimensional input information (e.g., loading history, loading conditions, and materials information) to a lower-dimensional hidden space and a decoder to map the hidden representation to the stress of interest. We evaluated various encoder architectures, including gated recurrent unit (GRU), GRU with attention, temporal convolutional network (TCN), and the Transformer encoder, on two complex stress-strain datasets that were designed to include a wide range of complex loading histories and loading conditions. All architectures achieved excellent test results with an root-mean-square error (RMSE) below 1 MPa. Additionally, we analyzed the capability of the different architectures to make predictions on out-of-domain applications, with an uncertainty estimation based on deep ensembles. The proposed approach provides a robust alternative to empirical/semi-empirical models for CRs modeling, offering the potential for more accurate and efficient materials design and optimization.

36 MATERIALS SCIENCE↗
DOE OSTI2022

Site-specific field management adaptation is key to feeding the world in the 21st century

Rapid climate change and growing population threaten global food security across the globe. Several studies have proposed early planting, increased irrigation, and increased fertilizer applications as climate adaptation strategies, yet none have considered combined and site-specific field management strategies as a comprehensive solution. Here, we analyzed non-irrigated wheat yield responses to climate change and field management adaptation using a mechanistic crop model evaluated against observed global non-irrigated wheat-yield over 3-year intervals spanning 13 years at 3749 sites (RMSE = 36 gC m -2 ). Early planting with later-maturing varieties provided the most benefit to future yields among the proposed field management adaptation strategies. Improved water use efficiency from increased CO 2 led to relatively low benefits of additional irrigation. Here we estimated that spatially heterogeneous adaption strategies had the potential to improve global wheat yields by 91% by 2100 compared to the present day. The yield improvements from combined field adaptation strategies were larger than the sum of improvements from the individual strategies. These synergistic benefits were shown to result from complementary processes regulating nutrient and water uptake, physiological tolerance to heat stress, and internal carbon and nutrient cycling.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2023

Can upscaling ground nadir SIF to eddy covariance footprint improve the relationship between SIF and GPP in croplands?

Ground solar-induced chlorophyll fluorescence (SIF) is important for the mechanistic understanding of the dynamics of vegetation gross primary production (GPP) at fine spatiotemporal scales. However, eddy covariance (EC) observations generally cover larger footprint areas than ground SIF observations (a bare fiber with nadir), and this footprint mismatch between nadir SIF and GPP could complicate the canopy SIF-GPP relationships. Here, we upscaled nadir SIF observations to EC footprint and investigated the change in SIF-GPP relationships after the upscaling in cropland. We included 13 site-years data in our study, with seven site-years corn, four site-years soybeans, and two site-years miscanthus, all located in the US Corn Belt. All sites’ crop nadir SIF observations collected from the automated FluoSpec2 system (a hemispheric-nadir system) were upscaled to the GPP footprint-based SIF using vegetation indices (VIs) calculated from high spatiotemporal satellite reflectance data. We found that SIF-GPP relationships were not substantially changed after upscaling nadir SIF to GPP footprint at our crop sites planted with corn, soybean, and miscanthus, with R 2 change after the upscaling ranging from -0.007 to 0.051 and root mean square error (RMSE) difference from -0.658 to 0.095 umol m –2 s –1 relative to original nadir SIF-GPP relationships across all the site-years. The variation of the SIF-GPP relationship within each species across different site-years was similar between the original nadir SIF and upscaled SIF. Different VIs, EC footprint models, and satellite data led to marginal differences in the SIF-GPP relationships when upscaling nadir SIF to EC footprint. Furthermore, our study provided a methodological framework to correct this spatial mismatch between ground nadir SIF and GPP observations for croplands and potentially for other ecosystems. Our results also demonstrated that the spatial mismatch between ground nadir SIF and GPP might not significantly affect the SIF-GPP relationship in cropland that are largely homogeneous.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2023

SIF-based GPP modeling for evergreen forests considering the seasonal variation in maximum photochemical efficiency

Solar-induced chlorophyll fluorescence (SIF) has shown great potential in estimating gross primary production (GPP). However, their quantitative relationship is not invariant, which undermines the reliability of empirical SIF-based GPP estimation at fine spatiotemporal scales, especially under extreme conditions. In this study, we developed a parsimonious mechanistic model for SIF-based GPP estimation in evergreen needle forests (ENF) by employing the Mechanistic Light Response framework and Eco-Evolutionary theory to describe the light and dark reactions during photosynthesis, respectively. Specifically, we found that considering the seasonal variation in a key parameter of the MLR framework, the maximum photochemical efficiency of photosystem II (Φ PSIImax ), can avoid the GPP overestimation in winter and early spring due to the relatively low environmental sensitivity of SIF. Compared to the estimates from other benchmark models, our GPP estimates were closer to the 1: 1 line and had higher accuracy (average R 2 = 0.86, RMSE=1.99 μmol m -2 s -1 ) across sites. Furthermore, the changes in the relationship between SIF and J (refers to the electron transport rate) contribute a lot to the dynamic SIF–GPP relationship in this study, while the J–GPP relationship is less variant when the temperature drops. Further, the seasonal variation in the SIF–J relationship, especially the reduction in its slope at low temperatures, is found largely explained by the Φ PSIImax . These results indicate the importance of the uncertainty caused by the variation in the SIF–J relationship for SIF-based GPP estimation, and the consideration of changes in Φ PSIImax under extreme conditions (such as severe winter in this study) is crucial for the improvement of GPP estimation via SIF.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2026

Maximizing machine learning interatomic potential transferability for the discovery of the novel stellated octadecagon Bi18-Pt24 cage structure

Achieving true transferability remains the central challenge for Machine Learning Interatomic Potentials (ML-IAPs) in modeling complex bimetallic nanoclusters across their vast potential energy surfaces. We systematically investigate data selection strategies to optimize the Chebyshev Interaction Model for Efficient Simulation (ChIMES) potential for the Bi-Pt nanoclusters by comparing three innovative sampling methods: Principal Component Analysis (PCA)/k-means (structural diversity), t-distributedStochasticNeighborEmbedding (t-SNE)/k-means (force-space diversity), and hierarchical clustering. Quantitatively, the PCA/k-means strategy proved most effective for global accuracy, yielding the lowest force errors and achieving energy root mean square errors (RMSE) values competitive with Density Functional Theory (DFT), demonstrating excellent accuracy (19.16meV/atom). Structural validation on 34 unique DFT-optimized isomers further confirmed the potential’s high fidelity, with the best model PCA/k-means reproducing structures with an average root mean square deviation (RMSD) of 0.10 Å. However, the t-SNE methods, by maximizing diversity in the force space, demonstrated superior extrapolative power, leading to the more precise prediction of a novel stellated octadecagon Bi18⁢Pt24 cage structure, demonstrating the potential for exploring previously unseen morphologies. Our results establish a clear methodology for strategic data sampling that successfully maximizes ML-IAP transferability, providing an accurate and computationally efficient tool that accelerates the theoretical discovery of complex bimetallic architectures.

Vangheluwe, Raphaël [Université Paris-Saclay, CNRS↗
DOE OSTI2025

Transfer learning-based soybean LAI estimations by integrating PROSAIL, UAV, and PlanetScope imagery

Accurate Leaf Area Index (LAI) estimations at the soybean plot scale is achievable using high-resolution Unmanned Aerial Vehicle (UAV) imagery and field measurement samples. However, the limited coverage of UAV flights restricts large-scale remote sensing monitoring in expansive soybean fields. This study leverages the broad coverage and 3-m resolution of PlanetScope satellite imagery to extend LAI prediction from UAV to satellite scales through transfer learning, using UAV-scale LAI estimates as a benchmark to validate cross-scale consistency. To address this challenge, this study proposed the LAI-TransNet, a two-stage transfer learning framework designed for precise and scalable soybean LAI prediction across large areas, demonstrating its effectiveness in cross-scale monitoring. In Stage 1, a UAV-scale benchmark is established using PROSAIL-simulated UAV reflectance data (UAV-Sim) and field-measured soybean LAI. Traditional machine learning, deep learning, and transfer learning models are trained on a hybrid UAV-Sim and field-measured dataset (UAV-Sim_Measured), with the transfer learning model CNN-TL, fine-tuned using pre-trained weights derived from UAV-Sim, achieving the highest accuracy (R 2 = 0.81, RMSE = 0.64 m 2 /m 2 , rRMSE = 11.5 %). In Stage 2, LAI-TransNet is developed by fine-tuning the CNN-TL model on PlanetScope simulated data (PS-Sim), preprocessed via cross-domain mapping to align UAV and satellite spectral features. Real PlanetScope imagery is corrected for reflectance consistency with reference to UAV imagery spectral profiles. LAI-TransNet outperforms other deep learning models trained directly on PS-Sim (R 2 = 0.69 vs. 0.60–0.63), ensuring robust cross-scale consistency. In conclusion, by bridging UAV and satellite scales, LAI-TransNet enables large-scale soybean LAI monitoring, enhancing precision agriculture management through improved monitoring with the PlanetScope imagery.

Leaf area index (LAI)↗
DOE OSTI2024

Maximizing energy efficiency in wastewater treatment plants: A data-driven approach for waste heat recovery and an economic analysis using Organic Rankine Cycle and thermal energy storage

Maximizing energy efficiency through waste heat recovery (WHR) processes is crucial for sustainable and ecofriendly operations across multiple industries, notably in wastewater treatment plants (WWTPs). This work proposes a comprehensive approach for assessing the WHR feasibility in WWTPs, structured in two main objectives. Firstly, an Artificial Neural Network (ANN) model is developed to accurately predict WHR based on operational data, including biogas temperature, biogas pressure, daily production in kWh, and WHR values in kWh th . The second objective focuses on economically evaluating the WHR feasibility based on the estimated WHR values obtained by the ANN model, and then realistically assessing the economic feasibility of integrating the Organic Rankine Cycle (ORC) and Seasonal Thermal Energy Storage (STES) systems. With an application to the As-Samra WWTP located in Jordan, the developed ANN model demonstrates promising results in the validation phase, with a root mean square error (RMSE) of 2206 kWh/day, a mean absolute error (MAE) of 1674 kWh/day, and an R-squared (R 2 ) value of 68%. On the other hand, the economic analysis reveals that an optimal ORC system of 412.14 kW e capacity yields a Net Present Value (NPV) of 2.09 million US dollars, a Levelized Cost of Energy (LCOE) of 0.0749 USD/kWh, a Payback Period (PBP) of 4.8 years, and annual revenues of 428 kUSD. This work also investigates the techno-economic feasibility of integrating ORC-STES. Results indicate that the LCOE and PBP are highly affected by the ORC's capital cost, and integrating STES increases the LCOE to 0.0824 USD/kWh, rendering its integration with ORC infeasible. In conclusion, this study aims to advance the understanding and application of WHR in WWTPs, paving the way for more efficient and sustainable practices in the field.

29 ENERGY PLANNING, POLICY, AND ECONOMY↗
DOE OSTI2025

Mapping tree height in complex terrain of northern China using ultra-high-resolution images

Tree height is a key parameter for estimating forest biomass and carbon sequestration. In recent years, notable progress has been made in mapping tree height using satellite imagery. However, existing tree height products show low accuracy in mountainous and complex terrains, and few studies typically addressed tree height estimations in mountain areas. This study examines the Mentougou district of Beijing, China, characterized by complex terrain and mountainous landscapes. We analyzed two methods for estimating tree height: one using only spectral features and another combining spectral features with topographic factors (elevation, slope, aspect). We used 3-m resolution PlanetScope 8-band multispectral imagery, with 710 field-measured individual tree heights averaged to obtain 471 pixel-level tree height values as ground-truth, to develop tree height prediction models using eXtreme Gradient Boosting (XGBoost), Random Forest (RF), and Gradient Boosting Machine (GBM) models. The results show that the XGBoost model consistently presented the highest accuracy for both methods evaluated. Specifically, the XGBoost model that combined spectral data with elevation and slope variables with an R² of 0.75 and an RMSE of 2.69 m. Using the XGBoost model, we generated the tree height map for the Mentougou area at 3 m resolution, showing tree heights ranging from 0.5 to 30.4 m, and the model’s prediction error standard deviations ranged from 2.50 to 4.71 m, indicating reliable performance across varied terrain. Additionally, we compared and evaluated the global tree height products, identifying limitations in the accuracy within complex terrains. This study demonstrates the potential for accurately predicting tree heights by combining high-resolution multispectral satellites with a terrain factor modeling approach.

Complex terrain↗
DOE OSTI2025

Uncertainty quantification and optimization of precipitating hydrometeor parameters for winter precipitation in a cloud microphysics scheme

The precipitating hydrometeor parameters used in cloud microphysics schemes carry inherent uncertainties. The quantification of these uncertainties, together with parameter optimization, can significantly improve precipitation forecasts. This study investigates the effects of 13 parameters in the Weather Research and Forecasting (WRF) Double-Moment 6-class (WDM6) microphysics scheme, which define the hydrometeor characteristics such as fall velocity–diameter and mass–diameter relationships, as well as the shape parameter of the drop size distribution for precipitating particles such as rain, snow, and graupel on simulated winter precipitation. A comparison between the model's pre-defined parameters and observations from the International Collaborative Experiments for the PyeongChang 2018 Olympic and Paralympic winter games (ICE-POP 2018) field campaign reveals that the fall velocity–diameter relationship for rain, the mass–diameter relationships for snow and graupel, and the shape parameters for all precipitating particles in the WDM6 scheme deviate from the median values observed by the two-dimensional video disdrometer (2DVD). To quantify parameter sensitivities, a perturbed parameter ensemble (PPE) of 256 simulations was conducted within parameter ranges constrained by 2DVD observations for three winter precipitation cases. Bayesian optimization was then applied to identify parameter sets that minimized the root mean square error (RMSE) for each case, achieving reductions of up to 30.2 %. These results demonstrate that ensemble-based uncertainty quantification and parameter optimization can help identify key parameters and provide a pathway to improving precipitation simulation performance. In addition, measurement sites can be strategically selected based on regions that show high sensitivity to variations in hydrometeor characteristic parameters.

Bayesian optimization↗
DOE OSTI2024

Adsorption of terbium (III) on DGA and LN resins: Thermodynamics, isotherms, and kinetics

Two commercially available extraction chromatography (EXC) resins containing N,N,N’,N’-tetra-n-octyldiglycolamide (DGA Resin, Normal, 50 – 100 μm) and Bis(2-ethylhexyl) phosphate (LN Resin, 100 – 150 μm) were used as adsorbents to study fundamental adsorption properties such as thermodynamic values, equilibrium isotherms, and kinetic uptake models for terbium(III) adsorption. Weight distribution ratios (D w ) for terbium on DGA and LN resins were measured using a [ 160 Tb]Tb 3+ radiometric tracer in nitric acid as a function of acidity, temperature, initial analyte concentration, and equilibrium time. The D w values showed increasing binding affinity for DGA resin at high nitric acid concentrations and decreasing binding affinity for LN resins. Thermodynamic studies for DGA and LN resins revealed that the Gibbs free energy (ΔG) increased consistently with temperature. To model equilibrium data, increasingly higher parameter equilibrium isotherm models (Henry (1) < Langmuir, Freundlich (2) < Redlich-Peterson (3) < Fritz-Schluender (4)) were compared on their root mean squared errors (RMSE) and adjusted determination coefficients to determine the most applicable model. In all cases, the empirical four-parameter Fritz-Schluender isotherm demonstrated a superior fit. Similar comparisons for reaction-based kinetic models (Pseudo-first-order < Pseudo-second-order < Pseudo-n-order) revealed that the higher-order PNO model yielded a superior fit of kinetic data for both resins. Furthermore, in some cases, adsorption isotherms and kinetic models could also be modeled by a lower-order model with minimal change in error parameters. Weber-Morris plots revealed that two linear sections are observed for each resin, where the first linear segment is attributed to fast (film diffusion) adsorption of terbium, followed by slower intraparticle diffusion of terbium through the pores as the rate-limiting step. Based on the Weber-Morris plot, both film and intraparticle diffusion are involved in controlling the kinetic rate of adsorption for DGA and LN resins.

38 RADIATION CHEMISTRY, RADIOCHEMISTRY, AND NUCLEA↗
DOE OSTI2026

From observation to replication: machine-learning-driven quantification and replication of fine-scale fish kinematics and behavior

Long-term quantification of fish behavior is essential for aquatic ecology, wildlife telemetry, and biomechanical device development. However, the observation duration required to obtain reliable behavioral and kinematic metrics remains unclear, and few tools exist to physically reproduce natural swimming motion for controlled experimentation. We address these challenges by developing a generalizable framework that models behavioral reliability (Spearman–Brown reliability index) as a function of observation duration and derives metric-specific monitoring thresholds. Using juvenile white sturgeon (Acipenser transmontanus) as a case study, we demonstrate that the minimum duration needed for reliable estimates varies substantially across kinematic features: to exceed a reliability of 0.8, total distance traveled requires 12 days, average curvature (mm?¹) 15 days, tail-beat frequency (Hz) 8 days, and average speed (body length/s) 17 days. We further bridge digital analysis and physical testing by developing a hardware-in-the-loop simulator that reconstructs machine-learning-derived swimming kinematics with high fidelity (correlation coefficient 0.98–0.99, RMSE 1.22–1.27 mm over a 5-minute segment). This platform enables realistic, repeatable motion stimuli for evaluating aquatic sensing technologies and bio-integrated devices under controlled conditions. Together, these contributions provide a scalable approach for designing long-term behavioral studies and a data-driven connection between ecological observation and robotic experimentation.

Hwang, SungJoo↗
DOE OSTI2023

Machine learning based simultaneous control of air handling unit discharge air and condenser water temperatures set-point for minimized cooling energy in an office building

In this study, an artificial intelligence based real-time prediction and control model to optimize condenser water temperature and discharge air temperature (DAT) set-points in water-cooled air handling unit (AHU) system has been developed. EnergyPlus-MATLAB co-simulation has been conducted to analyze the developed model's effectiveness. Here, to develop artificial neural networks (ANN) model, embedded neural network objects in MATLAB was utilized. The developed model could decide an optimal temperature set-points based on outdoor air wet-bulb temperature to reflect the Korean climate context. As a result, the developed ANN prediction model showed the predictive performance of Cv(RMSE) of approximately 21%. Compared to the conventional fixed temperature algorithm, which fixes AHU DAT at 14°C and condenser water temperature at 32°C, the ANN based optimized control showed a 22% total cooling energy reduction. These results show that significant energy savings can be achieved by simultaneously controlling condenser water temperature and AHU DAT set-points considering Korean climatic characteristics using AI technologies such as ANN models.

24 POWER TRANSMISSION AND DISTRIBUTION↗
DOE OSTI2024

A deep learning-based Bayesian framework for high-resolution calibration of building energy models

Calibrating building energy models (BEMs), i.e., closing discrepancy between modeling and field measurements, is of significance to support its applications in building sustainability and resilience analysis. However, as being widely used in practice, current Bayesian calibration is mostly performed in low-resolution (annual or monthly), instead of high-resolution (hourly or sub-hourly), which is crucial to support emerging BEM applications, such as building-renewable energy integration (demand response) and smart control. This is attributable to the gaps in current Bayesian calibration process, including (1) difficulty in supporting reliable high-resolution calibration with over-parameterization and multi-solution issues, (2) inadequacy of meta-model to capture temporal building dynamics in high-resolution, and (3) excessive computational burdens of covariance matrix calculation in Bayesian inference. Therefore, to close these gaps, this research proposes a novel deep learning-based Bayesian calibration framework, involving pre-calibration mechanism, Long Short-Term Memory as surrogate models, and simplified covariance matrix calculation, to calibrate BEMs in high temporal resolution (i.e., hourly) with enhanced accuracy and computational efficiency. Finally, the case study demonstrates its effectiveness to match modeling outcomes with measurements and realize CV-RMSE of < 30 % and NMBE of < 6 % in hourly resolution, as well as a significant reduction of calibration time (by > 99 %, from > 600 h to ~ 1.5 h).

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗
DOE OSTI2026

Machine learning-enhanced hybrid modeling approach for better identification of a building thermal network model and improved prediction

The gray-box modeling approach, which uses a semi-physical thermal network model, has been widely used in building prediction applications, such as model predictive control (MPC). However, unmeasured disturbances, such as occupants, lighting, and in/exfiltration loads, make it challenging to apply this approach to practical buildings. In this word, we propose a hybrid modeling approach that integrates the gray-box model with a model for unmeasured disturbance. After reviewing several system identification approaches, we systematically designed the unmeasured disturbance model with a model selection process based on statistical tests to make it robust. We generated data based on the building model calibrated by real operational data and then trained the hybrid model for two different weather conditions. The hybrid model approach demonstrates an RMSE reduction of approximately 0.2–0.9 °C and 0.3–2 °C on 1-day ahead temperature prediction compared to the Conventional approach for mild (Berkeley, CA) and cold (Chicago, IL) climates, respectively. In addition, this approach was applied to experimental data obtained from the laboratory building to be used for the MPC application, showing superior prediction performances.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗
DOE OSTI2023

Privacy-preserving federated learning: Application to behind-the-meter solar photovoltaic generation forecasting

Here, the growing usage of decentralized renewable energy sources has made accurate estimation of their aggregated generation crucial for maintaining grid flexibility and reliability. However, the majority of distributed photovoltaic (PV) systems are behind-the-meter (BTM) and invisible to utilities, leading to three challenges in obtaining an accurate forecast of their aggregated output. Firstly, traditional centralized prediction algorithms used in previous studies may not be appropriate due to privacy concerns. There is therefore a need for decentralized forecasting methods, such as federated learning (FL), to protect privacy. Secondly, there has been no comparison between localized, centralized, and decentralized forecasting methods for BTM PV production, and the trade-off between prediction accuracy and privacy has not been explored. Lastly, the computational time of data-driven prediction algorithms has not been examined. This article presents a FL power forecasting method for PVs, which uses federated learning as a decentralized collaborative modeling approach to train a single model on data from multiple BTM sites. The machine learning network used to design this FL-based BTM PV forecasting model is a multi-layered perceptron, which ensures privacy and security of the data. Comparing the suggested FL forecasting model to non-private centralized and entirely private localized models revealed that it has a high level of accuracy, with an RMSE that is 18.17% lower than localized models and 9.9% higher than centralized models.

14 SOLAR ENERGY↗