SEARCH · Engineering Papers

Results for “model structural error”

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 1,009 records · Page 56

All sources NASA NTRS DOE OSTI

DOE OSTI2022

Evaluation of obstacle modelling approaches for resource assessment and small wind turbine siting: case study in the northern Netherlands

Abstract. Growth in adoption of distributed wind turbines for energy generation is significantly impacted by challenges associated with siting and accurate estimation of the wind resource. Small turbines, at hub heights of 40 m or less, are greatly impacted by terrestrial obstacles such as built structures and vegetation that can cause complex wake effects. While some progress in high-fidelity complex fluid dynamics (CFD) models has increased the potential accuracy for modelling the impacts of obstacles on turbulent wind flow, these models are too computationally expensive for practical siting and resource assessment applications. To understand the efficacy of available models in situ, this study evaluates classic and commonly used methods alongside new state-of-the-art lower-order models derived from CFD simulations and machine learning approaches. This evaluation is conducted using a subset of an extensive original dataset of measurements from more than 300 operational wind turbines in the northern Netherlands. The results show that data-driven methods (e.g. machine learning and statistical modelling) are most effective at predicting production at real sites with an average error in annual energy production of 2.5 %. When sufficient data may not be available de novo to support these data-driven approaches, models derived from high-fidelity simulations show promise and reliably outperform classic methods. On average these models have 6.3 %–11.5 % error compared with 26 % for classic methods and 27 % baseline error for reanalysis data without obstacle correction. While more performant on average, these methods are also sensitive to the quality of obstacle descriptions and reanalysis inputs.

17 WIND ENERGY↗

DOE OSTI2023

Evaluation of the excitation spectra with diffusion Monte Carlo on an auxiliary bosonic ground state

We aim to improve upon the variational Monte Carlo (VMC) approach for excitations replacing the Jastrow factor by an auxiliary bosonic (AB) ground state and multiplying it by a fermionic component factor. The instantaneous change in imaginary time of an arbitrary excitation in the original interacting fermionic system is obtained by measuring observables via the ground-state distribution of walkers of an AB system that is subject to an auxiliary effective potential. The effective potential is used to (i) drive the AB system’s ground-state configuration space toward the configuration space of the excitations of the original fermionic system and (ii) subtract from a diffusion Monte Carlo (DMC) calculation contributions that can be included in conventional approximations, such as mean-field and configuration interaction (CI) methods. In this novel approach, the AB ground state is treated statistically in DMC, whereas the fermionic component of the original system is expanded in a basis. The excitation energies of the fermionic eigenstates are obtained by sampling a fermion–boson coupling term on the AB ground state. We show that this approach can take advantage of and correct for approximate eigenstates obtained via mean-field calculations or truncated interactions. We demonstrate that the AB ground-state factor incorporates the correlations missed by standard Jastrow factors, further reducing basis truncation errors. Relevant parts of the theory have been tested in soluble model systems and exhibit excellent agreement with exact analytical data and CI and VMC approaches. In particular, for limited basis set expansions and sufficient statistics, AB approaches outperform CI and VMC in terms of basis size for the same systems. The implementation of this method in current codes, despite being demanding, will be facilitated by reusing procedures already developed for calculating ground-state properties with DMC and excitations with VMC.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

DOE OSTI2022

Abundance matching analysis of the emission-line galaxy sample in the extended Baryon Oscillation Spectroscopic Survey

We present the measurements of the small-scale clustering for the emission-line galaxy (ELG) sample from the extended Baryon Oscillation Spectroscopic Survey (eBOSS) in the Sloan Digital Sky Survey IV (SDSS-IV). We use conditional abundance matching method to interpret the clustering measurements from 0.34 to 70 h -1 Mpc. In order to account for the correlation between properties of ELGs and their environment, we add a secondary connection between star formation rate of ELGs and halo accretion rate. Three parameters are introduced to model the ELG [O II ] luminosity and to mimic the target selection of eBOSS ELGs. The parameters in our models are optimized using Markov Chain Monte Carlo (MCMC) method. We find that by conditionally matching star formation rate of galaxies and the halo accretion rate, we are able to reproduce the eBOSS ELG small-scale clustering within 1σ error level. Our best-fitting model shows that the eBOSS ELG sample only consists of ~ 12 per cent of all star-forming galaxies, and the satellite fraction of eBOSS ELG sample is 19.3 per cent. We show that the effect of assembly bias is ~ 20 per cent on the two-point correlation function and ~ 5 per cent on the void probability function at scale of r ~ 20 h -1 Mpc.

79 ASTRONOMY AND ASTROPHYSICS↗

DOE OSTI2024

Shape restricted additive hazards models: Monotone, unimodal, and U‐shape hazard functions

We consider estimation of the semiparametric additive hazards model with an unspecified baseline hazard function where the effect of a continuous covariate has a specific shape but otherwise unspecified. Such estimation is particularly useful for a unimodal hazard function, where the hazard is monotone increasing and monotone decreasing with an unknown mode. A popular approach of the proportional hazards model is limited in such setting due to the complicated structure of the partial likelihood. Our model defines a quadratic loss function, and its simple structure allows a global Hessian matrix that does not involve parameters. Thus, once the global Hessian matrix is computed, a standard quadratic programming method can be applicable by profiling all possible locations of the mode. However, the quadratic programming method may be inefficient to handle a large global Hessian matrix in the profiling algorithm due to a large dimensionality, where the dimension of the global Hessian matrix and number of hypothetical modes are the same order as the sample size. We propose the quadratic pool adjacent violators algorithm to reduce computational costs. The proposed algorithm is extended to the model with a time‐dependent covariate with monotone or U‐shape hazard function. In simulation studies, our proposed method improves computational speed compared to the quadratic programming method, with bias and mean square error reductions. We analyze data from a recent cardiovascular study.

Mathematical & Computational Biology↗

DOE OSTI2022

Electronic structure factors and the importance of adsorbate effects in chemisorption on surface alloys

The chemisorption energy is an integral aspect of surface chemistry, central to numerous fields such as catalysis, corrosion, and nanotechnology. Electronic-structure-based methods such as the Newns-Anderson model are therefore of great importance in guiding the engineering of material surfaces with optimal properties. However, existing methods are inadequate for interpreting complex, multi-metallic systems. Herein, we introduce a physics-based chemisorption model for alloyed transition metal surfaces employing primarily metal d-band properties that accounts for perturbations in both the substrate and adsorbate electronic states upon interaction. Importantly, we show that adsorbate-induced changes in the adsorption site interact with its chemical environment leading to a second-order response in chemisorption energy with the d-filling of the neighboring atoms. We demonstrate the robustness of the model on a wide range of transition metal alloys with O, N, CH, and Li adsorbates yielding a mean absolute error of 0.13 eV versus density functional theory reference chemisorption energies.

36 MATERIALS SCIENCE↗

DOE OSTI2021

Charge Radius of Neutron-Deficient Ni 54 and Symmetry Energy Constraints Using the Difference in Mirror Pair Charge Radii

he nuclear root-mean-square charge radius of Ni 54 was determined with collinear laser spectroscopy to be R ( Ni 54 ) = 3.737 ( 3 ) fm . In conjunction with the known radius of the mirror nucleus Fe 54 , the difference of the charge radii was extracted as Δ R ch = 0.049 ( 4 ) fm . Based on the correlation between Δ R ch and the slope of the symmetry energy at nuclear saturation density ( L ), we deduced 21 ≤ L ≤ 88 MeV . The present result is consistent with the L from the binary neutron star merger GW170817, favoring a soft neutron matter EOS, and barely consistent with the PREX-2 result within 1 σ error bands. Our result indicates the neutron-skin thickness of Ca 48 as 0.15–0.21 fm.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

DOE OSTI2022

Predicting oxidation damage in ultra high-temperature borides: A machine learning approach

Ultra-high temperature (UHT) borides are ceramics materials with melting points above 3000 °C for structural applications in extreme environments. However, at temperatures exceeding 1600 °C and under oxidizing conditions, the material suffers from detrimental degradation. Optimized design and performance of diboride materials under such extreme conditions requires filling the missing composition-microstructure-oxidation gap. This study proposes a computational data-driven framework to connect the processing and microstructure of Ultra-high temperature borides with the oxidation damage. Random Forest Regressor (RFR) model is adopted to forecast the oxide scale thickness developed after oxidation testing based on processing variables and microstructural features. The model trained on a dataset consisting of 107 samples of experimental data extracted from the literature aims to predict oxidation damage. With proper data manipulation and fine model tuning, the predictor could forecast the oxide scale thickness of UHT diborides with a Mean Absolute Error of 37.45 μm and an R-square of 0.83. This model could be used as a high-throughput scheme to design and test new UHT diborides materials computationally. Furthermore, a model with larger composition capabilities could also be developed in the future as more experimental data become available.

36 MATERIALS SCIENCE↗

DOE OSTI2015

Delayed Gamma Measurements in Different Nuclear Research Reactors Bringing Out the Importance of the Delayed Contribution in Gamma Flux Calculations

Neutron and gamma flux levels are key parameters in nuclear research reactors. In Material Testing Reactors, such as the future Jules Horowitz Reactor, under construction at the French Alternative Energies and Atomic Energy Commission (CEA Cadarache, France), the expected gamma flux levels are very high (nuclear heating is of the order of 20 W/g at 100 MWth). As gamma rays deposit their energy in the reactor structures and structural materials it is important to take them into account when designing irradiation devices. There are only a few sensors which allow measurements of the nuclear heating ; a recent development at the CEA Cadarache allows measurements of the gamma flux using a miniature ionization chamber (MIC). The measured MIC response is often compared with calculation using modern Monte Carlo (MC) neutron and photon transport codes, such as TRIPOLI-4 and MCNP6. In these calculations only the production of prompt gamma rays in the reactor is usually modelled thus neglecting the delayed gamma rays. Hence calculations and measurements are usually in better accordance for the neutron flux than for the gamma flux. In this paper we study the contribution of delayed gamma rays to the total MIC signal in order to estimate the systematic error in gamma flux MC calculations. In order to experimentally determine the delayed gamma flux contributions to the MIC response, we performed gamma flux measurements with CEA developed MIC at three different research reactors: the OSIRIS reactor (MTR - 70 MWth at CEA Saclay, France), the TRIGA MARK II reactor (TRIGA - 250 kWth at the Jozef Stefan Institute, Slovenia) and the MARIA reactor (MTR - 30 MWth at the National Center for Nuclear Research, Poland). In order to experimentally assess the delayed gamma flux contribution to the total gamma flux, several reactor shut down (scram) experiments were performed specifically for the purpose of the measurements. Results show that on average about 30 % of the MIC signal is due to the delayed gamma rays. In this paper we describe experiments in each of the three reactors and how we estimate delayed gamma rays with MIC measurements. The results and perspectives are discussed. (authors)

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗

DOE OSTI2020

Collagen Turnover in Relation to Risk Factors and Hemodynamics in Human Intracranial Aneurysms

Determinants for molecular and structural instability, that is, impending growth or rupture, of intracranial aneurysms (IAs) remain uncertain. To elucidate this, we endeavored to estimate the actual turnover rates of the main molecular constituent in human IA (collagen) on the basis of radiocarbon ( 14 C) birth dating in relation to IA hemodynamics. Collagen turnover rates in excised human IA samples were calculated using mathematical modeling of 14 C birth dating data of collagen in relation to risk factors and histological markers for collagen maturity/turnover in selected IA. Hemodynamics were simulated using image-based computational fluid dynamics. Correlation, logistic regression, and receiver operating characteristic analyses were performed. Collagen turnover rates were estimated in 46 IA (43 patients); computational fluid dynamics could be performed in 20 IA (20 patients). The mean collagen turnover rate (γ) constituted 126% (±1% error) per year. For patients with arterial hypertension, γ was greater than 2600% annually, whereas γ was distinctly lower with 32% (±1% error) per year for patients without risk factors, such as smoking and hypertension. There was a distinct association between histological presence of rather immature collagen in human IA and the presence of modifiable risk factors. Spatial-temporal averaged wall shear stress predicted rapid collagen turnover (odds ratio, 1.6 [95% CI, 1.0–2.7]). Receiver operating characteristic analysis demonstrated a good test accuracy (area under the curve, 0.798 [95% CI, 0.598–0.998]) for average wall shear stress with a threshold ≥4.9 Pa for rapid collagen turnover. Our data indicate that turnover rates and stability of collagen in human IA are strongly associated with the presence of modifiable risk factors and aneurysmal hemodynamics. Finally, these findings underline the importance of strict risk factor modification in patients with unruptured IA. Future should include more detailed risk factor data to establish a more causal understanding of hemodynamics and the rupture risk of individual IA.

60 APPLIED LIFE SCIENCES↗

DOE OSTI2014

Global SAXS Data Analysis for Multilamellar Vesicles: Evolution of the Scattering Density Profile (SDP) Model

The highly successful scattering density profile (SDP) model, used to jointly analyze small-angle X-ray and neutron scattering data from unilamellar vesicles, has been adapted for use with data from fully hydrated, liquid crystalline multilamellar vesicles (MLVs). Using a genetic algorithm, this new method is capable of providing high-resolution structural information, as well as determining bilayer elastic bending fluctuations from standalone X-ray data. Structural parameters such as bilayer thickness and area per lipid were determined for a series of saturated and unsaturated lipids, as well as binary mixtures with cholesterol. The results are in good agreement with previously reported SDP data, which used both neutron and X-ray data. The inclusion of deuterated and non-deuterated MLV neutron data in the analysis improved the lipid backbone information but did not improve, within experimental error, the structural data regarding bilayer thickness and area per lipid.

Heftberger, Peter↗

DOE OSTI2019

A Decision Tree Based Machine Learning Algorithm for Rate Estimation

A subgraph isomorphic decision tree algorithm was developed for estimating chemical reaction rates. The algorithm essentially associates differences in rate estimation with differences in sub-structures on the reactant or reactants, and the tree splits reactions into groups with similar rates by identifying the substructural differences that affect rate estimation the most. The algorithm was trained on rates from the Reaction Mechanism Generator (RMG) database and compared with a much more traditional rate rule estimator based on the same data. Compared to the previous RMG rate rules, the prediction accuracy of the new algorithm, measured by median absolute error (MAE) factor, root mean squared error (RMSE) factor, and 2-sigma error factor, is consistently improved by a factor of two. Uncertainty estimation was also developed for the decision trees based on fitting a lognormal distribution to the leave-one-out errors at each node. Improvements in accuracy introduced by the algorithm should significantly reduce the amount of effort required to refine models and the need for additional quantum calculations and thus reduce the amount of time and cost required to create and refine biofuel models.

Johnson, Matthew S.↗

DOE OSTI2024

Improving gas adsorption modeling for MOFs by local calibration of Hubbard U parameters

While computational screening with density functional theory (DFT) is frequently employed for the screening of metal–organic frameworks (MOFs) for gas separation and storage, commonly applied generalized gradient approximations (GGAs) exhibit self-interaction errors, which hinder the predictions of adsorption energies. We investigate the Hubbard U parameter to augment DFT calculations for full periodic MOFs, targeting a more precise modeling of gas molecule–MOF interactions, specifically for N2, CO2, and O2. We introduce a calibration scheme for the U parameter, which is tailored for each MOF, by leveraging higher-level calculations on the secondary building unit (SBU) of the MOF. When applied to the full periodic MOF, the U parameter calibrated against hybrid HSE06 calculations of SBUs successfully reproduces hybrid-quality calculations of the adsorption energy of the periodic MOF. The mean absolute deviation of adsorption energies reduces from 0.13 eV for a standard GGA treatment to 0.06 eV with the calibrated U, demonstrating the utility of the calibration procedure when applied to the full MOF structure. Furthermore, attempting to use coupled cluster singles and doubles with perturbative triples calculations of isolated SBUs for this calibration procedure shows varying degrees of success in predicting the experimental heat of adsorption. It improves accuracy for N2 adsorption for cases of overbinding, whereas its impact on CO2 is minimal, and ambiguities in spin state assignment hinder consistent improvements of O2 adsorption. Our findings emphasize the limitations of cluster models and advocate the use of full periodic MOF systems with a calibrated U parameter, providing a more comprehensive understanding of gas adsorption in MOFs.

Chemistry↗

DOE OSTI2024

Material Model Parameters Optimization in Liquid Mercury Target Dynamics Simulation With Machine Learning Surrogates

A pulsed spallation target is subjected to very short (∼0.7μs) but intense loads (23.3 kJ) from repeated proton pulses, which knock away neutrons from the mercury atoms’ nuclei for a wide range application in physics, engineering, medicine, petroleum exploration, biology, chemistry, etc. The effect of this pulsed loading on the stainless-steel target module which contains the flowing mercury target material is difficult to predict not only due to its short but intense explosive-like physical reaction, but also the nonlinear material behavior of the liquid mercury in the structure. Injecting small helium bubbles in the mercury has been an efficient method of mitigating the pressure wave at high power level stage. However, prediction of the resultant loading on the target is more difficult when helium gas is intentionally injected into the mercury. A 2-phase material model that incorporates the Rayleigh-Plesset (R-P) model is expected to address this complex multi-physics dynamics problem by including the bubble dynamics in the liquid mercury. A parameter sensitivity study was firstly employed to understand their impact on the simulation strains. The investigated parameters included E, μ, γ, σ, n, VFgas, and gas cumulative volume curve control parameters a and b. Verification and validation results from sparse polynomial expansions (SPE) method and directional Gaussian smoothing (DGS) optimization show that the surrogate model had training error of ∼7% and validation error of ∼15%, indicating that machine learning methods and surrogate models can help optimize the uncertain parameters in the complex 2-phase material model. This approach is expected to fill the knowledge gap between unknown liquid-gas mixture material model and measured vessel strain responses.

Lin, Lianshan↗

DOE OSTI2024

SSTDR and FDR Detection of Un-Energized and Energized Cable Anomalies Including Thermal Degradation Using Machine Learning

Historically, cables are initially qualified for nuclear power plant use for 40 years. As plants extend their operating license to 60 and 80 years, continued use of these cables must shift to a performance-based approach since it is cost prohibitive to completely replace cables that are likely still capable of performing their design function. A variety of cable tests are available and are commonly applied during outages when the cables can be taken out of service. Frequency domain reflectometry (FDR) is one of these test methods that is being more broadly accepted and used because it not only detects anomalies along the cable with a low-voltage signal that does not stress the cable insulation, but the technique also locates the anomalies. This supports follow-up local inspection and local repair or partial replacement of a damaged cable segment. Currently, FDR testing is only applied to cables that are taken out of service since the test instrument would be damaged by operational voltages. A related technology that has found some acceptance in the aircraft and rail industry is spread spectrum time domain reflectometry (SSTDR). This technology has been implemented with a custom commercial instrument by LiveWire Innovation that is designed to operate on live cables up to 1000 volts and with a bandwidth of 48 MHz. Initial evaluation by the Pacific Northwest National Laboratory (PNNL) of the Live Wire system indicated that a broader bandwidth (BW) SSTDR may be better for many kinds of flaws. This led PNNL to develop an SSTDR laboratory instrument suitable for tests up to 500 MHz bandwidth. Testing on energized cables is also desirable for online monitoring systems so an inductive clamshell coupler was developed that allows energized cables to be tested up to at least 5 kV and likely higher voltage levels. Dielectric spectroscopy and tan delta testing plus various laboratory destructive tests were included in this data acquisition campaign directed to feed a machine learning (ML) study. With these kinds of developments, online energized cable tests may be possible with industrial adoption of such hardware advances but it will be completely impractical to have highly skilled data analysts continually examine these complex signals for indications of damage or compromised conditions. If online testing is to be implemented in new test hardware, it must be accompanied by software that can interpret the signals and alert plant operators of changing or degraded conditions. The thermally aged, shielded cable investigated here was separately treated for ML analysis. Visual analysis of electrical data showed generally increasing peaks where the cable entered and exited the oven. These peaks were not exactly aligned with expected locations, but these differences were attributed to velocity of propagation calibration errors. Only supervised ML was applied to the thermally aged data as this data was only available shortly before the committed publication date of this report. The supervised ML was structured to divide the 0 to 70-day responses as ‘normal’ from 0 to 35 days or ‘anomalous’ from 36 to 70 days, based on cable tensile elongation at break (EAB) insulation characterization. Using 80% of the data for training and 20% for testing, the supervised ML predicted normal versus anomalous was 70% accurate. Important conclusions include: • Accuracy to predict the presence of cable damage is improved from the 2023 effort by more training data. Weighted accuracies for comparisons among the instruments ranged from 67 to 89 % for unsupervised ML and 71 to 99% for supervised ML. • Based on the synthetic data tests, the unsupervised models are more generalizable to unseen anomalies. The Multi-Layer Perceptron classifier (MLP) model reported as high as 99.7% accuracy on the test data, but this dropped to 58.3% when tested on the synthetic data. In contrast, the unsupervised Pointwise model only achieved 89.7% accuracy on the experimental data but reported 78.3% accuracy on the synthetic data. • The best anomaly indicators are higher frequency (400 MHz BW) FDR data. Other tests may be interesting but for this study, this was the best predicter.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

DOE OSTI2018

Using Deep Machine Learning to Conduct Object-Based Identification and Motion Detection on Safeguards Video Surveillance

Video surveillance is one of the core monitoring technologies used by the International Atomic Energy Agency (IAEA) Department of Safeguards at safeguarded nuclear facilities worldwide. Current IAEA image-review software has functions for scene-change detection, black image detection and missing scene analysis, but their capabilities are not optimum. The current workflow for the detection of safeguards relevant events heavily depends on inspectors’ laborious visual examination of surveillance videos, which is a time-consuming process and prone to errors. To improve the accuracy of the process and reduce inspectors’ burden, the paper proposes using deep machine learning to detect objects of interest in video streams and to conduct object-based motion detection. The hypothesis of this work is that deep machine learning will reduce the burden on inspectors and reduce errors by automatically locating and identifying objects of interest in video streams. Objects of interest include casks and fuel assemblies that are typically monitored by inspectors. The algorithm being developed in this work is based on a computationally efficient deep machine learning algorithm – You Only Look Once (YOLO) – but is further devised to address specific challenges related to the operation of nuclear facilities. The developed model (which is called YOLO-NSG – YOLO for Nuclear Safeguards) is evaluated with data sets collected at the test facilities at Brookhaven National Laboratory (BNL) and Sandia National Laboratories (SNL). The initial focus of the research is for application at safeguarded nuclear reactors, such as pressurized heavy-water reactors, where video surveillance is broadly deployed, but can be extended to other use cases of nuclear safeguards. The detailed structure of YOLO-NSG model is introduced and the test results are reported in the paper.

36 MATERIALS SCIENCE↗

DOE OSTI2019

Assessing Electronic Structure Methods for Long-Range Three-Body Dispersion Interactions: Analysis and Calculations on Well-Separated Metal Atom Trimers

Three-body dispersion interactions are much weaker than their two-body counterpart. However, their importance grows quickly as the number of interacting monomers rises. To explore the numerical performance of correlation methods for long-range three-body dispersion, we performed calculations on eight very simple dispersion-dominated model metal trimers: Na 3 , Mg 3 , Zn 3 , Cd 3 , Hg 3 , Cu 3 , Ag 3 , and Au 3 . One encouraging aspect is that relatively small basis sets of augmented triple-ζ size appear to be adequate for three-body dispersion in the long-range. Coupled cluster calculations were performed at high levels to assess MP3, CCSD, CCSD(T), empirical density functional theory dispersion (D3), and the many-body dispersion (MBD) approach. We found that the accuracy of CCSD(T) was generally significantly lower than for two-body interactions, with errors sometimes reaching 20% in the investigated systems, while CCSD and particularly MP3 were generally more erratic. MBD is found to perform better than D3 at large distances, whereas the opposite is true at shorter distances. When computing reference numbers for three-body dispersion, care should be taken to appropriately represent the effect of the connected triple excitations, which are significant in most cases and incompletely approximated by CCSD(T).

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

DOE OSTI2026

Surrogate model evaluation and building energy benchmarking for commercial buildings

Building energy consumption benchmarking involves challenges associated with various energy patterns for different building types; heating, ventilating, and air-conditioning (HVAC) system types; and climates. Given significant variation in energy use patterns, accurate prediction of long-term energy use using surrogate models remains challenging. Multiple linear regression (MLR) is commonly used for building energy benchmarking because of its simple structure; however, it lacks accuracy compared to other black-box models. Although many studies have compared surrogate models and offer guidance on model selection based on metrics, they do not provide detailed analysis on improving the surrogate model accuracy. In this paper, we implement a surrogate model using polynomial ridge regression (i.e., MLR with interaction terms combined with ridge regularization) for small office and retail strip mall buildings across six HVAC system types and all climate zones, for electricity and natural gas in baseline and proposed scenarios. A simulation workflow is developed using OpenStudio TM /EnergyPlus TM to generate simulation data using measures over a wide range of efficiency inputs. Enhancements based on statistical insights are used for improving the model accuracy using filters, input transformations, and change points. Surrogate models achieved average coefficient of variation of the root mean squared error (CVRMSE) values of 2.17, 1.06, 2.05, and 3.26 for proposed electricity, proposed natural gas, baseline electricity, and baseline natural gas, respectively, with enhancements reducing CVRMSE by an average of 14.9% across all combinations. We provide model interpretation via Shapley additive explanations to determine which input variables most influence energy consumption and provide supportive arguments for enhancements.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

DOE OSTI2022

GraphAlign: Graph-Enabled Machine Learning for Seismic Event Filtering

This report summarizes results from a 2 year effort to improve the current automated seismic event processing system by leveraging machine learning models that can operated over the inherent graph data structure of a seismic sensor network. Specifically, the GraphAlign project seeks to utilize prior information on which stations are more likely to detect signals originating from particular geographic regions to inform event filtering. To date, the GraphAlign team has developed a Graphical Neural Network (GNN) model to filter out false events generated by the Global Associator (GA) algorithm. The algorithm operates directly on waveform data that has been associated to an event by building a variable sized graph of station waveforms nodes with edge relations to an event location node. This builds off of previous work where random forest models were used to do the same task using hand crafted features. The GNN model performance was analyzed using an 8 week IMS/IDC dataset, and it was demonstrated that the GNN outperforms the random forest baseline. We provide additional error analysis of which events the GNN model performs well and poorly against concluded by future directions for improvements.

58 GEOSCIENCES↗