Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “comparable input ranges”

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 145 records · Page 8
All sourcesNASA NTRSDOE OSTI
DOE OSTI2025

Exploring Novel Configuration Designs Alternatives for Lazy Wave Submarine Power Cables

Floating offshore wind turbines require a complex integrated design to balance the platform response and station-keeping requirements in a highly dynamic environment. Dynamic subsea power cables are one of the critical components that have to be considered in detail to ensure both extreme, as well as repeated operational conditions do not exceed the cable design limits. This paper presents the design feasibility limits of a 66kV dynamic cable for a lazy wave configuration. The work explores an innovative staggered cable configuration that has the potential to provide additional compliance in shallow water conditions. Both configurations are evaluated using the motions of a representative floating wind platform under development for deployment in relatively shallow water depth of 150m. The assessment was conducted using the hydrodynamic software Orcaflex, incorporating a range of design inputs and design variations. Key parameters analysed include cable curvature and effective tension and bending stress. The alternative cable configuration provides a more gradual crossing of the water column, which would allow to support increased platform excursions in shallow water depths. The design uses a combined buoyancy and bend protection element, offering designers additional options for challenging operational conditions. The paper will be of interest to designers and developers of floating wind projects, providing a comparative assessment between the conventional lazy wave configuration and an innovative alternative that enables larger cable and mooring excursion envelopes in challenging conditions.

16 TIDAL AND WAVE POWER↗
DOE OSTI2020

Shale as a Source of Organic Carbon in Floodplain Sediments of a Mountainous Watershed

Abstract Shales contain high levels of organic carbon (OC) and represent a large fraction of the Earth's reduced carbon stocks. While recent evidence suggests that shale‐derived OC may be actively cycled in riverine systems, this process is poorly understood and not currently considered in global C models. Through the use of sediment density fractionations, extractions, radiocarbon measurements, and chemical characterization, we provide information on the abundance, chemistry, and mobility of shale‐derived OC in floodplain sediments of a shale‐rich mountainous watershed. The heavy fraction of the sediment, representing mineral‐associated OC, is the largest (84 ± 6% of TOC) and oldest (Δ 14 C values −224 to −853‰) OC pool. Evidence of shale‐derived OC is observed in all sediment C pools (i.e., occluded light fraction, water‐soluble, and pyrophosphate‐extractable) except the free light fraction, which is entirely modern. Relatively consistent chemistry was observed across samples for extracted and density‐separated OC, despite wide ranges of Δ 14 C values. Carbon spectroscopy revealed that floodplain sediments had a higher degree of functionalized aromatic groups and lower carbonate content compared to shale collected nearby, consistent with chemical alteration and mixing with other C sources in the floodplain. We estimate that approximately 23–34% of sediment OC is derived from shale, with implications for other shale‐derived elements (e.g., N). This study demonstrates the important contribution of shale‐OC, particularly in environments with low litter inputs. The large impact of radiocarbon‐dead shale‐OC, which has a thermally altered chemical structure distinct from plant litter, on Δ 14 C values and reactivity of sediment‐OC must be considered.

Fox, Patricia M.↗
DOE OSTI2024

Implementation and validation of optimal start control strategy for air conditioners and heat pumps

Commercial buildings are responsible for approximately 20 % of the total energy consumption and greenhouse gas emissions in the United States. Over 85 % of these buildings lack building automation systems, and many are small (<50,000 square feet), underserved, and use rooftop units (RTUs) for heating, ventilation, and air-conditioning needs. Because these buildings lack proper energy management systems, several operational deficiencies lead to excess energy consumption. Studies have shown that managing the RTUs’ heating and cooling set points, schedules, setbacks, and optimal start can result in a 20 % to 25 % reduction in electricity consumption in small commercial buildings. These buildings typically use fixed schedules to start the RTUs 60 to 120 min before occupancy begins, which results in excess energy consumption. This paper presents research that demonstrates and evaluates the performance of four optimal start methods, which utilize data-based modeling as a key element in facilitating adaptive control in response to time-varying inputs while requiring minimal sensor inputs. The evaluation found energy savings in two commercial buildings equipped with RTUs by periodically alternating four different optimal start models during the cooling and heating season. The resulting energy savings are positive for all models and range from 2 to 5 kWh/day/unit. The units on the east side of the building showed higher savings, while interior units showed greater variability in savings due to the differences in capacities and room sizes. Savings were considerably greater during the heating season compared to the cooling season. The performance of all four models on Mondays was poor; models suggested a shorter optimal start time, which resulted in relatively larger errors. Finally, the future work will look at using a different model for the days after weekends and holidays.

42 ENGINEERING↗
DOE OSTI2025

Intercomparison of Deep Learning Model Architectures for Atmospheric River Prediction

With a rapid surge in the application of machine learning (ML) for a diverse range of tasks in climate science, the present study addresses a challenge for climate scientists when selecting the optimal ML or deep learning (DL) architecture for a given application. In particular, a DL intercomparison study was performed with a focus on forecasting the position of atmospheric rivers (ARs) on short-range time scales (up to 5-day lead times). AR predictions from multiple DL architectures, including various types of convolutional autoencoders and a vision transformer (ViT), were compared against ECMWF ERA5 reanalysis and hindcasts from a global climate model. DL models with similar trainable parameters were trained on ERA5 reanalysis data and AR positions derived from a thresholding algorithm to ensure a fair comparison among the DL models. Each model’s performance and accuracy in forecasting AR location and key input fields within a 5-day window were assessed using metrics of root-mean-square error, anomaly correlation, and mean intersection over union. The ViT architecture outperformed other autoencoder models in most of the metrics. Incorporating additional meteorological fields only yielded slight improvements in forecasting certain fields at longer lead times. The results also suggest that a smaller number of input time steps or smaller number of autoregressive steps can achieve better prediction skills, while also improving the overall computational efficiency. This research offers valuable insights into the strengths and weaknesses of different DL techniques for AR forecasting, hopefully guiding the development of improved models for forecasting this phenomenon.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2022

A Universal Converter for DC, Single-phase AC, and Multi-phase AC Systems (Final Report)

The goal of this project was to develop a new class of universal power converters that utilize the unique properties of WBG switches, including high reverse voltage blocking and fast switching, to achieve superior performance in universal converters. Some of the features of the proposed converter are as follows: 1) Significantly longer lifetime, higher power density, lower costs associated with bill of materials, and lower shipping, installation, and repair costs compared to the existing WBG-based solutions, 2) Very flexible operation for transferring power between any type of source (dc, single-phase ac, or multi-phase ac) and any type of load (dc, single-phase ac, or multi-phase ac) without using large passive components or additional stages/components, even when the instantaneous values of input and output power are not the same, and 3) High performance operation at all power levels. Considering its advantages and wide range of applications, the proposed converter has the potential to create a new paradigm in power electronics. Among numerous applications of this converter are electric vehicles, wind energy systems, PV systems, industrial motor drives, residential variable frequency drive systems, and nanogrid applications. This project focused on the application of this converter for residential PV, residential motor drive, and industrial motor drive applications, and evaluated the performance of this converter for these applications.

42 ENGINEERING↗
DOE OSTI2022

Cordon screen: A cordon-based congestion pricing policy evaluation method for U.S. cities

Global trends toward urbanization will exacerbate traffic congestion, delays in economic productivity, and air pollution issues for growing cities. Traffic congestion pricing is one method available to help ameliorate these concerns. New York City is on the verge of implementing a cordon-based traffic congestion pricing policy around its central business district. For budget-constrained municipalities, evaluating implementation of such policy could be costly. This article proposes a sketch-planning methodology, called Cordon Screen, for major U.S. cities to evaluate the net income, traffic mitigation, and avoided pollution emissions from cordon-based traffic congestion pricing. This method relies on national datasets and limited user-specific data inputs, along with a range of user-selectable assumptions informed by academic literature to deliver order-of-magnitude results. The numerous limitations of this method are acceptable for preliminary policy evaluation to determine if greater financial investment to obtain more accurate results is justified. The Denver metropolitan area is used to demonstrate Cordon Screen capabilities, with mid-range assumption results suggesting the policy is most effective at generating net income and increasing vehicle speeds on major interstates. For Denver, the policy is comparably less effective at reducing air pollution and increasing speeds on minor roadways. Validation against early implementation results from the London cordon are acceptable. Still, users should discount revenue generation projections. Choice of cordon area may be the most difficult obstacle when using the Cordon Screen. With refinement, Cordon Screen could serve as a low-cost, open-source planning evaluation tool for growing and congested U.S. cities.

29 ENERGY PLANNING, POLICY, AND ECONOMY↗
DOE OSTI2024

Analysis and Control of a Three-Port Converter with High-Frequency AC Link for Hybrid Unmanned/Autonomous Aerial Vehicles

In this study, a three-port resonant converter with a high-frequency isolation link for hybrid unmanned/autonomous aerial vehicle (UAV/AAV) applications is analyzed. The rectified engine-generator set output supplies the first input ports while the other is connected to a battery energy storage system. Input ports consist of half-bridge high-frequency inverters, whereas output port utilizes a semi-bridgeless active rectifier. This rectifier is formed by replacing rectifier low-side diodes with active switching devices. A phase-shift duty-cycle control is implemented for switching signals to regulate the power share and flow between ports. The proposed control system implemented achieves zero voltage switching (ZVS), reduces switching losses, and minimum current flow optimization ensures regulation in a wide range of output power. Compared to the existing bidirectional pulse width modulation (PWM) converters, the proposed system decreases the number of power conversion stages and provides centralized control without needing communications between the ports. The proposed control technique operating in closed-loop control achieves the power flow between ports under various load conditions. Finally, theoretical analysis demonstrates the proposed converter’s viability and laboratory experimental validations at 500 W full power.

33 ADVANCED PROPULSION SYSTEMS↗
DOE OSTI2025

3D-Printed mmWave Quasi-Holographic Antenna for 2-D Beamforming

This paper presents the design and manufacturing of a novel 2D-scanning antenna that integrates a 3D-printed Rotman lens with a quasi-holographic leaky-wave antenna (HLWA). The proposed design achieved beam-scanning capabilities by leveraging the beamforming of the Rotman lens and the high-gain directive properties of the quasi-HLWA. The Rotman lens (RL) enables beam steering in the elevation plane by switching between input ports. The quasi-HLWA, designed using holographic principles, achieves frequency-controlled beam scanning in the azimuth plane. The entire antenna structure was fabricated using additive manufacturing with an Ink1092 substrate and silver ink for the conductive traces. This approach provides greater control over material placement and design freedom compared to traditional methods. A 25° transmission linear substrate taper was used to ensure good impedance matching between the Rotman lens and the quasi-HLWA, allowing greater gain while maintaining a good scanning range. The experimental results validate the 2D scanning capability of the proposed antenna. The antenna system provides coverage from −54° to 54° in the elevation θ plane and −28° to 28° in the azimuth plane ϕ , with a maximum measured gain of 21.3 dBi at 28 GHz with an average radiation efficiency η=60 %. The fabricated prototype was tested, and the performance was in good agreement with the simulated performance.

Rotman lens↗
DOE OSTI2025

Mathematical Morphological Filtering with a Self-Adaptive Reconstruction Technique and Application to Local Seismic Data

Recorded seismic data are generally contaminated by noise from different sources, which masks the signals of interest. In the seismology community, frequency filtering (FF) is the standard method for noise suppression. However, when the signal of interest and noise share the same frequency band, the latter cannot be filtered out without infringing on the former. We implemented a noise suppression approach based on the mathematical morphology theorem. The method involves compound operations of dilation and erosion using structuring elements of varying lengths and decomposes an input noisy waveform into several time functions with differing characteristics. Further, the filtered waveform is constructed from the time functions using a self-adaptive reconstruction technique. Application to a data set of >4700 local waveforms suggests that the implemented mathematical morphological filtering (MMF) approach is efficient for data with low signal-to-noise ratio (SNR) and significantly outperforms FF in that SNR range. For most of the dataset, FF, machine learning (ML) denoising, and continuous wavelet transform (CWT) thresholding result in higher SNR values compared with the MMF method. However, for ~42% of the waveforms, MMF outperforms FF, and the SNR gain achieved with MMF is as large as ~23 dB. Compared to ML denoising and CWT thresholding, this proportion drops to only ~10%–14%. Our results suggests that in an operational setting, MMF cannot replace the other noise suppression methods; however, signal detection can be improved if MMF is used to supplement them in some scenarios. MMF could help detect signals in problematic low-SNR data, which are currently being missed particularly when using FF alone.

58 GEOSCIENCES↗
DOE OSTI2021

The future evolution of energy-water-agriculture interconnectivity across the US

Abstract Energy, water, and agricultural resources across the globe are highly interconnected. This interconnectivity poses science challenges, such as understanding and modeling interconnections, as well as practical challenges, such as efficiently managing interdependent resource systems. Using the US as an example, this study seeks to define and explore how interconnectivity evolves over space and time under a range of influences. Concepts from graph theory and input–output analysis are used to visualize and quantify key intersectoral linkages using two new indices: the ‘Interconnectivity Magnitude Index’ and the ‘Interconnectivity Spread Index’. Using the Global Change Analysis Model (GCAM-USA), we explore the future evolution of these indices under four scenarios that explore a range of forces, including socioeconomic and technological change. Analysis is conducted at both national and state level spatial scales from 2015 to 2100. Results from a Reference scenario show that resource interconnectivity in the US is primarily driven by water use amongst different sectors, while changes in interconnectivity are driven by a decoupling of the water and electricity systems, as power plants become more water-efficient over time. High population and GDP growth results in relatively more decoupling of sectors, as a larger share of water and energy is used outside of interconnected sector feedback loops. Lower socioeconomic growth results in the opposite trend. Transitioning to a low-carbon economy increases interconnectivity because of the expansion of purpose-grown biomass, which strengthens the connections between water and energy. The results highlight that while some regions may experience similar sectoral stress projections, the composition of the intersectoral connectivity leading to that sectoral stress may call for distinctly different multi-sector co-management strategies. The methodology we introduce here can be applied in diverse geographical and sectoral contexts to enable better understanding of where, when, and how coupling or decoupling between sectors could evolve and be better managed.

Khan, Zarrar (ORCID:0000000281478553)↗
DOE OSTI2023

Assessment of tank designs for hydrogen storage on heavy duty vehicles using metal hydrides

The objective of this project was to evaluate material-based hydrogen storage solutions as a replacement for high-pressure hydrogen gas or liquid hydrogen on Class 7 or 8 tractor fuel cell electric vehicles. The project focused on low-density main-group hydrides, a well-known class of materials for hydrogen storage. Prior research has considered metal amides as storage materials for light-duty vehicles but not for heavy-duty applications. The project established the basis for further development of storage systems of this type for heavy duty vehicles (HDV). Systems analysis of an HDV storage system comprised of a tank and associated balance of plant (piping, coolant tubes, burner) was performed to determine the usable hydrogen capacity. A composite storage material comprised of a metal hydride mixed with a high thermal-conductivity carbon is predicted to have a usable hydrogen volumetric capacity comparable to or exceeding that of 700 bar pressurized hydrogen gas. The gravimetric capacity of this material is also predicted to be competitive with pressurized gas, particularly if costly carbon fiber composite Type III or Type IV tanks are excluded. The storage system design parameters and material properties served as inputs to a second model that simulates fuel cell operation in conjunction with the storage system during an HDV drive cycle. The results show that sufficient hydrogen pressure can be produced to operate a Class 8 HDV, yielding a range of ~480 miles. These results are particularly relevant for high-impact regions, such as the South Coast Air Quality Management District, for which an economical vehicular hydrogen storage system with minimal impact on cargo capacity could accelerate adoption of heavy-duty fuel cell electric vehicles. An additional benefit is that knowledge generated by this project can assist in development of material-based storage for stationary applications such as microgrids and backup power for data centers.

08 HYDROGEN↗
DOE OSTI2023

Experimental and modelling study of the effect of pre-shock duration on the double shock initiation response of PBX 9501

Here, a series of double shock initiation gas-gun experiments have been performed on PBX 9501 (95% weight HMX, 5% binder) in which the input conditions were nominally the same, but the time delay between the two shocks was explicitly varied. The impact conditions generated pre- and main shock pressures of 2.9 and 6.2 GPa respectively, with time delays between them ranging from 0.56 to 1.10 µs. This resulted in different levels of reaction behind the pre-shock prior to the main shock traversing the explosive samples, with the growth of reaction monitored using embedded particle velocity and shock tracker gauges. Simulated results from three reactive burn models (CREST, AWSD and SURF) are compared with the measured data to assess their ability to predict the response of PBX 9501 under such double shock conditions. The described experiments add to the body of evidence to help improve our understanding of the double shock initiation response of explosives, and the data obtained provide an excellent test for reactive burn models.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗
DOE OSTI2025

Calculation of ion–ion mutual neutralization rate constants using Landau–Zener theory coupled with trajectory simulations for Ar + –Cl − , Br − , I −

In this computational study, we self-consistently calculate the rate constants of mutual neutralization reactions by incorporating the electron transfer probability, using Landau–Zener state transition theory with inputs derived from ab initio quantum chemistry calculations, into classical trajectory simulations. Electronic structure calculations are done using correlation consistent basis sets with multi-reference configuration interaction to map all the molecular electronic states below the ion-dissociation limit as a function of the distance between the reacting species. Our electronic structure calculations have been significantly improved from our previous work through improved selection of molecular electronic configurations maintaining a fine grid of 1a 0 over a wide range of bond lengths and accurate treatment of spin–orbit couplings. Non-adiabatic coupling matrix elements are calculated with the three-point central difference method near each avoided crossing to estimate the exact crossing point R x and coupling parameter H if , which are inputs to the multi-channel Landau–Zener theory to calculate the electron transition probability. Our approach is applied to estimate the mutual neutralization rate constants for the following ion pairs: Ar + –Cl − , Ar + –Br − , Ar + –I − at ∼133 Pa. Furthermore, our predictions are compared against the experimental data reported. It is seen that the improvement in the electronic structure calculation results in excellent agreement between the simulation results and the available experimental data to within a factor of ∼2 or ∼±50%.

Complete-active space self-consistent field↗
DOE OSTI2021

A comparative study of machine learning models for predicting the state of reactive mixing

Mixing phenomena are important mechanisms controlling flow, species transport, and reaction processes in fluids and porous media. Accurate predictions of reactive mixing are critical for many Earth and environmental science problems such as contaminant fate and remediation, macroalgae growth, and plankton biomass evolution. Here, to investigate the evolution of mixing dynamics under different scenarios (e.g., anisotropy, fluctuating velocity fields), a finite-element-based numerical model was built to solve the fast, irreversible bimolecular reaction-diffusion equations to simulate a range of reactive-mixing scenarios. A total of 2,315 simulations were performed using different sets of model input parameters comprising various spatial scales of vortex structures in the velocity field, time-scales associated with velocity oscillations, the perturbation parameter for the vortex-based velocity, anisotropic dispersion contrast (i.e., ratio of longitudinal-to-transverse dispersion), and molecular diffusion. The outputs comprised concentration profiles of reactants and products. The inputs to and outputs from these simulations were concatenated into feature and label matrices, respectively, to train 20 different machine learning (ML) models intended to emulate system behavior. These 20 ML emulators, based on linear methods, Bayesian methods, ensemble learning methods, and multilayer perceptrons (MLPs), were trained to classify the state of mixing and predict three quantities of interest (QoIs) characterizing species production, decay (i.e., average concentration, square of average concentration), and degree of mixing (i.e., variances of species concentration). Unsurprisingly, linear classifiers and regressors failed to reproduce the QoIs; however, ensemble methods (classifiers and regressors) and the MLP model accurately classified the state of reactive mixing and the QoIs. Among ensemble methods, random forest and decision-tree-based AdaBoost faithfully predicted the QoIs. At run time, trained ML emulators produced results times faster than the finite-element simulations. Due to their low computational expense and high accuracy, ensemble and MLP models are excellent emulators for these numerical simulations and great utilities in uncertainty quantification exercises, which can require 1,000s of forward model runs.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2026

A Dual-Active-Bridge Converter Employing a Variable Inductor Without an Auxiliary Circuit

This paper proposes a dual active bridge (DAB) converter employing a variable inductor (VI) without an auxiliary circuit. Unlike conventional VI-based designs that require an external DC bias circuit, the proposed DAB utilizes the input DC current itself as the bias source, enabling automatic inductance variation with load conditions. The inductance naturally increases at low power and decreases at high power, effectively extending the zero voltage switching (ZVS) range and reducing circulating current, respectively. The VI was experimentally implemented and characterized to obtain its inductance-current profile, which was then integrated into a PLECS model of the DAB converter for circuit and thermal simulations. Simulation results confirm that the proposed auxiliary-free VI-DAB converter achieves a wider ZVS range and lower circulating current compared with a conventional fixed-inductor DAB converter. By realizing a variable inductor without any auxiliary bias circuitry, the proposed approach maintains soft switching and reduces reactive current losses across a wide load range, leading to improved efficiency and simplified implementation.

Jo, Cheolhui [ORNL] (ORCID:0000000322692434)↗
DOE OSTI2024

Excitation functions and isomeric cross-section ratios of (d,xn) reactions on 86 Sr

Excitation functions of the 86 Sr(d,n) 87m,87g Y, 86 Sr(d,2n) 86m,86g Y and 86 Sr(d,3n) 85m,85g Y reactions on enriched 86 Sr target were measured by the activation technique up to deuteron energies of 49 MeV. The isomeric cross-section ratios as a function of projectile energy were deduced from the measured data for 87m Y, 87g Y(cum), 86m Y, 86g Y(cum), 85m Y and 85g Y pairs for the same energy range. All measurements are reported for the first time. The experimental data were compared with the data from the TENDL library which is based on TALYS calculation with default parameters. No satisfactory agreement was observed. Nuclear model calculations were then performed using the codes TALYS and EMPIRE with some parameter adjustments, and compared with the experimental data. The quality of the agreement between experimental data and model calculations was numerically quantified. In general, the data as well as the isomeric cross-section ratios are partially reproduced by the model calculations, provided the input model parameters are properly chosen and the level structure of the product nucleus is thoughtfully considered.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
DOE OSTI2022

Microstructural Development in Inconel 718 Nickel-Based Superalloy Additively Manufactured by Laser Powder Bed Fusion

Excellent weldability and high temperature stability make Inconel 718 (IN718) one of the most popular alloys to be produced by additive manufacturing. Here in this study, we investigated the effects of laser powder bed fusion (LPBF) parameters on the microstructure and relative density of IN718. The samples were fabricated with independently varied laser power (125 - 350W), laser scan speed (200 - 2200 mm/s), and laser scan rotation (0° - 90°). Archimedes’ method, optical microscopy, and scanning electron microscopy were employed to assess the influence of LPBF parameters on the relative density and microstructure. Optimal processing windows were identified for a wide range of processing parameters and relative density greater than 99.5% was achieved using volumetric energy density between 50 and 100 J/mm3. Microstructural features including melt pool geometry, lack of fusion defect, keyhole porosity, and sub-grain cellular microstructure were examined and quantified to correlate to LPBF parameters. A simple empirical model was postulated to relate relative sample density and LPBF volumetric energy density. Melt pool dimensions were quantitatively measured and compared to estimations based on Rosenthal solution, which yielded a good agreement with the width, but underestimated the depth, particularly at high energy input, due to lack of consideration for keyhole mode. In addition, the sub-grain cellular-dendritic microstructure in the as-built samples was observed to decrease with increasing laser scan speed. Quantification of the sub-micron cellular-dendritic microstructure yielded estimated cooling rate in the order of 10 5 – 10 7 K/s.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗
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↗