Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “machine learning algorithms”

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 775 records · Page 43
All sourcesNASA NTRSDOE OSTI
DOE OSTI2022

Anomaly Detection and Mitigation in FACTS-based Wide-Area Voltage Control Systems using Machine Learning

With the increasing deployment of Flexible AC Transmission System (FACTS) devices in wide-area voltage control systems (WAVCS) for achieving improved voltage stability of bulk power systems, the possibility for cyber attacks on these systems is also increasing. Successful stealthy cyber attacks that are difficult to detect by traditional informational technology (IT)-based cybersecurity solutions or threshold-based bad data detectors can lead to a voltage collapse in power grid. This paper presents the testbed-based attacks implementation and real-time evaluation of machine learning (ML) algorithm for detecting and mitigating stealthy cyber attacks on FACTS-based WAVCS on a hardware-in-the-loop (HIL) testbed. Initially, we discuss the implementation of a fuzzy logic controller (FLC) that controls a Static VAR Compensator (SVC) device deployed in a two-area four-machine Kundur power system for improving transient voltage stability. Later, the ML-based Anomaly Detection and Mitigation (ADM) system is implemented on the cyber-physical HIL testbed to detect and mitigate various stealthy cyber attacks, which are injected in real-time over the wide-area network (WAN). The experimental results show accurate and effective performance of ADM system in detecting and mitigating anomalies while keeping the grid stable and within the system operating limits, as defined by the North America Electric Reliability Corporation (NERC).

24 POWER TRANSMISSION AND DISTRIBUTION↗
DOE OSTI2022

Multi-Kernel Support Vector Machine based Predictive Maintenance of Circulating Water Pumps in Nuclear Power Plants

Multi-Kernel Support Vector Machine (MK-SVM) is a machine learning classification algorithm that can assist in the development of predictive maintenance strategies for nuclear power plant systems. Predictive maintenance can alleviate maintenance costs and enhance reliability of plant systems. In this work, MK-SVM is utilized for determining the health of the circulating water system (CWS) in a nuclear power plant.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗
NASA NTRS2000

Autonomous Performance Monitoring System: Monitoring and Self-Tuning (MAST)

Maintaining the long-term performance of software onboard a spacecraft can be a major factor in the cost of operations. In particular, the task of controlling and maintaining a future mission of distributed spacecraft will undoubtedly pose a great challenge, since the complexity of multiple spacecraft flying in formation grows rapidly as the number of spacecraft in the formation increases. Eventually, new approaches will be required in developing viable control systems that can handle the complexity of the data and that are flexible, reliable and efficient. In this paper we propose a methodology that aims to maintain the accuracy of flight software, while reducing the computational complexity of software tuning tasks. The proposed Monitoring and Self-Tuning (MAST) method consists of two parts: a flight software monitoring algorithm and a tuning algorithm. The dependency on the software being monitored is mostly contained in the monitoring process, while the tuning process is a generic algorithm independent of the detailed knowledge on the software. This architecture will enable MAST to be applicable to different onboard software controlling various dynamics of the spacecraft, such as attitude self-calibration, and formation control. An advantage of MAST over conventional techniques such as filter or batch least square is that the tuning algorithm uses machine learning approach to handle uncertainty in the problem domain, resulting in reducing over all computational complexity. The underlying concept of this technique is a reinforcement learning scheme based on cumulative probability generated by the historical performance of the system. The success of MAST will depend heavily on the reinforcement scheme used in the tuning algorithm, which guarantees the tuning solutions exist.

Peterson, Chariya↗
DOE OSTI2023

Enhanced Surgical Decision-Making Tools in Breast Cancer: Predicting 2-Year Postoperative Physical, Sexual, and Psychosocial Well-Being following Mastectomy and Breast Reconstruction (INSPiRED 004)

Background: We sought to predict clinically meaningful changes in physical, sexual, and psychosocial well-being for women undergoing cancer-related mastectomy and breast reconstruction 2 years after surgery using machine learning (ML) algorithms trained on clinical and patient-reported outcomes data. Patients and Methods: We used data from women undergoing mastectomy and reconstruction at 11 study sites in North America to develop three distinct ML models. We used data of ten sites to predict clinically meaningful improvement or worsening by comparing pre-surgical scores with 2 year follow-up data measured by validated Breast-Q domains. We employed ten-fold cross-validation to train and test the algorithms, and then externally validated them using the 11th site’s data. We considered area-under-the-receiver-operating-characteristics-curve (AUC) as the primary metric to evaluate performance. Results: Overall, between 1454 and 1538 patients completed 2 year follow-up with data for physical, sexual, and psychosocial well-being. In the hold-out validation set, our ML algorithms were able to predict clinically significant changes in physical well-being (chest and upper body) (worsened: AUC range 0.69–0.70; improved: AUC range 0.81–0.82), sexual well-being (worsened: AUC range 0.76–0.77; improved: AUC range 0.74–0.76), and psychosocial well-being (worsened: AUC range 0.64–0.66; improved: AUC range 0.66–0.66). Baseline patient-reported outcome (PRO) variables showed the largest influence on model predictions. Conclusions: Machine learning can predict long-term individual PROs of patients undergoing postmastectomy breast reconstruction with acceptable accuracy. This may better help patients and clinicians make informed decisions regarding expected long-term effect of treatment, facilitate patient-centered care, and ultimately improve postoperative health-related quality of life.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2021

From Machine Learning to Nuclear Digital Twins

This presentation provides multiple case studies of Machine Learning, Uncertainty Quantification, Reduced Order Modeling, to achieve digital transformation and provide the components for nuclear digital twinning. The presentation is for an international virtual event titled: Consultancy Meeting on Applications of AI and Pattern Recognition Techniques for Uncertainty Quantification in Nuclear Power Modelling and Simulation held Oct 21-22, 2021.

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗
DOE OSTI2022

An Efficient Bayesian Approach to Learning Droplet Collision Kernels: Proof of Concept Using “Cloudy,” a New n -Moment Bulk Microphysics Scheme

The small-scale microphysical processes governing the formation of precipitation particles cannot be resolved explicitly by cloud resolving and climate models. Instead, they are represented by microphysics schemes that are based on a combination of theoretical knowledge, statistical assumptions, and fitting to data (“tuning”). Historically, tuning was done in an ad hoc fashion, leading to parameter choices that are not explainable or repeatable. Recent work has treated it as an inverse problem that can be solved by Bayesian inference. The posterior distribution of the parameters given the data—the solution of Bayesian inference—is found through computationally expensive sampling methods, which require over $\mathcal{O}$(10 5 ) evaluations of the forward model; this is prohibitive for many models. We present a proof of concept of Bayesian learning applied to a new bulk microphysics scheme named “Cloudy,” using the recently developed Calibrate-Emulate-Sample (CES) algorithm. Cloudy models collision-coalescence and collisional breakup of cloud droplets with an adjustable number of prognostic moments and with easily modifiable assumptions for the cloud droplet mass distribution and the collision kernel. The CES algorithm uses machine learning tools to accelerate Bayesian inference by reducing the number of forward evaluations needed to $\mathcal{O}$(10 2 ). It also exhibits a smoothing effect when forward evaluations are polluted by noise. In a suite of perfect-model experiments, we show that CES enables computationally efficient Bayesian inference of parameters in Cloudy from noisy observations of moments of the droplet mass distribution. In an additional imperfect-model experiment, a collision kernel parameter is successfully learned from output generated by a Lagrangian particle-based microphysics model.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2020

Prediction of gas hydrate saturation using machine learning and optimal set of well-logs

We report resistivity and acoustic logs are widely used to estimate gas hydrate saturation in various sedimentary systems using one of the two popular methods ((1) acoustic velocity and (2) electrical resistivity), but the limitations of these two methods are often overlooked, which include (i) well-specific calibration of empirical exponents in the electrical resistivity method, (ii) assumption of known pore morphology for gas hydrates in the acoustic velocity method, and (iii) presence of unknown mineralogy and bulk modulus terms in the acoustic velocity method. NMR-density porosity-derived gas hydrate saturation based on the analysis of the transverse magnetization relaxation time (T2) is considered the most precise method, but acquisition of NMR-based logs is limited at relatively recent drilled sites; additionally, its use in conventional oil and gas reservoirs is not that common due to higher cost and operational deployment limitations associated with acquiring NMR well-logs. This study proposes a new method that predicts gas hydrate saturation (S h ) for any well using porosity, bulk density, and compressional wave (P wave) velocity well-logs with neural network (or stochastic gradient descent regression) without any well-specific calibration and/or other aforementioned shortcomings of the existing methods. The method is developed by examining the underlying dependency between S h and different combinations of well-logs, chosen from 6 routine logs, with 12 different machine learning (ML) algorithms. The accuracy of the proposed method in predicting S h is ~ 84%, which is better than the accuracy of seismic and electrical resistivity methods (≤ 75%) per the results reported by three different studies. The robustness of the method in the specific case of permafrost-associated gas hydrates is demonstrated with well-log data from two wells drilled on the Alaska North Slope.

58 GEOSCIENCES↗
DOE OSTI2020

Predicting boron coordination in multicomponent borate and borosilicate glasses using analytical models and machine learning

Accurate prediction of boron coordination in multicomponent glasses is critical in glass science and technology as it strongly affects the properties of borate and borosilicate glasses. We have collected a dataset containing 657 glasses from literature with boron coordination values and developed models using analytical functions based on the well accepted Dell, Xiao and Bray model. Good prediction of boron coordination with a R 2 value higher than 0.8 was obtained. The large variation of boron coordination from experiments, originated from sample preparations and characterizations, led to difficulties in obtaining models with better prediction performance. Various machine learning (ML) algorithms were evaluated and slightly better prediction performance was observed; however, interpretation of the ML models is less straight forward. In conclusion, this study developed various models capable of providing quantitative boron coordination predictions, providing insights into its structural roles in multi-component glasses, and suggesting fruitful areas for future research.

36 MATERIALS SCIENCE↗
DOE OSTI2020

Emerging materials intelligence ecosystems propelled by machine learning

We report that the age of cognitive computing and artificial intelligence (AI) is just dawning. Inspired by its successes and promises, several AI ecosystems are blossoming, many of them within the domain of materials science and engineering. These materials intelligence ecosystems are being shaped by several independent developments. Machine learning (ML) algorithms and extant materials data are utilized to create surrogate models of materials properties and performance predictions. Materials data repositories, which fuel such surrogate model development, are mushrooming. Automated data and knowledge capture from the literature (to populate data repositories) using natural language processing approaches is being explored. The design of materials that meet target property requirements and of synthesis steps to create target materials appear to be within reach, either by closed-loop active-learning strategies or by inverting the prediction pipeline using advanced generative algorithms. AI and ML concepts are also transforming the computational and physical laboratory infrastructural landscapes used to create materials data in the first place. Surrogate models that can outstrip physics-based simulations (on which they are trained) by several orders of magnitude in speed while preserving accuracy are being actively developed. Automation, autonomy and guided high-throughput techniques are imparting enormous efficiencies and eliminating redundancies in materials synthesis and characterization. The integration of the various parts of the burgeoning ML landscape may lead to materials-savvy digital assistants and to a human-machine partnership that could enable dramatic efficiencies, accelerated discoveries and increased productivity. Here, we review these emergent materials intelligence ecosystems and discuss the imminent challenges and opportunities. The materials research landscape is being transformed by the infusion of approaches based on machine learning. This Review discusses the emerging materials intelligence ecosystems and the potential of human-machine partnerships for fast and efficient virtual materials screening, development and discovery.

36 MATERIALS SCIENCE↗
DOE OSTI2024

Machine Learning Driven Sensitivity Analysis of E3SM Land Model Parameters for Wetland Methane Emissions

Methane (CH 4 ) is globally the second most critical greenhouse gas after carbon dioxide, contributing to 16%–25% of the observed atmospheric warming. Wetlands are the primary natural source of methane emissions globally. However, wetland methane emission estimates from biogeochemistry models contain considerable uncertainty. One of the main sources of this uncertainty arises from the numerous uncertain model parameters within various physical, biological, and chemical processes that influence methane production, oxidation, and transport. Sensitivity Analysis (SA) can help identify critical parameters for methane emission and achieve reduced biases and uncertainties in future projections. This study performs SA for 19 selected parameters responsible for critical biogeochemical processes in the methane module of the Energy Exascale Earth System Model (E3SM) land model (ELM). The impact of these parameters on various CH 4 fluxes is examined at 14 FLUXNET- CH 4 sites with diverse vegetation types. Given the extensive number of model simulations needed for global variance-based SA, we employ a machine learning (ML) algorithm to emulate the complex behavior of ELM methane biogeochemistry. We found that parameters linked to CH 4 production and diffusion generally present the highest sensitivities despite apparent seasonal variation. Comparing simulated emissions from perturbed parameter sets against FLUXNET-CH 4 observations revealed that better performances can be achieved at each site compared to the default parameter values. This presents a scope for further improving simulated emissions using parameter calibration with advanced optimization techniques.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2021

Towards Automatic and Agile AI/ML Accelerator Design with End-to-End Synthesis

Domain-specific designs offer greater energy efficiency and performance gain than general-purpose processors. For this reason, modern system-on-chips have a significant portion of their silicon area with custom accelerators. However, designing hardware by hand is laborious and time-consuming, given the large design space and the performance, power, and area constraints that are not realized in the software. Moreover, domain-specific algorithms (e.g., machine learning models) are evolving quickly, challenging the accelerator design further. To address these issues, this paper presents SODA Synthesizer, an automated open-source high-level ML framework to Verilog modular compiler targeting AI/ML Application-Specific Integrated Circuits (ASICs) accelerators. SODA tightly couples the Multi- Level Intermediate Representation (MLIR) compiler infrastructure [24] and open-source HLS approaches. Thus, SODA can support various ML frameworks and algorithms and can perform optimizations that combine specialized architecture templates and conventional HLS to generate the hardware modules. In addition, SODA’s closed-loop design space exploration (DSE) engine allows developers to perform end-to-end design space explorations on different metrics and technology nodes.

Zhang, Jeff↗
NASA NTRSDate not supplied

Algorithmic Classification of Raman Spectra Biosignatures: Improving Life Detection Confidence

“Agnostic” biosignatures – indicators of life (or the absence of life), independent of a particular biochemistry – are increasingly considered a high standard for life detection. The Ladder of Life Detection (2018) called for investigating how combinations of independent and different potential biosignatures affect confidence. To address this gap, statistical classification of elemental abundances, isotopic fractionation, and reflectance spectroscopy (VNIR) has been implemented. Raman spectroscopy, highly desirable due to its wide availability, has the potential to improve this predictive power. This work implemented biosignature classification algorithms on Raman data alone, in preparation for combination with the other data types. Raman spectroscopy data was collected from published databases and papers as part of a manually curated dataset of “indicative” and “non-indicative of life” samples. These currently include 61 non-indicative samples (meteorites, magnetite); 3 indicative living samples (bacteria); 20 indicative non-living samples (chalk, bone); and 12 indicative mixed (with non-indicative material) samples (soil, microbial mats). Laboratory work is ongoing to characterize additional samples, particularly a greater breadth of mixed systems. Spectra were interpolated, filtered with the Savitzsky-Golay filter, and de-noised. For a preliminary examination, agnostic features were manually extracted including mean intensity, number of peaks, and mean peak width. Different peak prominences and filtering polynomials were used to refine features. Classification algorithms were implemented: k-nearest neighbors (KNN), logistic regression (LR), linear support vector machines (SVM), random forest (RF), Gaussian naïve bayes (GNB). Lastly, Monte Carlo simulations on 1,000 50%-train-test-splits were used to validate classification performance and feature significance. The preliminary feature set achieved its highest AUC of 0.52 with LR, with no strongly discriminatory features. Work to improve feature extraction, such as through deep learning with back propagation, is planned. In future work, the Raman data will be combined with the other data types, and potentially new data types such as enantiomeric excess. This project was partially supported through the NASA Ames Project EXcellence (APEX) incubator program.

Astrobiology↗
DOE OSTI2023

Machine Learning-Enabled Quantitative Analysis of Optically Obscure Scratches on Nickel-Plated Additively Manufactured (AM) Samples

Additively manufactured metal components often have rough and uneven surfaces, necessitating post-processing and surface polishing. Hardness is a critical characteristic that affects overall component properties, including wear. This study employed K-means unsupervised machine learning to explore the relationship between the relative surface hardness and scratch width of electroless nickel plating on additively manufactured composite components. The Taguchi design of experiment (TDOE) L9 orthogonal array facilitated experimentation with various factors and levels. Initially, a digital light microscope was used for 3D surface mapping and scratch width quantification. However, the microscope struggled with the reflections from the shiny Ni-plating and scatter from small scratches. To overcome this, a scanning electron microscope (SEM) generated grayscale images and 3D height maps of the scratched Ni-plating, thus enabling the precise characterization of scratch widths. Optical identification of the scratch regions and quantification were accomplished using Python code with a K-means machine-learning clustering algorithm. The TDOE yielded distinct Ni-plating hardness levels for the nine samples, while an increased scratch force showed a non-linear impact on scratch widths. The enhanced surface quality resulting from Ni coatings will have significant implications in various industrial applications, and it will play a pivotal role in future metal and alloy surface engineering.

36 MATERIALS SCIENCE↗
NASA NTRSDate not supplied

From the Knowledge-based Digital Platform (KbDP) Concept for Advanced Air Mobility Research to a Preliminary Prototype

Advanced Air Mobility (AAM) encompasses a range of innovative operational and technological changes to aviation (electric aircraft, increasingly automated aircraft, increasingly automated airspace operations, etc.) that are transforming aviation’s role in everyday movement of people and goods. There are multiple associated concepts and use cases for AAM, all interrelated, including small Unmanned Aircraft System (UAS) Traffic Management (UTM), Upper-Class E Traffic Management (ETM), Extensible Traffic Management (xTM), Regional Air Mobility (RAM), and Urban Air Mobility (UAM). These AAM operations must integrate with traditional Air Traffic Management (ATM) operations, as well as non-aviation modes of transportation and logistics. National Aeronautics and Space Administration (NASA) is spearheading an innovative digital engineering approach to integrate, communicate, and facilitate the research of multi-modal transportation systems. The Knowledge-based Digital Platform (KbDP) is a concept being developed that ties the workflows of Project Managers (PM), Principal Investigators (PI), and System Engineers together across organizational boundaries. It does so through the management of an information database defined by mathematical, data science, and system engineering principles. Machine Learning (ML) algorithms play a key role in this concept by extracting meaningful knowledge from the information database, which the human user leverages to greatly improve the efficiency and effectiveness of their research. Expected benefits of this concept include improved technology transfers from research to production, improved research portfolio investments, and research outcomes that are more integrated with all aspects of the multi-modal transportation problem. The preliminary KbDP prototype has been realized using UAM as a pathfinder use case and developed by a team of system engineer, software developer, data scientist, and interns.

Systems Engineering↗
DOE OSTI2021

Informing nuclear physics via machine learning methods with differential and integral experiments

Information from differential nuclear-physics experiments and theory is often too uncertain to accurately define nuclear-physics observables such as cross sections or energy spectra. Integral experimental data, representing the applications of these observables, are often more precise but depend simultaneously on too many of them to unambiguously identify issues in the observable with human expert analysis alone. Here, we explore how we can leverage physics knowledge gained from differential experimental data, nuclear theory, integral experiments, and neutron-transport calculations to better understand nuclear-physics observables in the context of the application area represented by integral experiments. We support this task with machine-learning methods to discern trends in a large amount of convoluted data. Differential and integral information was used in an analysis augmented by the random forest and the Shapley additive explanations metric. We chose as an application area one that is represented by criticality measurements and pulsed-sphere neutron-leakage spectra. We show one representative example ( 241 Pu fission observables) where the combination of differential and integral information allowed to resolve issues in data representing these observables. As a starting point, the machine learning (ML) algorithms highlighted several observables as leading potentially to bias in simulating integral experiments. Differential information, paired with sensitivity to integral quantities, allowed us then to pinpoint one specific observable ( 241 Pu fission cross section) as the main driver of bias. The comparison to integral experiments, on the other hand, allowed us to indicate a likely reliable experiment among several discrepant ones for this observables. In other cases (e.g., 239 Pu observables), we were not able to resolve the confounding introduced by integral experiments but instead highlighted the need for targeted new experiments and theory developments to better constrain the nuclear-physics space for the application area represented by integral experiments. We were able to combine information from differential experimental data, nuclear-physics theory, integral experiments, and neutron-transport simulations of the latter experiments with the help of the random forest algorithm and expert judgment. This combination of knowledge allows to improve our description of nuclear-physics observables as applied to a particular application area.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
DOE OSTI2024

Physics-informed machine learning to predict solvatochromic parameters of designer solvents with case studies in CO 2 and lignin dissolution

The polarity of solvents plays a critical role in various research applications, particularly in their solubilities. Polarity is conveniently characterized by the Kamlet-Taft parameters that is, the hydrogen bonding acidity (α), the basicity (β), and the polarizability (π*). Obtaining Kamlet-Taft parameters is very important for designer solvents, namely ionic liquids (ILs) and deep eutectic solvents (DESs). However, given the unlimited theoretical number of combinations of ionic pairs in ILs and hydrogen-bond donor/acceptor pairs in DESs, experimental determination of their Kamlet-Taft parameters is impractical. To address this, the present study developed two different machine learning (ML) algorithms to predict Kamlet-Taft parameters for designer solvents using quantum chemically derived input features. The ML models developed in the present study showed accurate predictions with high R 2 and low RMSE values. Further, in the context of present interest in the circular bioeconomy, the relationship between the basicities and acidities of designer solvents and their ability to dissolve lignin and carbon dioxide (CO 2 ) is discussed. Our method thus guides the design of effective solvents with optimal Kamlet-Taft parameter values dissolving and converting biomass and CO 2 into valuable chemicals.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2023

Disentangling Electronic Transport and Hysteresis at Individual Grain Boundaries in Hybrid Perovskites via Automated Scanning Probe Microscopy

Underlying the rapidly increasing photovoltaic efficiency and stability of metal halide perovskites (MHPs) is the advancement in the understanding of the microstructure of polycrystalline MHP thin film. Over the past decade, intense efforts have been aimed at understanding the effect of microstructures on MHP properties, including chemical heterogeneity, strain disorder, phase impurity, etc. It has been found that grain and grain boundary (GB) are tightly related to lots of microscale and nanoscale behavior in MHP thin films. Atomic force microscopy (AFM) is widely used to observe grain and boundary structures in topography and subsequently to study the correlative surface potential and conductivity of these structures. For now, most AFM measurements have been performed in imaging mode to study the static behavior; in contrast, AFM spectroscopy mode allows us to investigate the dynamic behavior of materials, e.g., conductivity under sweeping voltage. However, a major limitation of AFM spectroscopy measurements is that they require manual operation by human operators, and as such only limited data can be obtained, hindering systematic investigations of these microstructures. Here, we designed a workflow combining the conductive AFM measurement with a machine learning (ML) algorithm to systematically investigate grain boundaries in MHPs. The trained ML model can extract GBs locations from the topography image, and the workflow drives the AFM probe to each GB location to perform a current–voltage (IV) curve automatically. Then, we are able to have IV curves at all GB locations, allowing us to systematically understand the property of GBs. Using this method, we discovered that the GB junction points are less conductive, potentially more photoactive, and can play critical roles in MHP stability, while most previous works only focused on the difference between GB and grains.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2020

Machine learning assisted hybrid models can improve streamflow simulation in diverse catchments across the conterminous US

Incomplete representations of physical processes often lead to structural errors in process-based (PB) hydrologic models. Machine learning (ML) algorithms can reduce streamflow modeling errors but do not enforce physical consistency. As a result, ML algorithms may be unreliable if used to provide future hydroclimate projections where climates and land use patterns are outside the range of training data. Here we test hybrid models built by integrating PB model outputs with a ML algorithm known as Long Short-Term Memory (LSTM) network on their ability to simulate streamflow in 531 catchments representing diverse conditions across the Conterminous United States. Model performance of hybrid models as measured by Nash-Sutcliffe efficiency (NSE) improved relative to standalone PB and LSTM models. More importantly, hybrid models provide highest improvement in catchments where PB models fail completely (i.e., NSE < 0). However, all models performed poorly in catchments with extended low flow periods, suggesting need for additional research.

54 ENVIRONMENTAL SCIENCES↗