Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “extrapolation”

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 325 records · Page 18
All sourcesNASA NTRSDOE OSTI
NASA NTRS1993

Improved Ephemerides of Pluto

The history of the Pluto ephemerides created at JPL is given. The uncertainties of present and possible future ephemerides are illustrated, and it is shown how rapidly the error grows for any present-day ephemeris of Pluto which is extrapolated into the future - tens of thousands of kilometers after only a decade. Continuing the observations into the future will not only reduce the extrapolation time but time will provide a substantial improvement to the ephemeris itself.

ephemerides↗
DOE OSTI2024

Disentangling Acid‐Base Chemistry through Blue Shifting Hydrogen Bond Contributions

The blue shifting of vibrational frequencies in hydrogen bonded molecules, as observed in aqueous environments, has been attributed to local partial charge transfer from solvation. Here, in this study, we extrapolate the blue shift model to the stronger ionic interactions between hydrogen bond acceptors associated with protonation through augmented pH levels and competitive interactions with counter ion pairing. The chemical model we utilize in this work is the aqueous pyridine‐pyridinium equilibrium to characterize the blue shifts observed in the pyridinium chloride ionic system. The observed agreement between observed experimental and calculated spectral shifts shows that the blue shifting model can be extrapolated to stronger interactions and accurately describe the nature of the hydrogen bond.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2024

Predicting responses to climate change using a joint species, spatially dependent physiologically guided abundance model

Abstract Predicting the effects of warming temperatures on the abundance and distribution of organisms under future climate scenarios often requires extrapolating species–environment correlations to climatic conditions not currently experienced by a species, which can result in unrealistic predictions. For poikilotherms, incorporating species' thermal physiology to inform extrapolations under novel thermal conditions can result in more realistic predictions. Furthermore, models that incorporate species and spatial dependencies may improve predictions by capturing correlations present in ecological data that are not accounted for by predictor variables. Here, we present a joint species, spatially dependent physiologically guided abundance (jsPGA) model for predicting multispecies responses to climate warming. The jsPGA model uses a basis function approach to capture both species and spatial dependencies. We apply the jsPGA model to predict the response of eight fish species to projected climate warming in thousands of lakes in Minnesota, USA. By the end of the century, the cold‐adapted species was predicted to have high probabilities of extirpation across its current range—with 10% of lakes currently inhabited by this species having an extirpation probability >0.90. The remaining species had varying levels of predicted changes in abundance, reflecting differences in their thermal physiology. Though the model did not identify many strong species dependencies, the variation in estimated spatial dependence across species suggested that accounting for both dependencies was important for predicting the abundance of these fishes. The jsPGA model provides a new tool for predicting changes in the abundance, distribution, and extirpation probability of poikilotherms under novel thermal conditions.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2024

Exploring the Tl 2H 2 potential energy surface: A comparative analysis with group 13 systems and experiment

Abstract Thallium chemistry is experiencing unprecedented importance. Therefore, it is valuable to characterize some of the simplest thallium compounds. Stationary points along the singlet and triplet TlH potential energy surface have been characterized. Stationary point geometries were optimized with the CCSD(T)/aug‐cc‐pwCVQZ‐PP method. Harmonic vibrational frequencies were computed at the same level of theory while anharmonic vibrational frequencies were computed at the CCSD(T)/aug‐cc‐pwCVTZ‐PP level of theory. Final energetics were obtained with the CCSDT(Q) method. Basis sets up to augmented quintuple‐zeta cardinality (aug‐cc‐pwCV5Z‐PP) were employed to obtain energetics in order to extrapolate to the complete basis set limits using the focal point approach. Zero‐point vibrational energy corrections were appended to the extrapolated energies in order to determine relative energies at 0 K. It was found that the planar dibridged isomer lies lowest in energy while the linear structure lies highest in energy. The results were compared to other group 13 MH (M = B, Al, Ga, In, and Tl) theoretical studies and some interesting variations are found. With respect to experiment, incompatibilities exist.

Chemistry↗
DOE OSTI2021

Entanglement and confinement in coupled quantum systems

We study some general properties of coupled quantum systems. We consider simple interactions between two copies of identical Hamiltonians such as the SYK model, Pauli spin chains with random magnetic field and harmonic oscillators. Such couplings make the ground states close to the thermofield double states of the uncoupled Hamiltonians. For the coupled SYK model, we push the numerical computation further towards the thermodynamic limit so that an extrapolation in the size of the system is possible. We find good agreement between the extrapolated numerical result and the analytic result in the large-q limit. We also consider the coupled gauged matrix model and vector model, and argue that the deconfinement is associated with the loss of the entanglement, similarly to the previous observation for the coupled SYK model. The understanding of the microscopic mechanism of the confinement/deconfinement transition enables us to estimate the quantum entanglement precisely, and backs up the dual gravity interpretation which relates the deconfinement to the disappearance of the wormhole. Our results demonstrate the importance of the entanglement between the color degrees of freedom in the emergence of the bulk geometry from quantum field theory via holography.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
DOE OSTI2025

On tests for baby universes in AdS/CFT

To address a puzzle by Antonini and Rath — where a single CFT state has two bulk duals, one with a baby universe and one without — Engelhardt and Gesteau recently devised a test for baby universes in AdS/CFT. Using the extrapolate dictionary, they showed that the boundary dual of a bulk swap test favored bulk spacetimes without a baby universe, providing evidence against their semiclassical validity. However, recent work suggests that holographic maps should post-select on such closed universes, and we argue that this is consistent with the extrapolate dictionary. We therefore construct a new holographic map for bulk states with baby universes and use this to show that the swap test cannot distinguish between Antonini and Rath’s two candidate bulk duals. This not only allows for a valid semiclassical description of the baby universe, but also enables the application of recent techniques for including observers in holographic maps.

AdS-CFT correspondence↗
DOE OSTI2024

Modeling delayed thermal runaway in nitric acid-soaked cat litter mixed with radioactive waste

Thermal ignition of radioactive waste within a 55-gallon drum was simulated by using a pressure-dependent waste decomposition model (Hobbs et al. in Process Saf Environ Prot https://doi.org/10.1016/j.psep.2022.09.047, 2022) calibrated with data from full-scale drum experiments (Parker et al. in The thermolytic response of a surrogate RNS waste mixture at the drum scale. Los Alamos National Laboratory Report LA-UR-16-21760, 2016) and validated with experiments from multiple laboratories (Hobbs et al. in Thermal analysis of aged nitric acid-soaked kitty litter in TRU waste drums-23370.WM2023 Conference, Phoenix, AZ, 2023). The acceleration of nitric acid chemistry reacting with an organic cat litter leading to thermal ignition was likely triggered by a restricted vent in the drum. Here, we address whether the form of the rate equation in (Hobbs et al. in Process Saf Environ Prot https://doi.org/10.1016/j.psep.2022.09.047, 2022) is sufficient to extrapolate thermal ignition within aged drums of similar content that have been stored in Texas for over nine years by investigating four different reaction rate forms for waste decomposition. A critical reaction rate reduction analysis is performed on each of these models to determine if delayed thermal runaway within vented aged waste is possible after nine years. We found that a pressure-dependent first-order rate expression not only predicted the accidental ignition of the waste drum, but the form also matches multiple experiments from different laboratories. Even though the waste composition decreases over time, the model predicts that acceleration leading to thermal runaway is possible if the waste is confined, even after 9 years. In conclusion, waste containing oxidizers such as nitric acid should not be mixed with organic adsorbents, especially if the waste is confined.

cookoff↗
DOE OSTI2020

Data-Driven Constitutive Model for the Inelastic Response of Metals: Application to 316H Steel

Here, predictions of the mechanical response of structural elements are conditioned by the accuracy of constitutive models used at the engineering length-scale. In this regard, a prospect of mechanistic crystal-plasticity-based constitutive models is that they could be used for extrapolation beyond regimes in which they are calibrated. However, their use for assessing the performance of a component is computationally onerous. To address this limitation, a new approach is proposed whereby a surrogate constitutive model (SM) of the inelastic response of 316H steel is derived from a mechanistic crystal plasticity-based polycrystal model tracking the evolution of dislocation densities on all slip systems. The latter is used to generate a database of the expected plastic response and dislocation content evolution associated with several instances of creep loading. From the database, a SM is developed. It relies on the use of orthogonal polynomial regression to describe the evolution of the dislocation content. The SM is then validated against predictions of the dead load creep response given by the polycrystal model across a range of temperatures and stresses. When the SM is used to predict the response of 316H during complex non monotonic loading, extrapolating to new loading conditions, it is found that predictions compare particularly well against those from the physics-based polycrystal model.

36 MATERIALS SCIENCE↗
DOE OSTI2022

Data coverage assessment on neural network based digital twins for autonomous control system

We report in a recently developed Nearly Autonomous Management and Control (NAMAC) system, neural networks (NNs) are used to develop digital twins for diagnosis (DT-Ds). However, NNs are not usually considered extrapolation models and may result in large errors if they are applied to unseen data outside the training data (uncovered). In this study, we propose a data coverage assessment (DCA) to determine if the NN-based DT-Ds are extrapolated based on their epistemic uncertainty. The uncertainty quantification algorithms and uncertainty thresholds are selected based on the confusion matrix of classifying evaluation data into covered or uncovered data. To demonstrate the adaptability of the proposed framework, we applied it to a basic feedforward neural network and a more advanced recurrent neural network based on a more nonlinear database. Case studies show that the proposed framework can distinguish unseen data for both basic and advanced applications with proper uncertainty quantification algorithms and thresholds.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗
DOE OSTI2024

Efficient continuous Energy-Multigroup hybrid depletion scheme using the Shift Monte Carlo code. Part I: Energy condensation sensitivity analysis

Monte Carlo (MC) codes coupled to depletion solvers are increasingly used to provide high fidelity fuel cycle modeling capabilities. Here, these coupled depletion-MC tools produce accurate results in general but can experience nonphysical spatial oscillations when time steps are large or when a system’s dominance ratio approaches unity. Two substepping techniques have been developed previously to remedy and dampen these spatial oscillations without needing to reduce step sizes. The first approach relied on higher-order techniques to account for spectral changes within steps (extrapolation and interpolation techniques). The second approach used the first order perturbation (FOP) theory to account for the change in the one-group spatial flux distribution within steps. This paper develops a hybrid depletion methodology which, in a way, combines how the flux is handled in both substepping techniques. Specifically, the multigroup (MG) MC Shift code is used to update the flux distribution within steps rather than a one-group FOP solver. A fully reflected pincell is investigated, which is not spatially dependent in the MG representation. Thus, the analysis in this paper is an initial demonstration of hybrid depletion. An upcoming companion paper will focus on how the hybrid depletion dampens spatial oscillations. The hybrid depletion approach is verified to be consistent with previous constant extrapolation depletion (CED) methods. This paper finds that the hybrid CED exhibits some error in the eigenvalue and one group constants within macro steps. To address this discrepancy, a simple interpolation scheme (CELI) is investigated. This work found that CELI sufficiently addresses the discrepancy in spectrum for macro steps up to 100 days. Overall, this work demonstrates that the hybrid depletion method can significantly reduce the number of high fidelity MC executions in a MC-coupled depletion with an acceptable eigenvalue error.

29 ENERGY PLANNING, POLICY, AND ECONOMY↗
DOE OSTI2021

Empirical estimation of densities in NaCl-KCl-UCl 3 and NaCl-KCl-YCl 3 molten salts using Redlich-Kister expansion

Densities of molten KCl-NaCl-UCl 3 and KCl-NaCl-YCl 3 ternary systems have been estimated using a multidimensional Redlich-Kister model. Temperature and composition dependent Redlich-Kister functions have been used to generate binary interaction parameters in the outlined ternary salt systems. These binary interactions have been used in the extrapolation to ternary system densities. The results of the density extrapolations by Muggianu interpolation scheme provide agreement within 2–3% for the NaCl-KCl-YCl 3 liquids and 11% in NaCl-KCl-UCl 3 liquids compared to the available experimental data. Modeling NaCl-KCl-UCl 3 molten phase density with a ternary interaction parameter improved the agreement within 4%. Thermophysical modeling used in this study has shown promising results for use in other material properties, such as viscosity, thermal conductivity, and heat capacity of the molten salts. Lastly, the outlined modeling method applied in these specific molten salt ternaries can be used for quaternary or higher multicomponent molten salt systems.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2024

On the applicability of the Redlich-Kister framework for viscosity estimation of molten halide salt mixtures

For molten halide salt mixtures already being utilized or under consideration for carbon-free energy production systems, it is crucial that their viscosity is well understood so that system thermal hydraulics can be reliably assessed. Because of the difficulty in accurately measuring molten halide viscosity and the sheer size of the matrix of possible higher order salt mixtures that may be of interest to the energy industry, there are several gaps in the quantified understanding of molten halide viscosity across this matrix. As such, both first-principles and semi-empirical modeling techniques may be crucial for rapidly assessing this broad, complex compositional domain. Herein, the Redlich-Kister framework is applied to assess the feasibility of broadly interpolating and estimating the viscosity of several pseudobinary and pseudoternary molten halide salt systems that may be of key interest to the energy industry. The framework is based on the assumption that an ideal component and a nonideal component collectively describe the viscosity as a function of composition and temperature for a given molten halide system. Three different ideal models were considered for the ideal component, including Grunburg-Nissan, Katti-Chaudhri, and Gambill methods. Regarding the pseudobinary interpolations, the Redlich-Kister models with either the Grunburg-Nissan or Katti-Chaudhri models as the ideal component resulted in either highly (average error less than 5%) or reasonably (average error less than 15%) accurate interpolations of pseudobinary halide viscosity; BeF 2 - or UF 4 -bearing salts tended to result in reasonably accurate interpolations, whereas other pseudobinary mixtures tended to show high accuracy. Regarding the pseudoternary extrapolations, the Redlich-Kister framework shows reasonable success at estimating the extent to which a pseudoternary system may indicate deviations from ideal Grunburg-Nissan mixing, where discrepancies with comparative experimental data generally stay within 30%. Finally, the primary reasons identified for such discrepancies are (1) inaccuracy in the underlying experimental data, (2) different complexation behavior in the higher order systems compared to the pseudobinary subsystems, and (3) extrapolation into temperatures too far out of the domain, which is valid for the underlying experimental data feeding the Redlich-Kister model.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2025

Spectrophotometric determination of the stability of La hydroxyl complexes at near neutral to alkaline pH from 25 to 75 °C

The hydrolysis of rare earth elements (REE) potentially controls their mobility during fluid-rock interaction in a broad range of pH and temperature conditions. However, there is still a lack of thermodynamic data for modeling accurately the stability of REE hydroxyl complexes in hydrothermal aqueous fluids. Here, in this study, UV–Vis spectrophotometric experiments were conducted from 25 to 75 °C in near-neutral to alkaline NaOH-bearing aqueous solutions with varying lanthanum (La) concentrations (0 to ∼0.23 mmol/kg). The color indicator m-cresol purple was used to determine in situ pH and derive the average OH− ligand number ($\overrightarrow{n}$) and formation constants for the La hydroxyl complexes (LaOH 2+ , La(OH) 2 + , and La(OH) 3 0 ). From 25 to 50 °C, $\overrightarrow{n}$ ranges between ∼1 and 2 at pH from 7.0 to 9.3. At 75 °C, $\overrightarrow{n}$ ranges between ∼1.5 and 3 at pH from 6.3 to 8.8. These results suggest the predominance of LaOH 2+ and La(OH) 2 + complexes from 25 to 50 °C, and an increased predominance of La(OH) 3 0 at 75 °C. The cumulative formation constants (β n °, n = 1 to 3) are derived for the reaction La 3+ + nOH − = La(OH) n 3-n , and fitted between 25 and 250 °C by combining the UV–Vis and literature solubility data. The resulting logβ n ° are expressed as function of temperature (T in Kelvin): logβ 1 ° = −1.786 + 0.0133 T + 1.049·10 3 /T; logβ 2 ° = −5.797 + 0.0267 T + 2.713·10 3 /T; logβ 3 ° = 6.435 + 0.0223 T + 512.7/T. A comparison between these new fits and existing extrapolations using the Helgeson-Kirkham-Flowers equation of state indicates significant differences in the predicted hydrolysis of La. The latter extrapolations should therefore be updated for the hydrolysis of REE.

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

Modular machine learning-based elastoplasticity: Generalization in the context of limited data

The development of highly accurate constitutive models for materials that undergo path-dependent processes continues to be a complex challenge in computational solid mechanics. Challenges arise both in considering the appropriate model assumptions and from the viewpoint of data availability, verification, and validation. Recently, data-driven modeling approaches have been proposed that aim to establish stress-evolution laws that avoid user-chosen functional forms by relying on machine learning representations and algorithms. However, these approaches not only require a significant amount of data but also need data that probes the full stress space with a variety of complex loading paths. Furthermore, they rarely enforce all necessary thermodynamic principles as hard constraints. Hence, they are in particular not suitable for low-data or limited-data regimes, where the first arises from the cost of obtaining the data and the latter from the experimental limitations of obtaining labeled data, which is commonly the case in engineering applications. In this work, we discuss a hybrid framework that can work on a variable amount of data by relying on the modularity of the elastoplasticity formulation where each component of the model can be chosen to be either a classical phenomenological or a data-driven model depending on the amount of available information and the complexity of the response. The method is tested on synthetic uniaxial data coming from simulations as well as cyclic experimental data for structural materials. The discovered material models are found to not only interpolate well but also allow for accurate extrapolation in a thermodynamically consistent manner far outside the domain of the training data. This ability to extrapolate from limited data was the main reason for the early and continued success of phenomenological models and the main shortcoming in machine learning-enabled constitutive modeling approaches. Training aspects and details of the implementation of these models into Finite Element simulations are discussed and analyzed.

42 ENGINEERING↗
DOE OSTI2023

Machine learning-based surrogate models and transfer learning for derivative free optimization of HT-PEM fuel cells

Widespread adoption of high-temperature polymer electrolyte membrane electrochemical systems, such as fuel cells (HT-PEMFCs), requires models and computational tools for accurate optimization and guiding new materials for enhancing performance and durability. In this contribution, knowledge-based modelling and data-driven modelling are combined using Few-Shot Learning and implementing an Automated Machine Learning framework for the generation of Machine Learning-based surrogate models. Applicability of the resulting model for derivative-free optimization is demonstrated. Additionally, a way of considering extrapolation in the optimization task is presented. Results show that although extrapolation is needed to achieve better solutions during optimization, it can be monitored and managed. As a result, tuning the electrode ionomer binder's properties, such as ionic conductivity, in the fuel cell represents a promising pathway for improving HT-PEMFC performance.

08 HYDROGEN↗
DOE OSTI2021

Oxidation and Associated Pore Structure Modification During Experimental Alteration of Granite

Weathering plays a crucial role in a number of environmental processes, and the microstructure and evolution of multi-scale pore space is a critically important factor in the weathering of geological formations. In igneous rocks the infiltration of meteoric water into initially relatively dry material can cause the rock to begin to disaggregate, increasing porosity and surface area, and allowing further disaggregation and weathering to occur. These processes, in turn, allow biota to colonize the rock, further enhancing the weathering rate. In some igneous rocks this may be driven by oxidation of primary minerals. One such mineral, biotite, has been repeatedly mentioned as a cause of cracking during oxidation. However, the scale-dependence of the processes by which this occurs are poorly understood. We cannot, therefore, accurately extrapolate laboratory reaction rates to the field in predictive numerical models. In order to better understand the effects of oxidation and test the hypothesis that fracture and disaggregation are initiated by swelling of oxidizing biotites, we reacted granite cores in a selenic acid-rich aqueous solution at 200°C for up to 438 days. Elevated temperatures and selenic acid were used to provide relatively fast reaction rates and highly oxidizing conditions in sealed reaction vessels. These experiments were analyzed using a combination of imaging, X-ray diffraction, Mössbauer spectroscopy, and small- and ultra-small angle neutron scattering to interrogate porosity and microfracture formation. The experimental results show little observable biotite swelling, but significant transport and growth of iron oxides and/or clays along grain boundaries throughout the sample. Significant increases in porosity were also observed at the sample rim, likely associated with feldspar alteration. Fractures and transport were observed throughout the core, suggesting that stresses due to crystallization pressures caused by the growing iron phases may be the initiating factors in granite weathering, possibly followed by biotite swelling after sufficient permeability is achieved.

Anovitz, Lawrence M.↗
DOE OSTI2021

Experimental and theoretical determinations of hydrogen isotopic equilibrium in the system CH 4 —H 2 —H 2 O from 3 to 200°C

The stable isotopic composition of methane (CH 4 ) is commonly used to fingerprint natural gas origins. Over the past 50 years, there have been numerous proposals that both microbial and thermogenic CH 4 can form in or later attain hydrogen isotopic equilibrium with water (H 2 O) and carbon isotopic equilibrium with carbon dioxide (CO 2 ). Evaluation of such proposals requires knowledge of the equilibrium fractionation factors between CH 4 and H 2 O or CO 2 at the temperatures where microbial and thermogenic CH 4 form in or are found in the environment, which is generally less than 200°C. Experimental determinations of these fractionation factors are only available above 200°C, requiring extrapolation of these results beyond the calibrated range or the use of theoretical calculations at lower temperatures. Here, we provide a calibration of the equilibrium hydrogen isotopic fractionation factor for CH 4 and hydrogen gas (H 2 ) ( D α CH4(g)–H2(g) ) based on experiments using γ-Al 2 O 3 and Ni catalysts from 3 to 200°C. Results were regressed as a 2 nd order polynomial of 1000 × ln D α CH4(g)–H2(g) vs. 1/T (K -1 ) yielding: 1000 × l n D α C H 4 ( g ) - H 2 ( g ) = 3.5317 × 10 7 T 2 + 2.7749 × 10 5 T - 179.48 We combine this calibration with previous experimental determinations of hydrogen isotope equilibrium between H 2 , H 2 O(g), and H 2 O(l) and we provide an interpolatable experimental calibration of 1000 × ln D α CH4(g)–H2O(l) from 3 to 200°C. Our resulting 4th order polynomial is the following equation: 1000 × l n D α C H 4 ( g ) - H 2 O l = - 7.9443 × 10 12 T 4 + 8.7772 × 10 10 T 3 - 3.4973 × 10 8 T 2 + 5.4398 × 10 5 T - 382.05 At 3°C, the value from our calibration differs by 93‰ relative to what would be calculated based on the extrapolation of the only experimental calibration currently available to temperatures below its calibrated range (lowest temperature of 200°C; Horibe and Craig, 1995). We additionally provide new theoretical estimates of hydrogen isotopic equilibrium between CH 4 (g), H 2 (g), and H 2 O(g) and carbon isotopic equilibrium between CH 4 (g) and CO 2 (g) using Path Integral Monte Carlo (PIMC) calculations. Our PIMC calculations for hydrogen isotopic equilibrium between CH 4 and H 2 agree 1:1 with our experiments. Finally, we compile carbon and hydrogen isotopic measurements of CH 4 , CO 2 , and H 2 O from various environmental systems and compare observed differences between carbon and hydrogen isotopes to those expected based on isotopic equilibrium. We find that isotopic compositions of some microbial gases from marine sedimentary, coalbed, and shale environments are consistent with those expected for CH 4 H 2 O(l) hydrogen and CH 4 CO 2 carbon isotopic equilibrium. In contrast, microbial terrestrial and pure culture gases are not consistent with both CH 4 H 2 O(l) hydrogen and CH 4 CO 2 carbon isotopic equilibrium. Overall, these results are explained qualitatively using previously developed conceptual models that link free energy gradients available to microorganisms to the degree that their enzymes can promote isotope-exchange reactions between CH 4 , CO 2 , and H 2 O.

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

A multiscale recurrent neural network model for predicting energy production from geothermal reservoirs

Optimization of energy production from geothermal reservoirs requires reliable prediction of energy production performance under alternative operation and development scenarios. Traditionally, reservoir simulation models are used for the evaluation and screening of alternative production and development plans. However, simulation models require extensive data collection and modeling efforts and are time-consuming to build, run, and update. Data-driven predictive models, on the other hand, can serve as efficient prediction tools that can be used for decision support and management of daily operations and surveillance activities. Data-driven models become particularly attractive when a reservoir simulation model for a field does not exist and/or is difficult to build. Machine learning (ML)-based data-driven models that have recently become popular in several fields exploit statistical patterns and relations in training data to generate predictions. As such, they tend to perform better in interpolation problems (that is, prediction within the training data range) than when they are used to extrapolate beyond the training data. Production data from geothermal reservoirs tend to exhibit short-term variabilities as well as long-term trends, such as monotonically declining production temperatures. Capturing both short-term features and long-term trends with ML-based models is not trivial. We evaluate the use of recurrent neural networks (RNN) for the prediction of energy production from geothermal reservoirs. RNN is a class of ML architectures that are used to represent and predict sequential/dynamic data. Thus, it can be challenging to apply RNN to problems where long-term trends must be captured and extrapolation beyond the training data range is needed. We introduce the multiscale RNN architecture to extend the application of RNN to detect and predict both short-term variabilities and long-term trends in geothermal data. The developed architecture consists of a long-term component to only capture low-frequency data patterns, and a short-term component to detect features with higher frequency and more nonlinearity. The final prediction is obtained by combining the long-term and short-term predictions. Both synthetic and field data are used to evaluate the presented multiscale RNN model. The prediction performance of the multiscale RNN is compared against those obtained from the regular RNN and the autoregressive (AR) model. The results suggest that the multiscale architecture improves the long-term prediction performance of the regular RNN and enhances its robustness against noise.

15 GEOTHERMAL ENERGY↗