Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “structured model reduction”

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 1,369 records · Page 76
All sourcesNASA NTRSDOE OSTI
NASA NTRS2002

Noise Transmission Characteristics of Damped Plexiglas Windows

Most general aviation aircraft utilize single layer plexiglas material for the windshield and side windows. Adding noise control treatments to transparent panels is a challenging problem. In this paper, damped plexiglas windows are evaluated for replacement of conventional windows in general aviation aircraft to reduce the structure-borne and airborne noise transmitted into the interior. In contrast to conventional solid windows, the damped plexiglas window panels are fabricated using two or three layers of plexiglas with transparent viscoelastic damping material sandwiched between the layers. Results from acoustic tests conducted in the NASA Langley Structural Acoustic Loads and Transmission (SALT) facility are used to compare different designs of the damped plexiglas panels with solid windows of the same nominal thickness. Comparisons of the solid and damped plexiglas panels show reductions in the radiated sound power of up to 8 dB at low frequency resonances and as large as 4.5 dB over a 4000 Hz bandwidth. The weight of the viscoelastic treatment was approximately 1% of the panel mass. Preliminary FEM/BEM modeling shows good agreement with experimental results for radiated sound power.

Gibbs, Gary P.↗
NASA NTRS2018

Microphysics and Radiation Effect of Dust on Saharan Air Layer: An HS3 Case Study

A Saharan air layer (SAL) event associated with a nondeveloping African easterly wave (AEW) over the main development region of the eastern Atlantic was sampled by the NASA Global Hawk aircraft on 24-25 August 2013 during the NASA Hurricane and Severe Storm Sentinel (HS3) campaign and was simulated with the NASA Unified Weather Research and Forecasting (NU-WRF) Model. Airborne, ground-based, and spaceborne measurements were used to evaluate the model performance. The microphysical and radiative effects of dust and other aerosols on the SAL structure and environment were investigated with the factor-separation method. The results indicate that relative to a simulation without dust-radiative and microphysical impacts, Saharan dust and other aerosols heated the SAL air mainly through shortwave heating by the direct aerosol-radiation (AR) effect, resulting in a warmer (up to 0.6 K) and drier (up to 5% RH reduction) SAL and maintaining the strong temperature inversion at the base of the SAL in the presence of predominant longwave cooling. Radiative heating of the dust accentuated a vertical circulation within the dust layer, in which air rose (sank) in the northern (southern) portions of the dust layer. Furthermore, above and to the south of the dust layer, both the microphysical and radiative impacts of dust tended to counter the vertical motions associated with the Hadley circulation, causing a small weakening and southward shift of convection in the intertropical convergence zone (ITCZ) and reduced anvil cloud to the north. Changes in moisture and cloud/precipitation hydrometeors were largely driven by the dust induced changes in vertical motion. Dust strengthened the African easterly jet by up to ~1ms(exp -1) at the southern edge of the jet, primarily through the AR effect, and produced modest increases in vertical wind shear within and in the vicinity of the dust layer. These modulations of the SAL and AEW environment clearly contributed to the nondevelopment of this AEW.

Tao, Zhining↗
DOE OSTI2015

Development of highly accurate approximate scheme for computing the charge transfer integral

The charge transfer integral is a key parameter required by various theoretical models to describe charge transport properties, e.g., in organic semiconductors. The accuracy of this important property depends on several factors, which include the level of electronic structure theory and internal simplifications of the applied formalism. The goal of this paper is to identify the performance of various approximate approaches of the latter category, while using the high level equation-of-motion coupled cluster theory for the electronic structure. The calculations have been performed on the ethylene dimer as one of the simplest model systems. By studying different spatial perturbations, it was shown that while both energy split in dimer and fragment charge difference methods are equivalent with the exact formulation for symmetrical displacements, they are less efficient when describing transfer integral along the asymmetric alteration coordinate. Since the “exact” scheme was found computationally expensive, we examine the possibility to obtain the asymmetric fluctuation of the transfer integral by a Taylor expansion along the coordinate space. By exploring the efficiency of this novel approach, we show that the Taylor expansion scheme represents an attractive alternative to the “exact” calculations due to a substantial reduction of computational costs, when a considerably large region of the potential energy surface is of interest. Moreover, we show that the Taylor expansion scheme, irrespective of the dimer symmetry, is very accurate for the entire range of geometry fluctuations that cover the space the molecule accesses at room temperature.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2019

Thermally Driven Structure and Performance Evolution of Atomically Dispersed FeN 4 Sites for Oxygen Reduction

Abstract FeN 4 moieties embedded in partially graphitized carbon are the most efficient platinum group metal free active sites for the oxygen reduction reaction in acidic proton‐exchange membrane fuel cells. However, their formation mechanisms have remained elusive for decades because the Fe−N bond formation process always convolutes with uncontrolled carbonization and nitrogen doping during high‐temperature treatment. Here, we elucidate the FeN 4 site formation mechanisms through hosting Fe ions into a nitrogen‐doped carbon followed by a controlled thermal activation. Among the studied hosts, the ZIF‐8‐derived nitrogen‐doped carbon is an ideal model with well‐defined nitrogen doping and porosity. This approach is able to deconvolute Fe−N bond formation from complex carbonization and nitrogen doping, which correlates Fe−N bond properties with the activity and stability of FeN 4 sites as a function of the thermal activation temperature.

Li, Jiazhan↗
DOE OSTI2023

Double-Atom Catalysts Featuring Inverse Sandwich Structure for CO 2 Reduction Reaction: A Synergetic First-Principles and Machine Learning Investigation

Electrocatalytic CO 2 reduction reactions (CO 2 RR) based on scalable and highly efficient catalysis provide an attractive strategy for reducing CO 2 emissions. Here in this work, we combined first-principles density functional theory (DFT) and machine learning (ML) to comprehensively explore the potential of double-atom catalysts (DACs) featuring an inverse sandwich structure anchored on defective graphene (gra) to catalyze CO 2 RR to generate C 1 products. We started with five homonuclear M 2 ⊥gra (M = Co, Ni, Rh, Ir, and Pt), followed by 127 heteronuclear MM'⊥gra (M = Co, Ni, Rh, Ir, and Pt, M' = Sc–Au). Stable DACs were screened by evaluating their binding energy, formation energy, and dissolution potential of metal atoms, as well as conducting first-principles molecular dynamics simulations with and without solvent water molecules. Based on DFT calculations, Rh 2 ⊥gra DAC was found to outperform the other four homonuclear DACs and the Rh-based single- and double-atom catalysts of noninverse sandwich structures. Out of the 127 heteronuclear DACs, 14 were found to be stable and have good catalytic performance. An ML approach was adopted to correlate key factors with the activity and stability of the DACs, including the sum of radii of metal and ligand atoms (d M–M' , d M–C , and d M'–C ), the sum and difference of electronegativity of two metal atoms (P M + P M' , P M – P M '), the sum and difference of first ionization energy of two metal atoms (I M + I M' , I M – I M '), the sum and difference of electron affinity of two metal atoms (A M + A M' , A M – A M '), and the number of d-electrons of the two metal atoms (Nd). The obtained ML models were further used to predict 154 potential electrocatalysts out of 784 possible DACs featuring the same inverse sandwich configuration. Overall, this work not only identified promising CO 2 RR DACs featuring the reported inverse sandwich structure but also provided insights into key atomic characteristics associated with high CO 2 RR activity.

30 DIRECT ENERGY CONVERSION↗
DOE OSTI2019

Experimental and Theoretical Trends of PGM-Free Electrocatalysts for the Oxygen Reduction Reaction with Different Transition Metals

Replacement of platinum-based electrocatalysts to facilitate the oxygen reduction reaction (ORR) in proton exchange fuel cells remains an outstanding challenge. Significant progress in the development of platinum group metal-free (PGM-free) electrocatalysts has been reported but a fundamental understanding of how these materials catalyze the ORR remains elusive. In this work, we report our efforts to synthesize and characterize PGM-free electrocatalysts with different transition metals (M = Fe, Co, and Mn) in order to better understand the most plausible active site structures and ORR reaction pathways. Our findings indicate that PGM-free electrocatalysts synthesized via a dual nitrogen precursor, polyaniline and cyanamide, synthesis process but with varied transition metal precursors resulted in the following ORR activity trend: Fe > Co > Mn. Furthermore, similar densities of active sites across the three synthesized materials were calculated using an established molecular probe technique. The relative experimental activities are consistent with trends determined via density functional theory (DFT) modeling that suggests spontaneous OH ligand modification of active sites.

25 ENERGY STORAGE↗
DOE OSTI2017

Revealing Anisotropic Spinel Formation on Pristine Li- and Mn-Rich Layered Oxide Surface and Its Impact on Cathode Performance

Surface properties of cathode particles play important roles in the transport of ions and electrons and they may ultimately dominate cathode's performance and stability in lithium-ion batteries. Through the use of carefully prepared Li 1.2 Ni 0.13 Mn 0.54 Co 0.13 O 2 crystal samples with six distinct morphologies, surface transition-metal redox activities and crystal structural transformation are investigated as a function of surface area and surface crystalline orientation. Complementary depth-profiled core-level spectroscopy, namely, X-ray absorption spectroscopy, electron energy loss spectroscopy, and atomic-resolution scanning transmission electron microscopy, are applied in the study, presenting a fine example of combining advanced diagnostic techniques with a well-defined model system of battery materials. Here, we report the following findings: (1) a thin layer of defective spinel with reduced transition metals, similar to what is reported on cycled conventional secondary particles in the literature, is found on pristine oxide surface even before cycling, and (2) surface crystal structure and chemical composition of both pristine and cycled particles are facet dependent. Oxide structural and cycling stabilities improve with maximum expression of surface facets stable against transition-metal reduction. Finally, the intricate relationships among morphology, surface reactivity and structural transformation, electrochemical performance, and stability of the cathode materials are revealed.

36 MATERIALS SCIENCE↗
DOE OSTI2022

Finding Hidden Patterns in High Resolution Wind Flow Model Simulations

Wind flow data is critical in terms of investment decisions and policy making. High resolution data from wind flow model simulations serve as a supplement to the limited resource of original wind flow data collection. Given the large size of data, finding hidden patterns in wind flow model simulations are critical for reducing the dimensionality of the analysis. In this work, we first perform dimension reduction with two autoencoder models: the CNN-based autoencoder (CNN-AE) [1], and hierarchical autoencoder (HIER-AE) [2], and compare their performance with the Principal Component Analysis (PCA). We then investigate the super-resolution of the wind flow data. By training a Generative Adversarial Network (GAN) with 300 epochs, we obtained a trained model with 2× resolution enhancement. We compare the results of GAN with Convolutional Neural Network (CNN), and GAN results show finer structure as expected in the data field images. Also, the kinetic energy spectra comparisons show that GAN outperforms CNN in terms of reproducing the physical properties for high wavenumbers and is critical for analysis where high-wavenumber kinetics play an important role.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2018

Machine learning for graph-based representations of three-dimensional discrete fracture networks

Structural and topological information play a key role in modeling flow and transport through fractured rock in the subsurface. Discrete fracture network (DFN) computational suites such as dfnWorks (Hyman et al. Comput. Geosci. 84, 10–19 2015) are designed to simulate flow and transport in such porous media. Flow and transport calculations reveal that a small backbone of fractures exists, where most flow and transport occurs. Restricting the flowing fracture network to this backbone provides a significant reduction in the network’s effective size. However, the particle-tracking simulations needed to determine this reduction are computationally intensive. Such methods may be impractical for large systems or for robust uncertainty quantification of fracture networks, where thousands of forward simulations are needed to bound system behavior. In this paper, we develop an alternative network reduction approach to characterizing transport in DFNs, by combining graph theoretical and machine learning methods. We consider a graph representation where nodes signify fractures and edges denote their intersections. Using random forest and support vector machines, we rapidly identify a subnetwork that captures the flow patterns of the full DFN, based primarily on node centrality features in the graph. Our supervised learning techniques train on particle-tracking backbone paths found by dfnWorks, but run in negligible time compared to those simulations. In conclusion, we find that our predictions can reduce the network to approximately 20% of its original size, while still generating breakthrough curves consistent with those of the original network.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2020

Calcium nitrate: A chemical admixture to inhibit aggregate dissolution and mitigate expansion caused by alkali-silica reaction

Alkali-silica reaction (ASR) is a significant cause of the degradation of concrete structures. ASR results from the dissolution of reactive silicate aggregates and the subsequent formation of damaging, expansive ASR gels. Despite the prevalence of ASR, so far only lithium-based chemical admixtures have been shown to be successful in mitigating its deleterious effects. Herein, dissolved calcium nitrate (Ca(NO 3 ) 2 : CN) is demonstrated to reduce ASR-induced expansion caused by the use of sodium borosilicate (NBS) glass as a model reactive aggregate. The significance of reductions of NBS-glass dissolution and mortar-bar expansion rates is discussed, with special focus on highlighting the mechanism by which calcium nitrate mitigates ASR, i.e., by the inhibition of aggregate dissolution. This study reveals a new and cost-effective pathway to suppress aggregate dissolution and ASR in concrete based on the dosage of soluble alkaline-earth additives.

36 MATERIALS SCIENCE↗
DOE OSTI2019

Local atomic structure and lattice defect analysis in heavily Co-doped ZnS thin films using X-ray absorption fine structure spectroscopy

Iin this work, X-ray absorption fine structure (XAFS) spectroscopy was used to address both the local atomic structure and the local electronic structure around Zn, Co, and S in Co-doped ZnS thin films. X-ray absorption near-edge spectroscopy (XANES) revealed changes in the strong pre-edge feature of the S K-edge spectra that depend on the Co concentration and are related to atomic bound-state transitions through hybridization between $3d$ transition metal and $sp$ host semiconductor electronic states. These changes reveal intrinsic effects, such as S vacancies and dopant effects. From the Co K-edge XANES spectra, substitution of Co 2+ into ZnS is observed. The local atomic structure around the dopant, obtained by XAFS spectroscopy, indicates a reduction of the Co–S bond length. Optical analysis shows deep absorption centers or intermediate states in the bandgap, ascribed to $d$-$d$ transitions in the Co atoms that modify the electronic band structure and cause a decrease in the bandgap as a function of Co concentration. This work provides direct evidence of dopant effects in heavily Co-doped ZnS, including bond distances and Debye-Waller factors, providing key elements to generate a realistic atomic structural model, which is crucial to understand the observed electronic properties of this material.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2022

The influence of temperature on the strain-hardening behavior of Fe-22/25/28Mn-3Al-3Si TRIP/TWIP steels

In this work, the influence of temperature and stacking fault energy (SFE) on the strain-hardening behavior and critical resolved shear stress for twinning was investigated for three Fe–22/25/28Mn–3Al–3Si wt.% transformation- and twinning-induced plasticity (TRIP/TWIP) steels. The SFEs were calculated by two different methods, density functional theory and statistical thermodynamic modeling. The dislocation structure, observed at low levels of plastic deformation, transitions from “planar” to “wavy” dislocation glide with an increase in temperature, Mn content, and/or SFE. The change in dislocation glide mechanisms from planar to wavy reduces the strain hardening rate, in part due to fewer planar obstacles and greater cross slip activity. Additionally, the alloys exhibit a large decrease in strength and ductility with increasing temperature from 25 to 200 °C, attributed to a substantial reduction in the thermally activated component of the flow stress, predominate suppression of TRIP and TWIP, and a significant increase in the critical resolved shear stress for mechanical twinning. Interestingly, the increase in SFE with temperature had a rather minor influence on the critical resolved shear stress for mechanical twinning, and other temperature dependent factors which likely play a more dominant role are discussed.

36 MATERIALS SCIENCE↗
DOE OSTI2018

RalGPS2 is involved in tunneling nanotubes formation in 5637 bladder cancer cells

Highlights: • RalGPS2 is expressed in 5637 cancer cells and is essential for their growth. • RalGPS2 is required for TNT formation. • The overexpression of RalGPS2 PH domain markedly increased the number and length of TNT. • Interaction between RalA and Sec5 is required for TNT formation in 5637 cells. • RalGPS2 interacts with LST1 and RalA leading to TNT formation. RalGPS2 is a Ras-independent Guanine Nucleotide Exchange Factor (GEF) for RalA containing a PH domain and an SH3-binding region and it is involved in several cellular processes, such as cytokinesis, control of cell cycle progression, differentiation, cytoskeleton organization and rearrangement. Up to now, few data have been published regarding RalGPS2 role in cancer cells, and its involvement in bladder cancer is yet to be established. In this paper we demonstrated that RalGPS2 is expressed in urothelial carcinoma-derived 5637 cancer cells and is essential for cellular growth. These cells produces thin membrane protrusions that displayed the characteristics of actin rich tunneling nanotubes (TNTs) and here we show that RalGPS2 is involved in the formation of these cellular protrusions. In fact the overexpression of RalGPS2 or of its PH-domain increased markedly the number and the length of nanotubes, while the knock-down of RalGPS2 caused a strong reduction of these structures. Moreover, using a series of RalA mutants impaired in the interaction with different downstream components (Sec5, Exo84, RalBP1) we demonstrated that the interaction of RalA with Sec5 is required for TNTs formation. Furthermore, we found that RalGPS2 interacts with the transmembrane MHC class III protein leukocyte specific transcript 1 (LST1) and RalA, leading to the formation of a complex which promotes TNTs generation. These findings allow us to add novel elements to molecular models that have been previously proposed regarding TNTs formation.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2022

Advancing Chemical Separations: Unraveling the Structure and Dynamics of Phase Splitting in Liquid–Liquid Extraction

Liquid-liquid extraction (LLE), the go-to process for a variety of chemical separations, is limited by spontaneous organic phase splitting upon sufficient solute loading, called third phase formation. In this study we explore the applicability of critical phenomena theory to gain insight into this deleterious phase behavior with the goal of improving separations efficiency and minimizing waste. Here, a series of samples representative of rare earth purification were constructed to include each of one light and one heavy lanthanide (cerium and lutetium) paired with one of two common malonamide extractants (DMDOHEMA and DMDBTDMA). The resulting postextraction organic phases are chemically complex and often form rich hierarchical structures whose statics and dynamics near the critical point were probed herein with small-angle X-ray scattering and high-speed X-ray photon correlation spectroscopy. Despite their different extraction behaviors, all samples show remarkably similar critical behavior with exponents well described by classical critical point theory consistent with the 3D Ising model, where the critical behavior is characterized by fluctuations with a single diverging length scale. This unexpected result indicates a significant reduction in relevant chemical parameters at the critical point, indicating that the underlying behavior of phase transitions in LLE rely on far fewer variables than are generally assumed. The obtained scalar order parameter is attributed to the extractant fraction of the extractant/diluent mixture, revealing that other solution components and their respective concentrations simply shift the critical temperature but do not affect the nature of the critical fluctuations. These findings point to an opportunity to drastically simplify studies of liquid-liquid phase separation and phase diagram development in general while providing insights into LLE process improvement.

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

Interface synergism and engineering of Pd/Co@N-C for direct ethanol fuel cells

Direct ethanol fuel cells have been widely investigated as nontoxic and low-corrosive energy conversion devices with high energy and power densities. It is still challenging to develop high-activity and durable catalysts for a complete ethanol oxidation reaction on the anode and accelerated oxygen reduction reaction on the cathode. The materials’ physics and chemistry at the catalytic interface play a vital role in determining the overall performance of the catalysts. Herein, we propose a Pd/Co@N-C catalyst that can be used as a model system to study the synergism and engineering at the solid-solid interface. Particularly, the transformation of amorphous carbon to highly graphitic carbon promoted by cobalt nanoparticles helps achieve the spatial confinement effect, which prevents structural degradation of the catalysts. The strong catalyst-support and electronic effects at the interface between palladium and Co@N-C endow the electron-deficient state of palladium, which enhances the electron transfer and improved activity/durability. The Pd/Co@N-C delivers a maximum power density of 438 mW cm -2 in direct ethanol fuel cells and can be operated stably for more than 1000 hours. This work presents a strategy for the ingenious catalyst structural design that will promote the development of fuel cells and other sustainable energy-related technologies.

25 ENERGY STORAGE↗
DOE OSTI2015

Low-current field-assisted assembly of copper nanoparticles for current collectors

Current collectors are essential features of batteries and many other electronic devices being responsible for efficient charge transport to active electrode materials. Three-dimensional (3D), high surface area current collectors considerably improve the performance of cathodes and anodes in batteries, but their technological implementation is impeded by the complexity of their preparation, which needs to be simple, fast, and energy efficient. In this paper we demonstrate that field-stimulated assembly of ~3 nm copper nanoparticles (NPs) enables the preparation of porous Cu NP films. The use of NP dispersions enables 30× reduction of the deposition current for making functional 3D coatings. In addition to high surface area, lattice-to-lattice connectivity in the self-assembly of NPs in 3D structures enables fast charge transport. The mesoscale dimensions of out-of-plane features and the spacing between them in Cu films made by field-stimulated self-assembly of NPs provides promising morphology for current collection in lithium ion batteries (LIBs). Half-cell electrochemical models based on self-assembled films show improved specific capacity, total capacity, and cycling performance compared to traditional flat and other 3D current collectors. While integration of active electrode material into the 3D topography of the current collector needs to be improved, this study indicates that self-assembled NP films represent a viable manufacturing approach for 3D electrodes.

36 MATERIALS SCIENCE↗
DOE OSTI2021

Steering the methanol steam reforming reactivity of intermetallic Cu–In compounds by redox activation: stability vs . formation of an intermetallic compound–oxide interface

To compare the inherent methanol steam reforming properties of intermetallic compounds and a corresponding intermetallic compound–oxide interface, we selected the Cu–In system as a model to correlate the stability limits, self-activation and redox activation properties with the catalytic performance. Three distinct intermetallic Cu–In compounds – Cu 7 In 3 , Cu 2 In and Cu 11 In 9 – were studied both in an untreated and redox-activated state resulting from alternating oxidation–reduction cycles. The stability of all studied intermetallic compounds during methanol steam reforming (MSR) operation is essentially independent of the initial stoichiometry and all accordingly resist substantial structural changes. The inherent activity under batch MSR conditions is highest for Cu 2 In, corroborating the results of a Cu 2 In/In 2 O 3 sample accessed through reactive metal–support interaction. Under flow MSR operation, Cu 7 In 3 displays considerable deactivation, while Cu 2 In and Cu 11 In 9 feature stable performance at simultaneously high CO 2 selectivity. Here, the missing significant self-activation is most evident in the operando thermogravimetric experiments, where no oxidation is detected for any of the intermetallic compounds. In situ X-ray diffraction allowed us to monitor the partial decomposition and redox activation of the Cu–In intermetallic compounds into Cu 0.9 In 0.1 /In 2 O 3 (from Cu 7 In 3 ), Cu 7 In 3 /In 2 O 3 (from Cu 2 In) and Cu 7 In 3 /Cu 0.9 In 0.1 /In 2 O 3 (from Cu 11 In 9 ) interfaces with superior MSR performance compared to the untreated samples. Although the catalytic profiles appear surprisingly similar, the latter interface with the highest indium content exhibits the least deactivation, which we explain by formation of stabilizing In 2 O 3 patches under MSR conditions.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2022

A Fast Time-Stepping Strategy for Dynamical Systems Equipped with a Surrogate Model

Simulation of complex dynamical systems arising in many applications is computationally challenging due to their size and complexity. Model order reduction, machine learning, and other types of surrogate modeling techniques offer cheaper and simpler ways to describe the dynamics of these systems but are inexact and introduce additional approximation errors. In order to overcome the computational difficulties of the full complex models, on one hand, and the limitations of surrogate models, on the other, this work proposes a new accelerated time-stepping strategy that combines information from both. This approach is based on the multirate infinitesimal general-structure additive Runge--Kutta framework. The inexpensive surrogate model is integrated with a small time step to guide the solution trajectory, and the full model is treated with a large time step to occasionally correct for the surrogate model error and ensure convergence. Here, we provide a theoretical error analysis, and several numerical experiments, to show that this approach can be significantly more efficient than using only the full or only the surrogate model for the integration.

Surrogate models↗