Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “error localization”

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 919 records · Page 51

Generic method for automatic bladder segmentation on cone beam CT using a patient-specific bladder shape model

Purpose: The aim of this study is to develop and validate a generic method for automatic bladder segmentation on cone beam computed tomography (CBCT), independent of gender and treatment position (prone or supine), using only pretreatment imaging data. Methods: Data of 20 patients, treated for tumors in the pelvic region with the entire bladder visible on CT and CBCT, were divided into four equally sized groups based on gender and treatment position. The full and empty bladder contour, that can be acquired with pretreatment CT imaging, were used to generate a patient-specific bladder shape model. This model was used to guide the segmentation process on CBCT. To obtain the bladder segmentation, the reference bladder contour was deformed iteratively by maximizing the cross-correlation between directional grey value gradients over the reference and CBCT bladder edge. To overcome incorrect segmentations caused by CBCT image artifacts, automatic adaptations were implemented. Moreover, locally incorrect segmentations could be adapted manually. After each adapted segmentation, the bladder shape model was expanded and new shape patterns were calculated for following segmentations. All available CBCTs were used to validate the segmentation algorithm. The bladder segmentations were validated by comparison with the manual delineations and the segmentation performance was quantified using the Dice similarity coefficient (DSC), surface distance error (SDE) and SD of contour-to-contour distances. Also, bladder volumes obtained by manual delineations and segmentations were compared using a Bland-Altman error analysis. Results: The mean DSC, mean SDE, and mean SD of contour-to-contour distances between segmentations and manual delineations were 0.87, 0.27 cm and 0.22 cm (female, prone), 0.85, 0.28 cm and 0.22 cm (female, supine), 0.89, 0.21 cm and 0.17 cm (male, supine) and 0.88, 0.23 cm and 0.17 cm (male, prone), respectively. Manual local adaptations improved the segmentation results significantly (p < 0.01) based on DSC (6.72%) and SD of contour-to-contour distances (0.08 cm) and decreased the 95% confidence intervals of the bladder volume differences. Moreover, expanding the shape model improved the segmentation results significantly (p < 0.01) based on DSC and SD of contour-to-contour distances. Conclusions: This patient-specific shape model based automatic bladder segmentation method on CBCT is accurate and generic. Our segmentation method only needs two pretreatment imaging data sets as prior knowledge, is independent of patient gender and patient treatment position and has the possibility to manually adapt the segmentation locally.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

Note: Rotaphone, a new self-calibrated six-degree-of-freedom seismic sensor

We have developed and tested (calibration, linearity, and cross-axis errors) a new six-degree-of-freedom mechanical seismic sensor for collocated measurements of three translational and three rotational ground motion velocity components. The device consists of standard geophones arranged in parallel pairs to detect spatial gradients. The instrument operates in a high-frequency range (above the natural frequency of the geophones, 4.5 Hz). Its theoretical sensitivity limit in this range is 10{sup -9} m/s in ground velocity and 10{sup -9} rad/s in rotation rate. Small size and weight, and easy installation and maintenance make the instrument useful for local-earthquake recording and seismic prospecting.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

TU-F-BRE-01: A High Resolution Micro Fiber Scintillator Detector Optimized for SRS and SBRT in Vivo Real Time Treatment Verification

Purpose: We have built a high resolution real time scintillating fiber detector prototype to determine in real time the accuracy of stereotactic radiosurgery (SRS) and stereotactic body radiotherapy (SBRT) treatments when only a fraction of the planned dose was delivered. The motivation of this work is to enhance dose delivery accuracy and to achieve error free radiosurgery. Methods: A high density array of scintillating fibers and a high speed photo detectors array were integrated to implement a high resolution real time dosimeter that can sample with high resolution pulsed SRS and SBRT beams cross sections. The high efficiency of the developed system allows to read each linac pulse in real time and to compute the accumulated dose and dose errors when only a fraction of the beam was delivered. The fibers are highly packed in a substrate that is directly coupled to two 128 pixel arrays with a pitch matching the fiber spacing to achieve accurate spatial localization. The small cross section of the fiber array allows stacking multiple fiber arrays to measure independent angular profiles that are digitally processed in parallel for real time dosimetry. Results: We implemented a high density array detector prototype with a pitch of 0.5 mm, readout speed of 1.2 msec, and a response time of 0.5 usec. The fast reading speed has the capability to determining the dose in flattening free filter beams. The detector can be installed in transmission mode at the output port of a micro-MLC. Treatment deviations smaller than 3% are detected when less than 1/100 of the planned dose was delivered. Conclusions: We built a prototype of a high resolution fiber scintillator array detector for SRS and SBRT in vivo dosimetry. Results show that the developed detector has the potential to assure error free SRS and SBRT treatments.

60 APPLIED LIFE SCIENCES↗

Determining the hubble constant from hubble parameter measurements

We use 28 measurements of the Hubble parameter, H(z), at intermediate redshifts 0.07⩽z⩽2.3 to determine the present-day Hubble constant H{sub 0} in four cosmological models. We measure H{sub 0}=68.3{sub −2.6}{sup +2.7},68.4{sub −3.3}{sup +2.9},65.0{sub −6.6}{sup +6.5}, and 67.9{sub −2.4}{sup +2.4} km s{sup −1} Mpc{sup −1} (1σ errors) in the ΛCDM (spatially flat and non-flat), ωCDM, and ϕCDM models, respectively. These measured H{sub 0} values are more consistent with the lower values determined from recent data on the cosmic microwave background and baryon acoustic oscillations, as well as with the value found from a median statistical analysis of Huchra’s compilation of H{sub 0} measurements, but include the higher local measurements of H{sub 0} within the 2σ confidence limits.

79 ASTRONOMY AND ASTROPHYSICS↗

Demonstration of a viable quantitative theory for interplanetary type II radio bursts

Between 29 November and 1 December 2013 the two widely separated spacecraft STEREO A and B observed a long lasting, intermittent, type II radio burst for the extended frequency range ≈ 4 MHz to 30 kHz, including an intensification when the shock wave of the associated coronal mass ejection (CME) reached STEREO A. We demonstrate for the first time our ability to quantitatively and accurately simulate the fundamental (F) and harmonic (H) emission of type II bursts from the higher corona (near 11 solar radii) to 1 AU. Our modeling requires the combination of data-driven three-dimensional magnetohydrodynamic simulations for the CME and plasma background, carried out with the BATS-R-US code, with an analytic quantitative kinetic model for both F and H radio emission, including the electron reflection at the shock, growth of Langmuir waves and radio waves, and the radiations propagation to an arbitrary observer. The intensities and frequencies of the observed radio emissions vary hugely by factors ≈ 10{sup 6} and ≈ 10{sup 3}, respectively; the theoretical predictions are impressively accurate, being typically in error by less than a factor of 10 and 20 %, for both STEREO A and B. We also obtain accurate predictions for the timing and characteristics of the shock and local radio onsets at STEREO A, the lack of such onsets at STEREO B, and the z-component of the magnetic field at STEREO A ahead of the shock, and in the sheath. Very strong support is provided by these multiple agreements for the theory, the efficacy of the BATS-R-US code, and the vision of using type IIs and associated data-theory iterations to predict whether a CME will impact Earth’s magnetosphere and drive space weather events.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Nonlinear neoclassical two-fluid theory of response of tokamak plasma to resonant error-field

A nonlinear, neoclassical, two-fluid theory of the interaction of a single-helicity magnetic island chain with a resonant error-field in a quasi-cylindrical, low-β, tokamak plasma is presented. In particular, the analysis of Fitzpatrick [Phys. Plasmas 25, 042503 (2018)] is generalized to take explicit time dependence into account. Aside from the ability to more accurately treat time-varying problems, the main physical effect that is introduced into the theory by the incorporation of explicit time dependence is ion inertia. The formalism developed in the paper is used to analyze two time-varying problems. First, the interaction of a pre-existing magnetic island chain with a resonant error-field. Second, an error-field-maintained magnetic island chain. The latter problem is of direct relevance to experiments in which deliberately applied, multi-harmonic, resonant magnetic perturbations are used to suppress edge localized modes (ELMs) in tokamak plasmas. Indeed, the predictions of the theory are strikingly similar to data recently obtained from ELM suppression experiments in the DIII-D tokamak [R. Nazikian et al., Nucl. Fusion 58, 106010 (2018)].

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

The effect of self-interaction error on electrostatic dipoles calculated using density functional theory

Spurious electron self-interaction in density functional approximations (DFAs) can lead to inaccurate predictions of charge transfer in heteronuclear molecules that manifest as errors in calculated electrostatic dipoles. Here we show the magnitude of these errors on dipoles computed for a diverse set of 47 molecules taken from the recent benchmark study of Hait and Head-Gordon. We compare results of Perdew-Wang local spin density approximation (PW92), Perdew-Burke-Ernzerhof (PBE) generalized gradient approximation (GGA), and strongly constrained and appropriatedly normed (SCAN) meta-GGA dipole calculations, along with those of their respective self-interaction-corrected (SIC) counterparts, to reference values from accurate wave function-based methods. The SIC calculations were carried out using the Fermi-Löwdin orbital (FLO-SIC) approach. We find that correcting for self-interaction generally increases the degree of charge transfer, thereby increasing the size of calculated dipole moments. The FLO-SIC-PW92 and FLO-SIC-PBE dipoles are in better agreement with reference values than their uncorrected DFA counterparts, particularly for strongly ionic molecules where significant improvement is seen. Applying FLO-SIC to SCAN does not improve dipole values overall. In addition, we show that removing self-interaction improves the description of the dipole for stretched-bond geometries and recovers the physically correct separated atom limit of zero dipole. Finally, we find that the best agreement between the FLO-SIC-DFA and reference dipoles occurs when the molecular geometries are optimized using the FLO-SIC-DFA.

74 ATOMIC AND MOLECULAR PHYSICS↗

STUDIES IN ASTRONOMICAL TIME SERIES ANALYSIS. VI. BAYESIAN BLOCK REPRESENTATIONS

This paper addresses the problem of detecting and characterizing local variability in time series and other forms of sequential data. The goal is to identify and characterize statistically significant variations, at the same time suppressing the inevitable corrupting observational errors. We present a simple nonparametric modeling technique and an algorithm implementing it-an improved and generalized version of Bayesian Blocks-that finds the optimal segmentation of the data in the observation interval. The structure of the algorithm allows it to be used in either a real-time trigger mode, or a retrospective mode. Maximum likelihood or marginal posterior functions to measure model fitness are presented for events, binned counts, and measurements at arbitrary times with known error distributions. Problems addressed include those connected with data gaps, variable exposure, extension to piecewise linear and piecewise exponential representations, multivariate time series data, analysis of variance, data on the circle, other data modes, and dispersed data. Simulations provide evidence that the detection efficiency for weak signals is close to a theoretical asymptotic limit derived by Arias-Castro et al. In the spirit of Reproducible Research all of the code and data necessary to reproduce all of the figures in this paper are included as supplementary material.

79 ASTRONOMY AND ASTROPHYSICS↗

Boundary dual of bulk local operators

We provide a procedure to determine if a given nonlocal operator in a large N holographic CFT is dual to a local bulk operator on the geometry associated with a particular code subspace of the CFT. This procedure does not presuppose knowledge of the bulk geometry. We are able to pick out local operators in a large region of the bulk, called the "localizable region,"' that can extend beyond event horizons in certain cases. The method relies heavily on the quantum-error correcting structure of AdS/CFT and, in particular, on entanglement wedge reconstruction. As a byproduct of this machinery, we are able to reconstruct the metric in the localizable region up to a conformal factor. This suggests a connection between our program and the recent light-cone cut approach to bulk reconstruction.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Predicting September Arctic Sea Ice: A Multimodel Seasonal Skill Comparison

This study quantifies the state of the art in the rapidly growing field of seasonal Arctic sea ice prediction. A novel multimodel dataset of retrospective seasonal predictions of September Arctic sea ice is created and analyzed, consisting of community contributions from 17 statistical models and 17 dynamical models. Prediction skill is compared over the period 2001–20 for predictions of pan-Arctic sea ice extent (SIE), regional SIE, and local sea ice concentration (SIC) initialized on 1 June, 1 July, 1 August, and 1 September. This diverse set of statistical and dynamical models can individually predict linearly detrended pan-Arctic SIE anomalies with skill, and a multimodel median prediction has correlation coefficients of 0.79, 0.86, 0.92, and 0.99 at these respective initialization times. Regional SIE predictions have similar skill to pan-Arctic predictions in the Alaskan and Siberian regions, whereas regional skill is lower in the Canadian, Atlantic, and central Arctic sectors. The skill of dynamical and statistical models is generally comparable for pan-Arctic SIE, whereas dynamical models outperform their statistical counterparts for regional and local predictions. The prediction systems are found to provide the most value added relative to basic reference forecasts in the extreme SIE years of 1996, 2007, and 2012. SIE prediction errors do not show clear trends over time, suggesting that there has been minimal change in inherent sea ice predictability over the satellite era. Overall, this study demonstrates that there are bright prospects for skillful operational predictions of September sea ice at least 3 months in advance.

54 ENVIRONMENTAL SCIENCES↗

Electronic energy gap closure and metal-insulator transition in dense liquid hydrogen

Here, using quantum Monte Carlo (QMC) calculations, we investigate the insulator-metal transition observed in liquid hydrogen at high pressure. Below the critical temperature of the transition from the molecular to the atomic liquid, the fundamental electronic gap closure occurs abruptly, with a small discontinuity reflecting the weak first-order transition in the thermodynamic equation of state. Above the critical temperature, molecular dissociation sets in while the gap is still open. When the gap closes, the decay of the off-diagonal reduced density matrix shows that the liquid enters a gapless, but localized, phase: there is a crossover between the insulating and the metallic liquids. Compared to different density functional theory (DFT) functionals, our QMC calculations provide larger values for the fundamental gap and the electronic density of states close to the band edges, indicating that optical properties from DFT potentially benefit from error cancellations.

36 MATERIALS SCIENCE↗

Leveraging the digital thread for physics-based prediction of microstructure heterogeneity in additively manufactured parts

A major limitation of additive manufacturing (AM) processes is that local conditions of material deposition frequently lead to unintentional heterogeneities in microstructure and properties within a single component, despite nominally uniform process conditions. Up to now, there has been no way to a priori determine the distribution of these heterogeneities, requiring expensive trial-and-error approaches to fabrication, testing, and characterization. Here, a physics-based framework for creating a digital representation of the laser powder bed fusion (PBF) process is proposed to predict the variation in solidification behavior that leads to heterogeneous microstructures in an as-built part. By leveraging in situ process data stored in the part’s digital thread, the scan path and process parameters were input into a heat transfer model which predicted solidification data at the melt pool scale. A two-step unsupervised clustering algorithm was used to first cluster the local solidification conditions (12.5µm 3 voxels) and then to cluster the regional behavior on the scale of multiple scan passes and print layers (250µm 3 super-voxels). This process was used to identify regions with similar solidification characteristics for multiple locations in a Stainless Steel 316-L component. The corresponding as-built part was sectioned and characterized using electron backscatter diffraction (EBSD). Quantitative analysis of the pole figures confirmed that the predicted regions of heterogeneity in the solidification conditions corresponded with differences in the observed microstructure. In conclusion, this work shows a viable path for estimating the microstructural heterogeneity for additively manufactured parts to either limit microstructural variation throughout a part or to enable functionality-based variation of the microstructure.

36 MATERIALS SCIENCE↗

Quantum Information for Fusion Energy Sciences (Final Technical Report)

The simulation of plasma dynamics is a critical area of Fusion Energy Sciences (FES) due to it’s usefulness in predicting, controlling, and confining plasmas in the context of potential fusion reactors. The simulation of plasmas is a computationally difficult problem in both classical and quantum physics, motivating investigation into the potential of quantum computers to simulate these systems. This project took several concrete steps towards this goal by developing tools for improving the control, characterization, and calibration of quantum gates on a superconducting quantum computer, developing error suppression and mitigation tools to reduce errors on the quantum computer, and utilizing these advancements to simulate reduced models of plasma dynamics on the quantum computer. In order to efficiently simulate plasma physics, an optimal control method which synthesizes, directly at the pulse level, any quantum gate on qubit and qutrit systems was developed. Using four superconducting transmon quantum processors at Rigetti and LLNL, it was demonstrated that any arbitrary quantum gate on qubits and qutrits could be implemented with high fidelity, leading to a significantly reduced length of a gate sequence. A problem of interest in FES is the nonlinear optical process of laser pulse compression within a plasma. Since quantum physics is linear, simulating nonlinear operations is not naturally feasible on a quantum computer, however it is possible to simulated a quantized version of the nonlinear process. A quantization approach to convert nonlinear wave-wave interaction problems to Hamiltonian simulation problems was developed and demonstrated using two qubits on a Rigetti device. In this experiment, a number of error suppression and mitigation techniques were investigated to determine how best to utilize the finite quantum resources. This study provides an example of how plasma problems may be solved on near-term, noisy quantum computing platforms and identified a promising set of techniques. Building on the insights of these experiments, the investigation turned to linear electron-plasma wave physics. A connection was identified between a local one-dimensional lattice spin model and linear wave phenomena, allowing a plasma physics problem to be efficiently mapped to the quantum computer. In this framework, reflection and transmission of plasma waves at a sharp boundary was studied, as well as the propagation of waves through an inhomogeneous plasma medium. In addition to the suite of error suppression and mitigation techniques developed, this experiment introduced the use of a digital-analog gate scheme designed to efficiently simulate the plasma Hamiltonian. With hardware available at the conclusion of the project, simulation at the scale of 9 qubits and 15 timesteps (60 entangling layers) was achieved.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

SU-E-J-13: Six Degree of Freedom Image Fusion Accuracy for Cranial Target Localization On the Varian Edge Stereotactic Radiosurgery System: Comparison Between 2D/3D and KV CBCT Image Registration

Purpose: To determine the 6 degree of freedom systematic deviations between 2D/3D and CBCT image registration with various imaging setups and fusion algorithms on the Varian Edge Linac. Methods: An anthropomorphic head phantom with radio opaque targets embedded was scanned with CT slice thicknesses of 0.8, 1, 2, and 3mm. The 6 DOF systematic errors were assessed by comparing 2D/3D (kV/MV with CT) with 3D/3D (CBCT with CT) image registrations with different offset positions, similarity measures, image filters, and CBCT slice thicknesses (1 and 2 mm). The 2D/3D registration accuracy of 51 fractions for 26 cranial SRS patients was also evaluated by analyzing 2D/3D pre-treatment verification taken after 3D/3D image registrations. Results: The systematic deviations of 2D/3D image registration using kV- kV, MV-kV and MV-MV image pairs were within ±0.3mm and ±0.3° for translations and rotations with 95% confidence interval (CI) for a reference CT with 0.8 mm slice thickness. No significant difference (P>0.05) on target localization was observed between 0.8mm, 1mm, and 2mm CT slice thicknesses with CBCT slice thicknesses of 1mm and 2mm. With 3mm CT slice thickness, both 2D/3D and 3D/3D registrations performed less accurately in longitudinal direction than thinner CT slice thickness (0.60±0.12mm and 0.63±0.07mm off, respectively). Using content filter and using similarity measure of pattern intensity instead of mutual information, improved the 2D/3D registration accuracy significantly (P=0.02 and P=0.01, respectively). For the patient study, means and standard deviations of residual errors were 0.09±0.32mm, −0.22±0.51mm and −0.07±0.32mm in VRT, LNG and LAT directions, respectively, and 0.12°±0.46°, −0.12°±0.39° and 0.06°±0.28° in RTN, PITCH, and ROLL directions, respectively. 95% CI of translational and rotational deviations were comparable to those in phantom study. Conclusion: 2D/3D image registration provided on the Varian Edge radiosurgery, 6 DOF-based system provides accurate target positioning for frameless image-guided cranial stereotactic radiosurgery.

60 APPLIED LIFE SCIENCES↗

On the surface energy balance closure at different temporal scales

Measurements of the surface energy fluxes (turbulent and radiative) and other ancillary atmospheric/soil parameters made in the Columbia River Basin (Oregon) in an area of complex terrain during a 10-month long portion of the second Wind Forecast Improvement Project (WFIP 2) field campaign are used to study the surface energy budget (SEB) and surface fluxes over different temporal scales. This study analyzes and discusses SEB closure based on half-hourly, daily, monthly, seasonal, and sub-annual (~10-month) temporal averages. The data were collected over all four seasons for different states of the underlying ground surface (dry, wet, and frozen). Our half-hourly direct measurements of energy balance show that the sum of the turbulent sensible and latent heat fluxes systematically underestimate positive net radiation by around 20–30% during daytime and overestimate negative net radiation at night. This imbalance of the surface energy budget is comparable to other terrestrial sites. However, on average, the residual energy imbalance is significantly reduced at daily, weekly, and monthly averaging timescales, and moreover, the SEB can be closed for this site within reasonable limits on seasonal and sub-annual timescales (311-day averaging for the entire field campaign dataset). Increasing the averaging time to daily and longer time intervals substantially reduces the ground heat flux and storage terms, because energy locally entering the soil, air column, and vegetation in the morning is released in the afternoon and evening. Averaging on daily to sub-annual timescales also reduces random instrumental measurement errors and other uncertainties as well as smooths out a hysteresis effect (phase lag) in the SEB relationship between different components. This study shows that SEB closure is better for dry soils compared to wet soils and the statistical dependence of the turbulent fluxes and net radiation for freezing soil surfaces appears weak, if not non-existent, apparently due to lack of the latent heat of fusion term in the traditional SEB equation.

54 ENVIRONMENTAL SCIENCES↗

Enforcing the Courant–Friedrichs–Lewy condition in explicitly conservative local time stepping schemes

In this study, an optimally efficient explicit numerical scheme for solving fluid dynamics equations, or any other parabolic or hyperbolic system of partial differential equations, should allow local regions to advance in time with their own, locally constrained time steps. However, such a scheme can result in violation of the Courant-Friedrichs-Lewy (CFL) condition, which is manifestly non-local. Although the violations can be considered to be "weak" in a certain sense and the corresponding numerical solution may be stable, such calculation does not guarantee the correct propagation speed for arbitrary waves . We use an experimental fluid dynamics code that allows cubic "patches" of grid cells to step with independent, locally constrained time steps to demonstrate how the CFL condition can be enforced by imposing a condition on the time steps of neighboring patches. We perform several numerical tests that illustrate errors introduced in the numerical solutions by weak CFL condition violations and show how strict enforcement of the CFL condition eliminates these errors. In all our tests the strict enforcement of the CFL condition does not impose a significant performance penalty.

79 ASTRONOMY AND ASTROPHYSICS↗

Dyn$\mathrm{AMO}$: Multi-agent reinforcement learning for dynamic anticipatory mesh optimization with applications to hyperbolic conservation laws

Here we introduce DynAMO, a reinforcement learning paradigm for Dynamic Anticipatory Mesh Optimization. Adaptive mesh refinement is an effective tool for optimizing computational cost and solution accuracy in numerical methods for partial differential equations. However, traditional adaptive mesh refinement approaches for time-dependent problems typically rely only on instantaneous error indicators to guide adaptivity. As a result, standard strategies often require frequent remeshing to maintain accuracy. In the DynAMO approach, multi-agent reinforcement learning is used to discover new local refinement policies that can anticipate and respond to future solution states by producing meshes that deliver more accurate solutions for longer time intervals. By applying DynAMO to discontinuous Galerkin methods for the linear advection and compressible Euler equations in two dimensions, we demonstrate that this new mesh refinement paradigm can outperform conventional threshold-based strategies while also generalizing to different mesh sizes, remeshing and simulation times, and initial conditions.

97 MATHEMATICS AND COMPUTING↗

Geospatial modeling of near subsurface temperatures of the contiguous United States for assessment of materials degradation

Abstract Understanding subsurface temperature variations is crucial for assessing material degradation in underground structures. This study maps subsurface temperatures across the contiguous United States for depths from 50 to 3500 m, comparing linear interpolation, gradient boosting (LightGBM), neural networks, and a novel hybrid approach combining linear interpolation with LightGBM. Results reveal heterogeneous temperature patterns both horizontally and vertically. The hybrid model performed best achieving a root mean square error of 2.61 °C at shallow depths (50–350 m). Model performance generally decreased with depth, highlighting challenges in deep temperature prediction. State-level analyses emphasized the importance of considering local geological factors. This study provides valuable insights for designing efficient underground facilities and infrastructure, underscoring the need for depth-specific and region-specific modeling approaches in subsurface temperature assessment.

Science & Technology - Other Topics↗