Search NASA⌕ Search

SEARCH · Search NASA

Results for “gradient methods”

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 1,387 records · Page 77

The effects of applied current on one-dimensional interdiffusion between copper and nickel in spark plasma sintering

Spark plasma sintering (SPS) is a powder metallurgy technique that employs the use of fast sintering kinetics to produce final consolidated components in a matter of minutes. In order to use blended powders in SPS to obtain fully alloyed parts, diffusion during sintering must be understood. An investigation into the effects of current on the diffusion of copper and nickel was performed using SPS. Bulk specimens were used to generate diffusion couples in SPS in alternating orientations with respect to the direction of the current. Control samples were produced using a horizontal insertion vacuum furnace. Experiments were performed at temperatures between 850°C and 1000°C for 3 h. Concentration profiles were obtained by the use of both energy-dispersive spectroscopy and a Monte Carlo simulated correction curve. Diffusion coefficients and activation energies were calculated for samples produced by SPS and annealing without current. It was shown that, at temperatures near 0.9 T{sub m}, the application of current in SPS inhibits diffusion between copper and nickel due to the re-orientation of electrons caused by the loss of ferromagnetism in nickel. Activation energy for diffusion is, however, decreased due to the temperature gradients arising from the difference in resistivity between the two species.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

Mapping tree height in complex terrain of northern China using ultra-high-resolution images

Tree height is a key parameter for estimating forest biomass and carbon sequestration. In recent years, notable progress has been made in mapping tree height using satellite imagery. However, existing tree height products show low accuracy in mountainous and complex terrains, and few studies typically addressed tree height estimations in mountain areas. This study examines the Mentougou district of Beijing, China, characterized by complex terrain and mountainous landscapes. We analyzed two methods for estimating tree height: one using only spectral features and another combining spectral features with topographic factors (elevation, slope, aspect). We used 3-m resolution PlanetScope 8-band multispectral imagery, with 710 field-measured individual tree heights averaged to obtain 471 pixel-level tree height values as ground-truth, to develop tree height prediction models using eXtreme Gradient Boosting (XGBoost), Random Forest (RF), and Gradient Boosting Machine (GBM) models. The results show that the XGBoost model consistently presented the highest accuracy for both methods evaluated. Specifically, the XGBoost model that combined spectral data with elevation and slope variables with an R² of 0.75 and an RMSE of 2.69 m. Using the XGBoost model, we generated the tree height map for the Mentougou area at 3 m resolution, showing tree heights ranging from 0.5 to 30.4 m, and the model’s prediction error standard deviations ranged from 2.50 to 4.71 m, indicating reliable performance across varied terrain. Additionally, we compared and evaluated the global tree height products, identifying limitations in the accuracy within complex terrains. This study demonstrates the potential for accurately predicting tree heights by combining high-resolution multispectral satellites with a terrain factor modeling approach.

Complex terrain↗

SU-G-TeP4-03: A Multileaf Collimator Calibration and Quality Assurance Technique Using An Electronic Portal Imaging Device

Purpose: To develop an accurate and quick multileaf collimator (MLC) calibration and quality assurance technique using an electronic portal imaging device (EPID) Methods: The MLC models used include the MLCi and Agility (Elekta Ltd). This technique consists of two 22(L)x10(W) cm{sup 2} fields with 0{sup 0} and 180{sup 0} collimator angles centered to an offset EPID. The MLC opening is estimated by calculating the profile at the image’s center in the image’s horizontal direction. Scans in the image’s vertical direction were calculated every 20 pixels in the inner 70% of estimated MLC opening. The profiles’ edges were fitted with linear equations to determine the image’s rotation angle. Then, crossline profiles were scanned at the center of each leaf taking into account the leaf’s width at isocenter and the rotation angle. The profiles’ edges determine the location of the leaves’ edges and these were subtracted from the reference leaf’s position in order to determine the relative leaf offsets. The edge location of all profiles was determined by using the parameterized gradient of the penumbra region. The technique was tested against an established diode array-based method, and for different MLC systems, patterns, gantry angles, days, energies, beam modalities and MLC openings. Results: The differences between the proposed and established methods were 0.26±0.19mm. The leaf offsets’ deviation was <0.3mm (5 months period). For pattern fields, the differences between predetermined and calculated offsets were 0.18±0.18mm. The leaf offset deviation of measurements with different energies and MLC openings were <0.1mm and <0.3mm, respectively. The differences between offsets of FF and FFF beams were 0.01±0.02mm (<0.07mm). The differences between the offsets at different gantry angles were 0.08±0.15mm. Conclusion: The proposed method proved to be accurate and efficient in calculating the relative leaf offsets. Parameterized field edge is essential to obtain accurate result by eliminating the noise from EPID.

60 APPLIED LIFE SCIENCES↗

The Discrete Equation Method (DEM) for Fully Compressible, Two-Phase Flows in Ducts of Spatially Varying Cross-Section

For the simulation of light water nuclear reactor coolant flows, general two-phase models (valid for all volume fractions) have been generally used which, while allowing for velocity disequilibrium, normally force pressure equilibrium between the phases (see, for example, the numerous models of this type described in H. Städtke, Gasdynamic Aspects of Two-Phase Flow, Wiley-VCH, 2006). These equations are not hyperbolic, their physical wave dynamics are incorrect, and their solution algorithms rely on dubious truncation error induced artificial viscosity to render them numerically well posed over a portion of the computational spectrum. The inherent problems of the traditional approach to multiphase modeling, which begins with an averaged system of (ill-posed) partial differential equations (PDEs) which are then discretized to form a numerical scheme, are avoided by employing a new homogenization method known as the Discrete Equation Method (DEM) (R. Abgrall and R. Saurel, Discrete Equations for Physical and Numerical Compressible Multiphase Mixtures, J. Comp. Phys. 186, 361-396, 2003). This method results in well-posed hyperbolic systems, this property being important for transient flows. This also allows a clear treatment of non-conservative terms (terms involving interfacial variables and volume fraction gradients) permitting the solution of interface problems without conservation errors, this feature being important for the direct numerical simulation of two-phase flows. Unlike conventional methods, the averaged system of PDEs for the mixture are not used, and the DEM method directly obtains a well-posed discrete equation system from the single-phase conservation laws, producing a numerical scheme which accurately computes fluxes for arbitrary number of phases and solves non-conservative products. The method effectively uses a sequence of single phase Riemann problem solutions. Phase interactions are accounted for by Riemann solvers at each interface. Non-conservative terms are correctly approximated. Some of the closure relations missing from the traditional approach are automatically obtained. Lastly, the continuous equation system resulting from the discrete equations can be identified by taking the continuous limit with weak-wave assumptions. In this work, this approach is tested by constructing a DEM model for the flow of two compressible phases in 1-D ducts of spatially varying cross-section with explicit time integration. An analytical equation of state is included for both water vapor and liquid phases, and a realistic interphase mass transfer model is developed based on interphase heat transfer. A robust compliment of boundary conditions are developed and discussed. Though originally conceived as a first step toward implict time integration of the DEM method (to relieve time step size restrictions due to stiffness and to achieve tighter coupling of equations) in multidimensions, this model offers some unique capabilities for incorporation into next generation light water reactor safety analysis codes. We demonstrate, on a converging-diverging two-phase nozzle, that this well-posed, 2-pressure, 2-velocity DEM model can be integrated to a realistic and meaningful steady-state with both phases treated as compressible.

42 ENGINEERING↗

Gradient-informed Hamiltonian Monte Carlo for multicomponent CALPHAD model optimization and uncertainty quantification

CALPHAD model parameter optimization is inherently challenging due to non-smooth objective functions, high-dimensional parameter spaces, and the need for uncertainty quantification (UQ). Traditional weighted nonlinear least squares approaches are computationally efficient but local, whereas black-box global optimizers and ensemble Markov Chain Monte Carlo (MCMC) methods provide broader exploration at substantial computational cost. The objective of this work is to combine the global exploration capability of gradient-informed Hamiltonian Monte Carlo – specifically the No-U-Turn Sampler (NUTS) – with local deterministic refinement using BFGS to efficiently optimize multicomponent CALPHAD models with minimal manual intervention. Analytic gradients are computed via the Jansson derivative framework. The methodology is demonstrated on the Cr—Fe binary system and extended to the Cr—Fe—Ni ternary system with 32 degrees of freedom. For Cr—Fe, NUTS achieves comparable or superior optimality relative to ensemble MCMC while requiring over an order-of-magnitude fewer likelihood evaluations. Parameter uncertainties are quantified through NUTS sampling and propagated to thermodynamic observables using local expansion, demonstrating a novel modular approach that combines binary and ternary parameter subsets without requiring global relaxation. These results establish gradient-informed exploration as a scalable strategy for multicomponent CALPHAD optimization and provide a practical route towards efficient higher-order database development with quantified uncertainty.

36 MATERIALS SCIENCE↗

Source-encoded waveform inversion in the Northern Hemisphere

SUMMARY We use source-encoded waveform inversion to image Earth’s Northern Hemisphere. The encoding method is based on measurements of Laplace coefficients of stationary wavefields. By assigning to each event a unique frequency, we compute Fréchet derivatives for all events simultaneously based on one ‘super’ forward and one ‘super’ adjoint simulation for a small fraction of the computational cost of classical waveform inversion with the same data set. No cross-talk noise is introduced in the process, and the method does not require all events to be recorded by all stations. Starting from global model GLAD_M25, we performed 100 conjugate gradient iterations using a data set consisting of 786 earthquakes recorded by 9846 stations. Synthetic inversion tests show that we achieve good convergence based on this data set, and we see a consistent misfit reduction during the inversion. The new model, named SE100, has much higher spatial resolution than GLAD_M25, revealing details of the Yellowstone and Iceland hotspots, subduction beneath the Western United States and the upper mantle structure beneath the Arctic Ocean.

Cui, Congyue↗

Talbot-Lau based Moiré deflectometry with non-coherent sources as potential High Energy Density plasma diagnostic

X-ray phase-contrast radiography could better characterize highly localized density gradients expected in High Energy Density (HED) plasma experiments than conventional attenuation radiography. In particular, the Talbot-Lau (TL) grating interferometer, which works with extended and polychromatic x-ray sources, is a potentially attractive HED diagnostic due to its high sensitivity. For HED characterization the TL setup and imaging techniques must be changed from the recently studied medical system. The object magnification must be increased greatly in order to resolve μm scale gradients while the Talbot magnification must be increased in order to keep the gratings away from the plasma. Additionally, techniques for retrieving the density profile from a single plasma image must be developed. We thus study the performance of high magnification TL interferometers, in conjunction with Moiré fringe deflectometry for single image phase retrieval. The results show a very good interferometer contrast (≤30%) at high magnification. The Moiré technique enables measuring both sharp and mild density gradients with good accuracy and spatial resolution. Both the laboratory and simulation studies indicate that the TL based Moiré deflectometry is more sensitive than the propagation phase-contrast method when utilizing an extended x-ray source (∼80 μm). In HED experiments this would allow for less demanding X-ray backlighters than those used at present.

36 MATERIALS SCIENCE↗

The Role of Postintervention Pullback Pressure Gradient in Percutaneous Transluminal Angioplasty for Central Vein Stenosis in Dialysis Patients

Purpose: The severity of residual stenosis (RS) sometimes cannot be accurately measured by angiography during central vein intervention. This study evaluated the role of pullback pressure measurement during central vein stenosis (CVS) intervention. Methods: A retrospective review enrolled 94 consecutive dialysis patients who underwent CVS interventions but not stenting procedures. Patients were classified into 2 groups by either angiography or pressure gradient (PG) criteria, respectively. Groups divided by angiographic result were successful group (RS {<=}30 %) and acceptable group (50 % {>=} RS > 30 %), while groups divided by PG were low PG group (PG {<=}5 mmHg) and high PG group (PG >5 mmHg). Baseline characteristics and 12-month patency rates between the groups were analyzed. Results: The angiography results placed 63 patients in the successful group and 31 patients in the acceptable group. The patency rate at 12 month was not statistically different (P = 0.167). When the patients were reclassified by the postintervention pullback PG, the patency rate at 12 months was significant (P = 0.048). Further analysis in groups redivided by different combinations of RS and PG criteria identified significant differences in the group with both RS {<=}30 % and PG {<=}5 mmHg compared with those with either RS >30 % (P = 0.047) or PG >5 mmHg (P = 0.027). In addition, there was a significant difference between those with both RS {<=}30 % and PG {<=}5 mmHg compared with those with both RS >30 % and PG >5 mmHg (P = 0.027). Conclusion: Postintervention PG can better predict long-term outcomes after angioplasty for CVS in nonstented dialysis patients than angiography.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

Estimation of the Surface Fluxes for Heat and Momentum in Unstable Conditions with Machine Learning and Similarity Approaches for the LAFE Data Set

Abstract Measurements of three flux towers operated during the land atmosphere feedback experiment (LAFE) are used to investigate relationships between surface fluxes and variables of the land–atmosphere system. We study these relations by means of two machine learning (ML) techniques: multilayer perceptrons (MLP) and extreme gradient boosting (XGB). We compare their flux derivation performance with Monin–Obukhov similarity theory (MOST) and a similarity relationship using the bulk Richardson number (BRN). The ML approaches outperform MOST and BRN. Best agreement with the observations is achieved for the friction velocity. For the sensible heat flux and even more so for the latent heat flux, MOST and BRN deviate from the observations while MLP and XGB yield more accurate predictions. Using MOST and BRN for latent heat flux, the root mean square errors (RMSE) are 107 Wm $$^{-2}$$ - 2 and 121 Wm $$^{-2}$$ - 2 , respectively, as well as the intercepts of the regression lines are $$\approx 110$$ ≈ 110 Wm $$^{-2}$$ - 2 . For the ML methods, the RMSEs reduce to 31 Wm $$^{-2}$$ - 2 for MLP and 33 Wm $$^{-2}$$ - 2 for XGB as well as the intercepts to just 4 Wm $$^{-2}$$ - 2 for MLP and $$-1$$ - 1 Wm $$^{-2}$$ - 2 for XGB with slopes of the regression lines close to 1, respectively. These results indicate significant deficiencies of MOST and BRN, particularly for the derivation of the latent heat flux. In fact, in contrast to the established theories, feature importance weighting demonstrates that the ML methods base their improved derivations on net radiation, the incoming and outgoing shortwave radiations, the air temperature gradient, and the available water contents, but not on the water vapor gradient. The results imply that further studies of surface fluxes and other turbulent variables with ML techniques provide great promise for deriving advanced flux parameterizations and their implementation in land–atmosphere system models.

54 ENVIRONMENTAL SCIENCES↗

Intrinsic dissipation in a nano-mechanical resonator

We investigate the effect of size on intrinsic dissipation in nano-structures. We use molecular dynamics simulation and study dissipation under two different modes of deformation: stretching and bending mode. In the case of stretching deformation (with uniform strain field), dissipation takes place due to Akhiezer mechanism. For bending deformation, in addition to the Akhiezer mechanism, the spatial temperature gradient also plays a role in the process of entropy generation. Interestingly, we find that the bending modes have a higher Q factor in comparison with the stretching deformation (under the same frequency of operation). Furthermore, with the decrease in size, the difference in Q factor between the bending and stretching deformation becomes more pronounced. The lower dissipation for the case of bending deformation is explained to be due to the surface scattering of phonons. A simple model, for phonon dynamics under an oscillating strain field, is considered to explain the observed variation in dissipation rate. We also studied the scaling of Q factor with initial tension, in a beam under flexure. We develop a continuum theory to explain the observed results.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

CO 2 storage site characterization using ensemble-based approaches with deep generative models

Estimating spatially distributed properties such as permeability from available sparse measurements is a great challenge in efficient subsurface CO 2 storage operations. In this paper, a deep generative model that can accurately capture complex subsurface structure is tested with an ensemble-based inversion method for accurate and accelerated characterization of CO 2 storage sites. We chose Wasserstein Generative Adversarial Network with Gradient Penalty (WGAN-GP) for its realistic reservoir property representation and Ensemble Smoother with Multiple Data Assimilation (ES-MDA) for its robust data fitting and uncertainty quantification capability. WGAN-GP are trained to generate high-dimensional permeability fields from a low-dimensional latent space and ES-MDA then updates the latent variables by assimilating available measurements. Several subsurface site characterization examples including Gaussian, channelized, and fractured reservoirs are used to evaluate the accuracy and computational efficiency of the proposed method and the main features of the unknown permeability fields are characterized accurately with reliable uncertainty quantification. Furthermore, the estimation performance is compared with a widely-used variational, i.e., optimization-based, inversion approach, and the proposed approach outperforms the variational inversion method in several benchmark cases. We explain such superior performance by visualizing the objective function in the latent space: because of nonlinear and aggressive dimension reduction via generative modeling, the objective function surface becomes extremely complex while the ensemble approximation can smooth out the multi-modal surface during the minimization. This suggests that the ensemble-based approach works well over the variational approach when combined with deep generative models at the cost of forward model runs unless convergence-ensuring modifications are implemented in the variational inversion.

42 ENGINEERING↗

ELECTRON ACCELERATION BY CASCADING RECONNECTION IN THE SOLAR CORONA. I. MAGNETIC GRADIENT AND CURVATURE DRIFT EFFECTS

We investigate the electron acceleration by magnetic gradient and curvature drift effects in cascading magnetic reconnection of a coronal current sheet via a test particle method in the framework of the guiding center approximation. After several Alfvén transit times, most of the electrons injected at the current sheet are still trapped in the magnetic islands. A small fraction of the injected electrons precipitate into the chromosphere. The acceleration of trapped electrons is dominated by the magnetic curvature drifts, which change the parallel momentum of the electron, and appears to be more efficient than the acceleration of precipitating electrons, which is dominated by the perpendicular momentum change caused by the magnetic gradient drifts. With the resulting trapped energetic electron distribution, the corresponding hard X-ray (HXR) radiation spectra are calculated using an optically thin Bremsstrahlung model. Trapped electrons may explain flare loop top HXR emission as well as the observed bright spots along current sheets trailing coronal mass ejections. The asymmetry of precipitating electrons with respect to the polarity inversion line may contribute to the observed asymmetry of footpoint emission.

79 ASTRONOMY AND ASTROPHYSICS↗

Shape optimization using a NURBS-based interface-enriched generalized FEM

This study presents a gradient-based shape optimization over a fixed mesh using a non-uniform rational B-splines-based interface-enriched generalized finite element method, applicable to multi-material structures. In the proposed method, non-uniform rational B-splines are used to parameterize the design geometry precisely and compactly by a small number of design variables. An analytical shape sensitivity analysis is developed to compute derivatives of the objective and constraint functions with respect to the design variables. Subtle but important new terms involve the sensitivity of shape functions and their spatial derivatives. As a result, verification and illustrative problems are solved to demonstrate the precision and capability of the method.

42 ENGINEERING↗

Local bilinear computation of Jacobi sets

Abstract We propose a novel method for the computation of Jacobi sets in 2D domains. The Jacobi set is a topological descriptor based on Morse theory that captures gradient alignments among multiple scalar fields, which is useful for multi-field visualization. Previous Jacobi set computations use piecewise linear approximations on triangulations that result in discretization artifacts like zig-zag patterns. In this paper, we utilize a local bilinear method to obtain a more precise approximation of Jacobi sets by preserving the topology and improving the geometry. Consequently, zig-zag patterns on edges are avoided, resulting in a smoother Jacobi set representation. Our experiments show a better convergence with increasing resolution compared to the piecewise linear method. We utilize this advantage with an efficient local subdivision scheme. Finally, our approach is evaluated qualitatively and quantitatively in comparison with previous methods for different mesh resolutions and across a number of synthetic and real-world examples.

97 MATHEMATICS AND COMPUTING↗

Estimation and uncertainty quantification for the output from quantum simulators

The problem of estimating certain distributions over {0, 1} d is considered here. The distribution represents a quantum system of d qubits, where there are non-trivial dependencies between the qubits. A maximum entropy approach is adopted to reconstruct the distribution from exact moments or observed empirical moments. The Robbins Monro algorithm is used to solve the intractable maximum entropy problem, by constructing an unbiased estimator of the un-normalized target with a sequential Monte Carlo sampler at each iteration. In the case of empirical moments, this coincides with a maximum likelihood estimator. A Bayesian formulation is also considered in order to quantify uncertainty a posteriori. Several approaches are proposed in order to tackle this challenging problem, according to recently developed methodologies. In particular, unbiased estimators of the gradient of the log posterior are constructed and used within a provably convergent Langevin-based Markov chain Monte Carlo method. The methods are illustrated on classically simulated output from quantum simulators.

97 MATHEMATICS AND COMPUTING↗

A regularized-interface method as a unified formulation for simulations of high-pressure multiphase flows

The injection of multi-species fluids into high-pressure and high-temperature environments beyond the species' critical points is commonly found in engineering applications. At these conditions, for immiscible species, both subcritical interfacial dynamics and supercritical mixing can coexist due to variations in temperature around the mixture critical point. The modeling of these complex transcritical phenomena for large-scale configurations is so far not possible. To address this issue, we propose the Regularized-Interface Method (RIM) as a unified formulation that can describe both sub- and supercritical processes as well as the transition between them. The proposed method is derived via filtering of the nanoscale interface-resolving formulation based on van der Waals' linear gradient theory. Thus, this approach allows for the consistent modeling of interfacial dynamics that vanishes at supercritical conditions, while significantly reducing the temporal and spatial resolution constraints of the original nanoscale formulation. The resulting RIM formulation is examined in interface-capturing simulations of sub-, trans-, and supercritical fuel injection processes, involving droplets and jets. Furthermore, these results highlight the importance of resolving spatio-temporal transitions from subcritical interfacial dynamics to supercritical mixing in high-pressure multiphase simulations, in contrast to commonly employed diffused-interface methods, where interfacial dynamics are often neglected.

Interface capturing↗

SU-F-T-539: Dosimetric Comparison of Volumetric Modulated Arc Therapy and Intensity Modulated Radiation Therapy for Whole Brain Hippocampal Sparing Radiation Therapy Treatments

Purpose: To compare the treatment plan quality and dose gradient near the hippocampus between VMAT (RapidArc) and IMRT delivery techniques for whole brain radiation therapy. Methods: Fifteen patients were evaluated in this retrospective study. All treatments were planned on Varian Eclipse TPS, using 3-Arc VMAT and 9-Field IMRT, following NRG Oncology protocol NRG-CC001 guidelines evaluated by a single radiation oncologist. Prescribed doses in all plans were 30 Gy delivered over 10 fractions normalized to a minimum of 100% of the dose covering 95% of the target volume. Identical contour sets and dose-volume constraints following protocol guidelines were also applied in all plans. A paired t-test analysis was used to compare VMAT and IMRT plans. Results: NRG-CC001 protocol dose-volume constraints were met for all VMAT and IMRT plans. For the planning target volume (PTV), the average values for D2% and D98% were 6% lower and 4% higher in VMAT than in IMRT, respectively. The average mean and maximum hippocampus doses in Gy for VMAT vs IMRT plans were (11.85±0.81 vs. 12.24±0.56, p=0.10) and (16.27±0.78 vs. 16.59±0.71, p=0.24), respectively. In VMAT, the average mean and maximum chiasm doses were 3% and 1% higher than in IMRT plans, respectively. For the left optic nerve, the average mean and maximum doses were 10% and 5% higher in VMAT than in IMRT plans, respectively. These values were 12% and 3% for the right optic nerve. The average percentage of dose gradient around the hippocampus in the 0–5mm and 5–10mm abutted regions for VMAT vs. IMRT were (4.42%±2.22% /mm vs. 3.95%±2.61% /mm, p=0.43) and (4.54%±1.50% /mm vs. 4.39%±1.28% /mm, p=0.73), respectively. Conclusion: VMAT plans can achieve higher hippocampus sparing with a faster dose fall-off than IMRT plans. Though statistically insignificant, VMAT offers better PTV coverage with slightly higher doses to OARs.

60 APPLIED LIFE SCIENCES↗

A finite elements method to solve the Bloch–Torrey equation applied to diffusion magnetic resonance imaging

The complex transverse water proton magnetization subject to diffusion-encoding magnetic field gradient pulses in a heterogeneous medium can be modeled by the multiple compartment Bloch–Torrey partial differential equation (PDE). In addition, steady-state Laplace PDEs can be formulated to produce the homogenized diffusion tensor that describes the diffusion characteristics of the medium in the long time limit. In spatial domains that model biological tissues at the cellular level, these two types of PDEs have to be completed with permeability conditions on the cellular interfaces. To solve these PDEs, we implemented a finite elements method that allows jumps in the solution at the cell interfaces by using double nodes. Using a transformation of the Bloch–Torrey PDE we reduced oscillations in the searched-for solution and simplified the implementation of the boundary conditions. The spatial discretization was then coupled to the adaptive explicit Runge–Kutta–Chebyshev time-stepping method. Our proposed method is second order accurate in space and second order accurate in time. We implemented this method on the FEniCS C++ platform and show time and spatial convergence results. Finally, this method is applied to study some relevant questions in diffusion MRI.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗