Search NASA⌕ Search

SEARCH · Search NASA

Results for “algorithmic differentiation”

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 973 records · Page 54

Bison Robustness and Performance Improvements

In last year’s Bison robustness report, the authors demonstrated that improving the Jacobian matrix can drastically increase solution convergence rates. The development of an automatic differentiation (AD) capability in the MOOSE framework has created the possibility of having exact Jacobian matrices for all of our material models. A large focus in this year’s report is the application of AD to MOOSE physics modules, such as the tensor mechanics module, and the application to Bison fuel and cladding material models. A second main focus in this year’s report is the investigation of a set of algorithmic solver changes to improve the robustness and performance of Bison LWR simulations. Beyond providing increased solver robustness, the ability to solve strongly coupled multiphysics problems, and solving large deformation problems with ease, these automatically generated perfect Jacobians also drastically reduce development time for new models. In models converted to AD, Jacobian computation made up the majority of the code base. With this code removed, both readability and maintainability of the MOOSE/Bison code base are improved. The work documented here significantly improved the robustness of the Bison LWR assessment case suite. Through algorithmic changes in solver infrastructure, such as the application of predictors and improvements in timestepping, tangible improvements in performance were observed. These improvements are documented using a new standard set of metrics that are collected in all assessment cases, and a suite of statistical analysis tools developed for this report.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗

SU-E-J-27: Appropriateness Criteria for Deformable Image Registration and Dose Propagation

Purpose: Several commercial software packages have been recently released that allow the user to apply deformable registration algorithms (DRA) for image fusion and dose propagation. Although the idea of anatomically tracking the daily patient dose in the context of adaptive radiotherapy or merely adding the dose from prior treatment to the current one is very intuitive, the accuracy and applicability of such algorithms needs to be investigated as it remains somewhat subjective. In our study, we used true anatomical data where we introduced changes in the density, volume and location of segmented structures to test the DRA for its sensitivity and accuracy. Methods: The CT scan of a prostate patient was selected for this study. The CT images were first segmented to define structure such as the PTV, bladder, rectum, intestines and pelvic bone anatomy. To perform our study, we introduced anatomical changes in the reference patient image set in three different ways: (i) we kept the segmented volumes constant and changed the density of rectum and bladder in increments of 5% (ii) we changed the volume of rectum and bladder in increments of 5% and (iii) we kept the segmented volumes constant but changed their location by moving their COM in increments of 3mm. Using the Velocity software, we evaluated the accuracy of the DRA for each incremental change in all three scenarios. Results: The DRA performs reasonably well when the differential density difference against the background is more than 5%. For the volume change study, the DRA results became unreliable for relative volume changes greater than 10%. Finally for the location study, the DRA performance was acceptable for shifts below 9mm. Conclusion: Site specific and patient specific QA for DRA is an important step to evaluate such algorithms prior to their use for dose propagation.

60 APPLIED LIFE SCIENCES↗

A novel implicit hybrid machine learning model and its application for reinforcement learning

A novel methodology to develop implicit hybrid models is presented. PyTorch is used to integrate physics-based equations with machine learning models. Automatic differentiation of the hybrid model is leveraged to solve the implicit equations. Iterative solving enables gradient based updates to the machine learning model. The novel methodology is compared to an explicit hybrid approach on a continuously stirred tank reactor (CSTR). The novel method results in a lower modelling error. Both hybrid models effectively train with noisy data. To test the implicit hybrid model, it is employed as a reinforcement learning (RL) training model. The RL algorithm trained on the hybrid model outperforms real time optimization of the CSTR and performs nearly as well as RL trained directly on the CSTR and a traditional gradient based approach. Training RL directly on the CSTR requires over 60,000 system interactions compared to 6000 historical data points for hybrid model development.

42 ENGINEERING↗

Bounded motion design in the Earth zonal problem using differential algebra based normal form methods

Establishing long-term relative bounded motion between orbits in perturbed dynamics is a key challenge in astrodynamics to enable cluster flight with minimum propellant expenditure. Herein, we present an approach that allows for the design of long-term relative bounded motion considering a zonal gravitational model. Entire sets of orbits are obtained via high-order Taylor expansions of Poincarè return maps about reference fixed points. The high-order normal form algorithm is used to determine a change in expansion variables of the map into normal form space, in which the phase space behavior is circular and can be easily parameterized by action–angle coordinates. The action–angle representation of the normal form coordinates is then used to parameterize the original Poincarè return map and average it over a full phase space revolution by a path integral along the angle parameterization. As a result, the averaged nodal period and drift in the ascending node are obtained, for which the bounded motion conditions are straightforwardly imposed. Sets of highly accurate bounded orbits are obtained, extending over several thousand kilometers, and valid for decades.

79 ASTRONOMY AND ASTROPHYSICS↗

Retroperitoneal Sarcoma Target Volume and Organ at Risk Contour Delineation Agreement Among NRG Sarcoma Radiation Oncologists

Purpose: The purpose of this study was to evaluate the variability in target volume and organ at risk (OAR) contour delineation for retroperitoneal sarcoma (RPS) among 12 sarcoma radiation oncologists. Methods and Materials: Radiation planning computed tomography (CT) scans for 2 cases of RPS were distributed among 12 sarcoma radiation oncologists with instructions for contouring gross tumor volume (GTV), clinical target volume (CTV), high-risk CTV (HR CTV: area judged to be at high risk of resulting in positive margins after resection), and OARs: bowel bag, small bowel, colon, stomach, and duodenum. Analysis of contour agreement was performed using the simultaneous truth and performance level estimation (STAPLE) algorithm and kappa statistics. Results: Ten radiation oncologists contoured both RPS cases, 1 contoured only RPS1, and 1 contoured only RPS2 such that each case was contoured by 11 radiation oncologists. The first case (RPS 1) was a patient with a de-differentiated (DD) liposarcoma (LPS) with a predominant well-differentiated (WD) component, and the second case (RPS 2) was a patient with DD LPS made up almost entirely of a DD component. Contouring agreement for GTV and CTV contours was high. However, the agreement for HR CTVs was only moderate. For OARs, agreement for stomach, bowel bag, small bowel, and colon was high, but agreement for duodenum (distorted by tumor in one of these cases) was fair to moderate. Conclusions: For preoperative treatment of RPS, sarcoma radiation oncologists contoured GTV, CTV, and most OARs with a high level of agreement. HR CTV contours were more variable. Further clarification of this volume with the help of sarcoma surgical oncologists is necessary to reach consensus. More attention to delineation of the duodenum is also needed.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

A Semi-Analytical Approach for State-Space Electromagnetic Transient Simulation

Here, this paper proposes a semi-analytical approach for efficient and accurate electromagnetic transient (EMT) simulation of a power grid. The approach first derives a high-order semi-analytical solution (SAS) of the grid’s state-space EMT model using the differential transformation (DT), and then evaluates the solution over enlarged, variable time steps to significantly accelerate the simulations while maintaining its high accuracy on detailed fast EMT dynamics. The approach also addresses switches during large time steps by using a limit violation detection algorithm with a binary search-enhanced quadratic interpolation. Case studies are conducted on EMT models of the IEEE 39-bus system and large-scale systems to demonstrate the merits of the new simulation approach against traditional numerical methods.

electromagnetic transient↗

Semi-supervised Learning of Dynamical Systems with Neural Ordinary Differential Equations: A Teacher-Student Model Approach

Modeling dynamical systems is crucial for a wide range of tasks, but it remains challenging due to complex nonlinear dynamics, limited observations, or lack of prior knowledge. Recently, data-driven approaches such as Neural Ordinary Differential Equations (NODE) have shown promising results by leveraging the expressive power of neural networks to model unknown dynamics. However, these approaches often suffer from limited labeled training data, leading to poor generalization and suboptimal predictions. On the other hand, semi-supervised algorithms can utilize abundant unlabeled data and have demonstrated good performance in classification and regression tasks. We propose TS-NODE, the first semi-supervised approach to modeling dynamical systems with NODE. TS-NODE explores cheaply generated synthetic pseudo rollouts to broaden exploration in the state space and to tackle the challenges brought by lack of ground-truth system data under a teacher-student model. TS-NODE employs an unified optimization framework that corrects the teacher model based on the student's feedback while mitigating the potential false system dynamics present in pseudo rollouts. TS-NODE demonstrates significant performance improvements over a baseline Neural ODE model on multiple dynamical system modeling tasks.

Wang, Yu↗

Resonant Control for Fermilab's PXIE RFQ

The RFQ for Fermilab's PXIE test program is designed to accelerate a < 10 mA H⁻ CW beam to 2.1 MeV. The RFQ has a four-vane design, with four modules brazed together for a total of 4.45 m in length. The RF power required is < 130 kW at 162.5 MHz. A 3 kHz limit on the maximum allowable frequency error is imposed by the RF amplifiers. This frequency constraint must be managed entirely through differential cooling of the RFQ's vanes and outer body and associated material expansion. Simulations indicate that the body and vane coolant temperature should be controlled to within 0.1 degrees C. We present the design of the cooling network and the resonant control algorithm for this structure, as well as results from initial operation.

43 PARTICLE ACCELERATORS↗

A C++ TPSA/DA Library with Python Wrapper

Truncated power series algebra (TPSA) or differential algebra (DA) is often used by accelerator physicists to generate a transfer map of a dynamic system. The map then can be used in dynamic analysis of the system or in particle tracking study. TPSA/DA can also be used in some fast algorithms, eg. the fast multipole method, for collective effect simulation. This paper reports a new TPSA/DA library written in C++. This library is developed based on Dr. Lingyun Yang?s TPSA code, which has been used in MAD-X and PTC. Compared with the original code, the updated version has the following changes: (1) The memory management has been revised to improve the efficiency; (2) A new data type of DA vector is defined and supported by most frequently used operators; (3) Support of inverse trigonometric functions and hyperbolic functions for the DA vector has been added; (4) function composition is revised for better efficiency; (5) a python wrapper is provided. The code is hosted at github and available to the public.

Zhang, Y.↗

Towards a Verifiable Domain-Specific Language for Hardware-Accelerated Stencils

Defining a domain-specific language (DSL) that supports vector-calculus abstractions eases the porting of partial differential equation (PDE) solvers to specialized architectures. Sufficiently high-level abstractions empower users to express universal laws with sufficient generality that the laws must always hold true within their domain of validity. A broad class of PDE solvers employs stencil-based algorithms, the target domain of Berkeley Lab's stencil accelerator chip co-design project. First released as open-source in January 2026, the Formal software framework lays a foundation for defining an embedded DSL based on composable operators that implement mimetic numerical methods -- stencil algorithms that guarantee satisfaction of discrete versions of important vector calculus theorems. The Formal DSL will be the frontend to a new class of stencil-PDE accelerators developed jointly by LBNL, UHCL, and UC Berkeley through the DOE Competitive Portfolios for Computer Science Project. This offers the potential of an order of magnitude acceleration for this important category of computational methods to serve the DOE mission. Future work on the Formal DSL will facilitate software verification via type-safe templates that enable problem-specific correctness proofs relying upon generic function theory and carefully crafted unit tests.

Rouson, Damian↗

CACTI CSAPR2 Taranis Retrievals

Taranis is an end-to-end processing chain for radar data written in Python with C extension for computation performance. Features include: masking for quality control, specific differential phase (Kdp), attenuation correction for reflectivity factor (Z) and differential reflectivity (Zdr) in rain, and additional geophysical retrievals. Retrievals are mostly drawn from literature or open-source software when appropriate, and have been tested, tuned, and modified to work with one another cohesively rather than using isolated off-the-shelf algorithms. Incorporated algorithms include hydrometeor (echo) identification, rain water content, raindrop mass-weighted mean diameter (gamma size distribution assumption), and rainfall rate (QPE). Taranis data sets exist for CSAPR2 PPI, HSRHI, and sector RHI scans. Cartesian-gridded data sets were also produced as well as a near-surface rain rate retrieval. More details can be found in the README.

54 ENVIRONMENTAL SCIENCES↗

Energy stable discontinuous Galerkin methods for Maxwell's equations in nonlinear optical media

The propagation of electromagnetic waves in general media is modeled by the time-dependent Maxwell's partial differential equations (PDEs), coupled with constitutive laws that describe the response of the media. In this work, we focus on nonlinear optical media whose response is modeled by a system of first order nonlinear ordinary differential equations (ODEs), which include a single resonance linear Lorentz dispersion, and the nonlinearity comes from the instantaneous electronic Kerr response and the residual Raman molecular vibrational response. To design efficient, accurate, and stable computational methods, we apply high order discontinuous Galerkin discretizations in space to the hybrid PDE-ODE Maxwell system with several choices of numerical fluxes, and the resulting semi-discrete methods are shown to be energy stable. Under some restrictions on the strength of the nonlinearity, error estimates are also established. When we turn to fully discrete methods, the challenge to achieve provable stability lies in the temporal discretizations of the nonlinear terms. To overcome this, novel strategies are proposed to treat the nonlinearity in our model within the framework of the second-order leap-frog and implicit trapezoidal time integrators. The performance of the overall algorithms are demonstrated through numerical simulations of kink and antikink waves, and third-harmonic generation in soliton propagation.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Investigation of Portable Event-Based Monte Carlo Transport Using the NVIDIA Thrust Library

Power consumption considerations are driving future high performance computing platforms toward many-core computing architectures. Los Alamos National Laboratory's Trinity machine, available in 2016, will use both Intel Xeon Haswell processors and Intel Xeon Phi Knights Landing many integrated core (MIC) architecture coprocessors. Lawrence Livermore National Laboratory's Sierra machine, available in 2018, will use an IBM PowerPC architecture along with Nvidia graphics processing unit (GPU) architecture accelerators. These different advanced architectures make the computing landscape in upcoming years complex. Traditional approaches to Monte Carlo transport do not work efficiently on these new computing platforms. MIC architectures require vectorization to operate efficiently, and vectorization is difficult to achieve in Monte Carlo transport. GPU architectures require additional code to explicitly use the hardware, requiring significant code changes or hardware specific branches in the source code. A significant challenge for Monte Carlo transport projects is to simultaneously support within a single source code base efficient simulations for both the current generation of architectures and the different advanced computing architectures. In order to address these challenges, two important changes are typically required: a new algorithmic approach for solving Monte Carlo transport, and explicit use of hardware specific software. In this paper, we describe initial research investigations of an event-based Monte Carlo transport algorithm implemented using the Nvidia Thrust library on a GPU for a Monte Carlo test code. The event-based algorithm targets many-core architectures by increasing SIMD (single instruction multiple data) parallelism, while Thrust potentially provides portable performance by allowing one source code base to compile code targeted for both CPUs and GPUs. We described preliminary investigations of portable event-based Monte Carlo algorithms implemented using the Nvidia Thrust library in a research Monte Carlo test code. We found that an explicit CUDA implementation of an event-based Monte Carlo algorithm performed significantly more efficiently than a Thrust implementation on GPU platforms, most likely as a result of additional flexibility in access to different memory spaces on the GPU. Additionally, we showed that on GPU platforms and at large enough problem sizes the event-based implementations perform more efficiently than the serial history-based implementation running on the host CPU. While investigating this problem, we also discovered that the performance of the event-based algorithm is affected by what tallies are being used. A zonal scalar flux tally requires atomic operations that significantly impacted the performance of the code, in some cases producing slowdowns instead of speedups. We decided to remove the tally in order to focus on the effectiveness of the event-based algorithm. Future work will be required to research more effective ways of handling such tallies. Additionally, we would like to consider new ways for optimizing both the Thrust and CUDA versions, in order to see how much performance we have yet to achieve. The potential trade-off between portability and performance was demonstrated in this investigation. Thrust provides both CPU and GPU versions of the code in one code base, but it does so at a cost. We discovered approximately a factor of 2-5 performance difference between Thrust and CUDA on each GPU platform we tested. Thrust provides a tool to access the GPUs with less effort and specialization, but it does so by giving up fine grained control where extra performance can be found for this application. Future work will be required to determine whether this performance differential between Thrust and CUDA can be reduced for event-based Monte Carlo transport. (authors)

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

An Explicitly Correlated Wavelet Method for the Electronic Schroedinger Equation

A discretization for an explicitly correlated formulation of the electronic Schroedinger equation based on hyperbolic wavelets and exponential sum approximations of potentials is described, covering mathematical results as well as algorithmic realization, and discussing in particular the potential of methods of this type for parallel computing.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Calibration of elastoplastic constitutive model parameters from full-field data with automatic differentiation-based sensitivities

Here, we present a framework for calibration of parameters in elastoplastic constitutive models that is based on the use of automatic differentiation (AD). The model calibration problem is posed as a partial differential equation-constrained optimization problem where a finite element (FE) model of the coupled equilibrium equation and constitutive model evolution equations serves as the constraint. The objective function quantifies the mismatch between the displacement predicted by the FE model and full-field digital image correlation data, and the optimization problem is solved using gradient-based optimization algorithms. Forward and adjoint sensitivities are used to compute the gradient at considerably less cost than its calculation from finite difference approximations. Through the use of AD, we need only to write the constraints in terms of AD objects, where all of the derivatives required for the forward and inverse problems are obtained by appropriately seeding and evaluating these quantities. We present three numerical examples that verify the correctness of the gradient, demonstrate the AD approach's parallel computation capabilities via application to a large-scale FE model, and highlight the formulation's ease of extensibility to other classes of constitutive models.

42 ENGINEERING↗

Properties of g &#x2192; b b &#xAF; at small opening angles in p p collisions with the ATLAS detector at s = 13 TeV

The fragmentation of high energy gluons at small opening angles is largely unconstrained by present measurements. Gluon splitting to b-quark pairs is a unique probe into the properties of gluon fragmentation because identified b-tagged jets provide a proxy for the quark daughters of the initial gluon. In this study, key differential distributions related to the g->bb process are measured using 33/fb of sqrt(s)=13 TeV pp collision data recorded by the ATLAS experiment at the LHC in 2016. Jets constructed from charged-particle tracks, clustered with the jet anti-kt algorithm with radius parameter R = 0.2, are used to probe angular scales below the R = 0.4 jet radius. The observables are unfolded to particle level in order to facilitate direct comparisons with predictions from present and future simulations.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

SU-E-I-17: Comparison of Two Novel Algorithms for the Modulation Transfer Function of CT Using a Simple Cylindrical Phantom

Purpose: To compare and analyze two novel algorithms for the assessment of modulation transfer functions (MTFs) of computed tomography (CT) systems using a simple acrylic cylindrical phantom Method and Materials: Images of the acrylic cylindrical phantom were acquired by a GE LightSpeed 16 RT (GE Healthcare, Milwaukee, WI) using 120 kVp, 330 mA, 2.5 mm slice thickness, 10 cm field-of view (FOV), four reconstruction kernels (e.g. standard, soft, detail, bone, and lung). Two different algorithms were used to analyze images for MTF assessment. First, Richard et al. suggested a task-based MTF assessment method through an edge spread function (ESF) which described pixel intensities as a function of distance from the center. The MTF was obtained as the absolute value of Fourier transform of the differentiated ESF. Second, Ohkubo et al. devised an effective method to determine the point spread function (PSF) of CT system accompanied with verification. The line spread function (LSF), which was the one-dimensional integration of the PSF, was used to obtain the MTF. We validated the reliability of two above-mentioned methods through the comparison with a conventional method using a thin tungsten wire phantom. Results: The measured MTFs by two methods were mostly similar each other for standard, soft, and detail kernels. In 0.6 lp/mm, the MTF difference between two methods were 0.012(standard), 0.004(soft), and 0.037(detail). They also coincided with the MTF by the conventional method well. However, there were considerable distinctions for bone and lung kernels containing edge enhancement that might cause undershoots near the peak of the LSF. Conclusions: We compared two novel methods to assess task-based MTFs for clinical CT systems especially using a simple acrylic cylindrical phantom with high-convenience and low-cost, and validated them against a conventional method. This work can provide a practical solution to users for the quality assurance of CT.

60 APPLIED LIFE SCIENCES↗

On the Convergence of Overlapping Schwarz Decomposition for Nonlinear Optimal Control

Here, we study the convergence properties of an overlapping Schwarz decomposition algorithm for solving nonlinear optimal control problems (OCPs). The algorithm decomposes the time domain into a set of overlapping subdomains, and solves all subproblems defined over subdomains in parallel. The convergence is attained by updating primal-dual information at the boundaries of overlapping subdomains. We show that the algorithm exhibits local linear convergence, and that the convergence rate improves exponentially with the overlap size. We also establish global convergence results for a general quadratic programming, which enables the application of the Schwarz scheme inside second-order optimization algorithms (e.g., sequential quadratic programming). The theoretical foundation of our convergence analysis is a sensitivity result of nonlinear OCPs, which we call "exponential decay of sensitivity" (EDS). Intuitively, EDS states that the impact of perturbations at domain boundaries (i.e., initial and terminal time) on the solution decays exponentially as one moves into the domain. Here, we expand a previous analysis available in the literature by showing that EDS holds for both primal and dual solutions of nonlinear OCPs, under uniform second-order sufficient condition, controllability condition, and boundedness condition. We conduct experiments with a quadrotor motion planning problem and a partial differential equations (PDE) control problem to validate our theory, and show that the approach is significantly more efficient than alternating direction method of multipliers and as efficient as the centralized interior-point solver.

42 ENGINEERING↗