Search NASA⌕ Search

SEARCH · Search NASA

Results for “Adjoint 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 397 records · Page 22
All sourcesNASA NTRSDOE OSTI
DOE OSTI2016

Sensitivity-Uncertainty Based Nuclear Criticality Safety Validation

These are slides from a seminar given to the University of Mexico Nuclear Engineering Department. Whisper is a statistical analysis package developed to support nuclear criticality safety validation. It uses the sensitivity profile data for an application as computed by MCNP6 along with covariance files for the nuclear data to determine a baseline upper-subcritical-limit for the application. Whisper and its associated benchmark files are developed and maintained as part of MCNP6, and will be distributed with all future releases of MCNP6. Although sensitivity-uncertainty methods for NCS validation have been under development for 20 years, continuous-energy Monte Carlo codes such as MCNP could not determine the required adjoint-weighted tallies for sensitivity profiles. The recent introduction of the iterated fission probability method into MCNP led to the rapid development of sensitivity analysis capabilities for MCNP6 and the development of Whisper. Sensitivity-uncertainty based methods represent the future for NCS validation – making full use of today’s computer power to codify past approaches based largely on expert judgment. Validation results are defensible, auditable, and repeatable as needed with different assumptions and process models. The new methods can supplement, support, and extend traditional validation approaches.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗
DOE OSTI2019

Nonlocal transport in bounded two-dimensional systems: An iterative method

Here, the concept of transport mediated through the dynamics of 'jumping' particles is used to develop an iterative method for obtaining steady-state solutions to the nonlocal transport equation in two dimensions (2D). The technique is self-adjoint and capable of correctly treating spatially non-uniform, asymmetric systems. An appropriate reduced version of the iteration method is used to compare with results obtained with a self-adjoint 1D transport matrix approach. The transport 'jump' probability distribution functions (pdfs) are based on Lévy α-stable distributions. The technique can handle the entire Lévy α-parameter range from one (Lorentz distributions) to two (Gaussian distributions). Cases with α = 2 (standard diffusion) are used to establish the validity of the iterative method. The capabilities of the iterative method are demonstrated by presenting examples from systems with various source configurations, boundary shapes, boundary conditions and spatial variations in parameters.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2013

Nonlinear diffusion acceleration for the multigroup transport equation discretized with S{sub N} and continuous FEM with rattlesnake

Nonlinear diffusion acceleration (NDA) can improve the performance of a neutron transport solver significantly especially for the multigroup eigenvalue problems. The high-order transport equation and the transport-corrected low-order diffusion equation form a nonlinear system in NDA, which can be solved via a Picard iteration. The consistency of the correction of the low-order equation is important to ensure the stabilization and effectiveness of the iteration. It also makes the low-order equation preserve the scalar flux of the high-order equation. In this paper, the consistent correction for a particular discretization scheme, self-adjoint angular flux (SAAF) formulation with discrete ordinates method (S{sub N}) and continuous finite element method (CFEM) is proposed for the multigroup neutron transport equation. Equations with the anisotropic scatterings and a void treatment are included. The Picard iteration with this scheme has been implemented and tested with RattleS{sub N}ake, a MOOSE-based application at INL. Convergence results are presented. (authors)

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2022

Application of Monte Carlo code SHIFT for MSBR dose rate calculations

This paper presents a study that applied the Monte Carlo code SHIFT to calculate the radiological environment within a detailed Molten Salt Breeder Reactor (MSBR) model. It represents one of the first applications of the SHIFT code in stand-alone simulations for non-light water reactors to help demonstrate the code's potential uses in the design, licensing, and operation of advanced reactors. The radiological conditions of the MSBR were modeled when the reactor is at two different operation modes: normal full power and a drained state. The Forward Weighted-Consistent Adjoint Driven Importance Sampling (FW-CADIS) hybrid method in SHIFT was applied successfully to calculate the ex-core neutron and gamma dose rates for the MSBR at full power. Neutron and gamma dose rates within the drain cell were also calculated for the MSBR at the drained state by integrating the source terms obtained from an ORIGEN-S depletion calculation into the SHIFT simulation. The results indicate that in the drained state, the delayed gammas from the depleted fuel salt are the main contributors to the dose rates. (authors)

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗
DOE OSTI2013

An iterative stochastic ensemble method for parameter estimation of subsurface flow models

Parameter estimation for subsurface flow models is an essential step for maximizing the value of numerical simulations for future prediction and the development of effective control strategies. We propose the iterative stochastic ensemble method (ISEM) as a general method for parameter estimation based on stochastic estimation of gradients using an ensemble of directional derivatives. ISEM eliminates the need for adjoint coding and deals with the numerical simulator as a blackbox. The proposed method employs directional derivatives within a Gauss–Newton iteration. The update equation in ISEM resembles the update step in ensemble Kalman filter, however the inverse of the output covariance matrix in ISEM is regularized using standard truncated singular value decomposition or Tikhonov regularization. We also investigate the performance of a set of shrinkage based covariance estimators within ISEM. The proposed method is successfully applied on several nonlinear parameter estimation problems for subsurface flow models. The efficiency of the proposed algorithm is demonstrated by the small size of utilized ensembles and in terms of error convergence rates.

97 MATHEMATICS AND COMPUTING↗
NASA NTRS1988

On nonconvergence of adjoint semigroups for control systems with delays

The spline scheme of Banks and Kappel (1979) is used as an aproximation method for calculating optimal feedback gains for control systems governed by functional differential equations. It is shown that the adjoints of the spline-based approximation scheme for delay equations do not converge strongly.

Burns, John↗
DOE OSTI2019

An Efficient Parallel-in-Time Method for Optimization with Parabolic PDEs

To solve optimization problems with parabolic PDE constraints, often methods working on the reduced objective functional are used. They are computationally expensive due to the necessity of solving both the state equation and a backward-in-time adjoint equation to evaluate the reduced gradient in each iteration of the optimization method. In this study, we investigate the use of the parallel-in-time method PFASST in the setting of PDE-constrained optimization. In order to develop an efficient fully time-parallel algorithm, we discuss different options for applying PFASST to adjoint gradient computation, including the possibility of doing PFASST iterations on both the state and the adjoint equations simultaneously. Here, we also explore the additional gains in efficiency from reusing information from previous optimization iterations when solving each equation. Numerical results for both a linear and a nonlinear reaction-diffusion optimal control problem demonstrate the parallel speedup and efficiency of different approaches.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2018

Evaluating importance maps for TRIPOLI-4{sup R} using deterministic or on-line methods - 25446

Variance reduction is a key ingredient for solving radiation-protection problems with Monte-Carlo particle-transport codes. Many variance-reduction methods require the definition of an importance map and exhibit optimal performance if the importance map is given by the adjoint flux, the solution of the adjoint Boltzmann equation. This paper presents the implementation of the Consistent Adjoint-Driven Importance Sampling (CADIS) methodology via a coupling between the TRIPOLI-4 R Monte-Carlo particle-transport code and the IDT deterministic solver, Additionally, we describe the implementation of a new TRIPOLI-4 R score that makes it possible to estimate the adjoint flux during a direct Monte-Carlo calculation. These new features are expected to simplify the solution of difficult shielding problems. (authors)

61 RADIATION PROTECTION AND DOSIMETRY↗
DOE OSTI2016

On stability criteria for kinetic magnetohydrodynamics

The existence of a potential energy functional in the zero-Larmor-radius collisionless plasma theory of Kruskal & Oberman (Phys. Fluids, vol. 1, 1958 p. 275), Rosenbluth & Rostoker (Phys. Fluids, vol. 2, 1959, p. 23) allows us to derive easily sufficient conditions for linear stability. However, this kinetic magnetohydrodynamics (KMHD) theory does not have a self-adjointness property, making it difficult to derive necessary conditions. In particular, the standard methods to prove that an instability follows if some trial perturbation makes the incremental potential energy negative, which rely on the self-adjointness of the force operator or on the existence of a complete basis of normal modes, are not applicable to KMHD. This paper investigates KMHD linear stability criteria based on the time evolution of initial-value solutions, without recourse to the classic bounds or comparison theorems of Kruskal–Oberman and Rosenbluth–Rostoker for the KMHD potential energy. The adopted approach does not solve the kinetic equations by integration along characteristics and does not require that the particle orbits be periodic or nearly periodic. Most importantly, the investigation of a necessary condition for stability does not require the self-adjointness of the force operator or the existence of a complete basis of normal modes. Thus, it is thereby shown that stability in isothermal ideal-MHD is a sufficient condition for stability in KMHD and that, with a proviso on the long-time behaviour of oscillations about stable equilibria, stability in the double-adiabatic fluid theory, including the variation of the parallel fluid displacement, would be a necessary condition for stability in KMHD.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
NASA NTRSDate not supplied

Adjoint-based Sensitivities of Flutter Predictions based on the Linearized Frequency-domain Approach

Flutter is a critical factor in designing and certifying aircraft. The linearized frequency-domain method offers a lower cost alternative to time-marching computational fluid dynamics for high-fidelity flutter analysis. In this work, adjoint-based sensitivities are added to a flutter analysis based on the linearized frequency-domain method to efficiently compute derivatives of flutter cost functions with respect to design variables or uncertain parameters. The derivation of the adjoint equations, which involve complications such as derivatives of a nonlinear generalized eigenvalue problem with complex-valued inputs and derivatives of the linearized Navier-Stokes equations, is provided. The implemented adjoint terms and derivatives are verified before demonstrating the approach for derivatives of flutter dynamic pressure with respect to Mach number for the AGARD 445.6 wing.

Aeroelasticity↗
NASA NTRS2016

Thermodynamics of Gas Turbine Cycles with Analytic Derivatives in OpenMDAO

A new equilibrium thermodynamics analysis tool was built based on the CEA method using the OpenMDAO framework. The new tool provides forward and adjoint analytic derivatives for use with gradient based optimization algorithms. The new tool was validated against the original CEA code to ensure an accurate analysis and the analytic derivatives were validated against finite-difference approximations. Performance comparisons between analytic and finite difference methods showed a significant speed advantage for the analytic methods. To further test the new analysis tool, a sample optimization was performed to find the optimal air-fuel equivalence ratio, , maximizing combustion temperature for a range of different pressures. Collectively, the results demonstrate the viability of the new tool to serve as the thermodynamic backbone for future work on a full propulsion modeling tool.

Propulsion↗
NASA NTRS1996

Aerodynamic Shape Optimization of Supersonic Aircraft Configurations via an Adjoint Formulation on Parallel Computers

This work describes the application of a control theory-based aerodynamic shape optimization method to the problem of supersonic aircraft design. The design process is greatly accelerated through the use of both control theory and a parallel implementation on distributed memory computers. Control theory is employed to derive the adjoint differential equations whose solution allows for the evaluation of design gradient information at a fraction of the computational cost required by previous design methods. The resulting problem is then implemented on parallel distributed memory architectures using a domain decomposition approach, an optimized communication schedule, and the MPI (Message Passing Interface) Standard for portability and efficiency. The final result achieves very rapid aerodynamic design based on higher order computational fluid dynamics methods (CFD). In our earlier studies, the serial implementation of this design method was shown to be effective for the optimization of airfoils, wings, wing-bodies, and complex aircraft configurations using both the potential equation and the Euler equations. In our most recent paper, the Euler method was extended to treat complete aircraft configurations via a new multiblock implementation. Furthermore, during the same conference, we also presented preliminary results demonstrating that this basic methodology could be ported to distributed memory parallel computing architectures. In this paper, our concern will be to demonstrate that the combined power of these new technologies can be used routinely in an industrial design environment by applying it to the case study of the design of typical supersonic transport configurations. A particular difficulty of this test case is posed by the propulsion/airframe integration.

Reuther, James↗
NASA NTRS1996

Aerodynamic Shape Optimization of Supersonic Aircraft Configurations via an Adjoint Formulation on Parallel Computers

This work describes the application of a control theory-based aerodynamic shape optimization method to the problem of supersonic aircraft design. The design process is greatly accelerated through the use of both control theory and a parallel implementation on distributed memory computers. Control theory is employed to derive the adjoint differential equations whose solution allows for the evaluation of design gradient information at a fraction of the computational cost required by previous design methods (13, 12, 44, 38). The resulting problem is then implemented on parallel distributed memory architectures using a domain decomposition approach, an optimized communication schedule, and the MPI (Message Passing Interface) Standard for portability and efficiency. The final result achieves very rapid aerodynamic design based on higher order computational fluid dynamics methods (CFD). In our earlier studies, the serial implementation of this design method (19, 20, 21, 23, 39, 25, 40, 41, 42, 43, 9) was shown to be effective for the optimization of airfoils, wings, wing-bodies, and complex aircraft configurations using both the potential equation and the Euler equations (39, 25). In our most recent paper, the Euler method was extended to treat complete aircraft configurations via a new multiblock implementation. Furthermore, during the same conference, we also presented preliminary results demonstrating that the basic methodology could be ported to distributed memory parallel computing architectures [241. In this paper, our concem will be to demonstrate that the combined power of these new technologies can be used routinely in an industrial design environment by applying it to the case study of the design of typical supersonic transport configurations. A particular difficulty of this test case is posed by the propulsion/airframe integration.

Reuther, James↗
DOE OSTI2016

Adjoint-based sensitivity analysis for high-energy density radiative transfer using flux-limited diffusion

Uncertainty quantification and sensitivity analyses are a vital component for predictive modeling in the sciences and engineering. The adjoint approach to sensitivity analysis requires solving a primary system of equations and a mathematically related set of adjoint equations. The information contained in the equations can be combined to produce sensitivity information in a computationally efficient manner. Here in this work, sensitivity analyses are performed on systems described by flux-limited radiative diffusion using the adjoint approach. The sensitivities computed are shown to agree with standard perturbation theory and require significantly less computational time. The adjoint approach saves the computational cost of one forward solve per sensitivity, making the method attractive when multiple sensitivities are of interest.

42 ENGINEERING↗
DOE OSTI2012

Nodal weighting factor method for ex-core fast neutron fluence evaluation

The nodal weighting factor method is developed for evaluating ex-core fast neutron flux in a nuclear reactor by utilizing adjoint neutron flux, a fictitious unit detector cross section for neutron energy above 1 or 0.1 MeV, the unit fission source, and relative assembly nodal powers. The method determines each nodal weighting factor for ex-core neutron fast flux evaluation by solving the steady-state adjoint neutron transport equation with a fictitious unit detector cross section for neutron energy above 1 or 0.1 MeV as the adjoint source, by integrating the unit fission source with a typical fission spectrum to the solved adjoint flux over all energies, all angles and given nodal volume, and by dividing it with the sum of all nodal weighting factors, which is a normalization factor. Then, the fast neutron flux can be obtained by summing the various relative nodal powers times the corresponding nodal weighting factors of the adjacent significantly contributed peripheral assembly nodes and times a proper fast neutron attenuation coefficient over an operating period. A generic set of nodal weighting factors can be used to evaluate neutron fluence at the same location for similar core design and fuel cycles, but the set of nodal weighting factors needs to be re-calibrated for a transition-fuel-cycle. This newly developed nodal weighting factor method should be a useful and simplified tool for evaluating fast neutron fluence at selected locations of interest in ex-core components of contemporary nuclear power reactors. (authors)

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗
DOE OSTI2018

Development of a Library for Conducting Monte Carlo Tallies on Heterogeneous Systems

A library is introduced that enables users to conduct Monte Carlo tallies on heterogeneous computing systems. The library can be built with or linked to a Monte Carlo transport code to provide tally capabilities for the project. The library provides a set of tools and an interface for constructing tallies that can be computed on CPUs or GPUs regardless of where the sample data was generated. The library can be incorporated into existing Monte Carlo transport codes at both the research and production level with minimal intrusion into the transport code. The tally library is capable of computing a variety of tallies including track-length histogram (mesh) tallies, collisionbased kernel density estimators, functional expansion tallies, eigenvalue sensitivity coefficients via iterated fission probability, and sensitivities of reaction rates or adjoint-weighted quantities via differential operator sampling. The parallelization methods are discussed and speedups are shown for 2-D and 3-D problems with comparisons to both the tally library CPU implementations as well as the tally capabilities in the host transport code. Performance results and examples are shown for MCNP, MCATK, and OpenMC.

97 MATHEMATICS AND COMPUTING↗
NASA NTRS1974

Rotation in vibration, optimization, and aeroelastic stability problems

The effects of rotation in the areas of vibrations, dynamic stability, optimization, and aeroelasticity were studied. The governing equations of motion for the study of vibration and dynamic stability of a rapidly rotating deformable body were developed starting from the nonlinear theory of elasticity. Some common features such as the limitations of the classical theory of elasticity, the choice of axis system, the property of self-adjointness, the phenomenon of frequency splitting, shortcomings of stability methods as applied to gyroscopic systems, and the effect of internal and external damping on stability in gyroscopic systems are identified and discussed, and are then applied to three specific problems.

Kaza, K. R. V.↗
NASA NTRS1977

Comparison of some optimal control methods for the design of turbine blades

This paper attempts a comparative study of some numerical methods for the optimal control design of turbine blades whose vibration characteristics are approximated by Timoshenko beam idealizations with shear and incorporating simple boundary conditions. The blade was synthesized using the following methods: (1) conjugate gradient minimization of the system Hamiltonian in function space incorporating penalty function transformations, (2) projection operator methods in a function space which includes the frequencies of vibration and the control function, (3) epsilon-technique penalty function transformation resulting in a highly nonlinear programming problem, (4) finite difference discretization of the state equations again resulting in a nonlinear program, (5) second variation methods with complex state differential equations to include damping effects resulting in systems of inhomogeneous matrix Riccatti equations some of which are stiff, (6) quasi-linear methods based on iterative linearization of the state and adjoint equation. The paper includes a discussion of some substantial computational difficulties encountered in the implementation of these techniques together with a resume of work presently in progress using a differential dynamic programming approach.

Desilva, B. M. E.↗