Search NASA⌕ Search

SEARCH · Search NASA

Results for “multiphysics modeling”

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 847 records · Page 47

A configurational force for adaptive re-meshing of gradient-enhanced poromechanics problems with history-dependent variables

Here, we introduce a mesh-adaption framework that employs a multi-physical configurational force and Lie algebra to capture multiphysical responses of fluid-infiltrating geological materials while maintaining the efficiency of the computational models. To resolve sharp gradients of both displacement and pore pressure, we introduce an energy-estimate-free re-meshing criterion by extending the configurational force theory to consider the energy dissipation due to the fluid diffusion and the gradient-dependent plastic flow. To establish new equilibria after remeshing, the local tensorial history-dependent variables at the integration points are first decomposed into spectral forms. Then, the principal values and directions are projected onto smooth fields interpolated by the basis function of the finite element space via the Lie-algebra mapping. Our numerical results indicate that this Lie algebra operator in general leads to a new trial state closer to the equilibrium than the ones obtained from the tensor component mapping approach. A new configurational force for dissipative fluid-infiltrating porous materials that exhibit gradient-dependent plastic flow is introduced such that the remeshing may accommodate the need to resolve the sharp pressure gradient as well as the strain localization. The predicted responses are found to be not influenced by the mesh size due to the micromorphic regularization, while the adaptive meshing enables us to capture the width of deformation bands without the necessity of employing fine mesh everywhere in the domain.

42 ENGINEERING↗

Pixel-Resolved Long-Context Learning for Turbulence at Exascale: Resolving Small-scale Eddies Toward the Viscous Limit

Turbulence plays a crucial role in multiphysics applications, including aerodynamics, fusion, and combustion. Accurately capturing turbulence's multiscale characteristics is essential for reliable predictions of multiphysics interactions, but remains a grand challenge even for exascale supercomputers and advanced deep learning models. The extreme-resolution data required to represent turbulence, ranging from billions to trillions of grid points, pose prohibitive computational costs for models based on architectures like vision transformers. To address this challenge, we introduce a multiscale hierarchical Turbulence Transformer that reduces sequence length from billions to a few millions and a novel RingX sequence parallelism approach that enables scalable long-context learning. We perform scaling and science runs on the Frontier supercomputer. Our approach demonstrates excellent performance up to 1.1 EFLOPS on 32,768 AMD GPUs, with a scaling efficiency of 94\%. To our knowledge, this is the first AI model for turbulence that can capture small-scale eddies down to the dissipative range in three-dimensional turbulence at high Reynolds numbers.

Yin, Junqi [ORNL] (ORCID:0000000338435520)↗

AMReX: a framework for block-structured adaptive mesh refinement

AMReX is a C++ software framework that supports the development of block-structured adaptive mesh refinement (AMR) algorithms for solving systems of partial differential equations (PDEs) with complex boundary conditions on current and emerging architectures. AMR reduces the computational cost and memory footprint compared to a uniform mesh while preserving the local descriptions of different physical processes in complex multiphysics algorithms. Current AMReX-based application codes span a number of areas, including atmospheric modeling, astrophysics, combustion, cosmology, fluctuating hydrodynamics, multiphase flows, and particle accelerators. In particular, the AMReX-Astro GitHub repository holds a number of astrophysical modeling tools based on AMReX. The origins of AMReX trace back to the BoxLib software framework. AMReX supports a number of different time-stepping strategies and spatial discretizations. Solution strategies supported by AMReX range from level-by-level approaches (with or without subcycling in time) with multilevel synchronization to full-hierarchy approaches, and any combination thereof. User-defined kernels that operate on patches of data can be written in C++ or Fortran; there is also a Fortran-interface functionality which wraps the core C++ data structures and operations in Fortran wrappers so that an application code based on AMReX can be written entirely in Fortran.

97 MATHEMATICS AND COMPUTING↗

ECP ST Project 2.3.1.06-STPM08-RAJA (Final Report)

Modern high-performance computing architectures are diverse, with hardware designs that feature multicore CPUs and CPUs plus accelerators, such as GPUs or FPGAs. Achieving good application performance often requires using a programming model and managing application data in ways that are best suited to a particular platform. For large codes that are used daily in production, and are under continual development, architecture-specific ports are untenable. Maintainability requires single-source code that is performance portable across a range of architectures and programming model back-ends. This report describes RAJA, Umpire, and CHAI, three software libraries that provide complementary capabilities to enable application portability for high performance computing (HPC) platforms. In particular, these libraries provide software abstractions that insulate application source code from hardware architecture and programming model-specific implementation details. These development of these libraries was motivated by the needs of multiphysics applications in the Weapons Simulation and Computing Program (WSC) at Lawrence Livermore National Laboratory (LLNL). Funding though the Exascale Computing Project (ECP) Software Technology (ST) project 2.3.1.06 RAJA has fostered outreach and collaboration with the ECP application and software technology development communities. In this report, we describe ECP adoption and integration of these tools to this point. The larger goal of the report is to describe to the ECP community what these projects offer, how they are used, and how to approach adoption and integration of them.

97 MATHEMATICS AND COMPUTING↗

Analysis of Anderson Acceleration on a Simplified Neutronics/Thermal Hydraulics System

ABSTRACT A standard method for solving coupled multiphysics problems in light water reactors is Picard iteration, which sequentially alternates between solving single physics applications. This solution approach is appealing due to simplicity of implementation and the ability to leverage existing software packages to accurately solve single physics applications. However, there are several drawbacks in the convergence behavior of this method; namely slow convergence and the necessity of heuristically chosen damping factors to achieve convergence in many cases. Anderson acceleration is a method that has been seen to be more robust and fast converging than Picard iteration for many problems, without significantly higher cost per iteration or complexity of implementation, though its effectiveness in the context of multiphysics coupling is not well explored. In this work, we develop a one-dimensional model simulating the coupling between the neutron distribution and fuel and coolant properties in a single fuel pin. We show that this model generally captures the convergence issues noted in Picard iterations which couple high-fidelity physics codes. We then use this model to gauge potential improvements with regard to rate of convergence and robustness from utilizing Anderson acceleration as an alternative to Picard iteration.

Toth, Alex↗

Enabling scientific machine learning in MOOSE using Libtorch

A neural-network-based machine learning interface has been developed for the Multiphysics Object-Oriented Simulation Environment (MOOSE). The interface relies on Libtorch, the C++ front-end of PyTorch, and enables an online interaction between modern machine learning algorithms and all the existing simulation, modeling, and analysis processes available in MOOSE. New capabilities in MOOSE include the native generation and training of artificial neural networks together with options to load pretrained neural networks in TorchScript format. Furthermore, the MOOSE stochastic tools module (MOOSE-STM) has been enhanced with neural network-based surrogate and reduced-order model generation options for efficient stochastic analyses. Lastly, a reinforcement learning capability has been added to MOOSE-STM for the interactive control and optimization of complex multiphysics problems.

97 MATHEMATICS AND COMPUTING↗

A Multicomponent Reactive Transport Model for Integrated Surface‐Subsurface Hydrology Problems

Abstract Despite the widespread use of integrated hydrology models in a variety of applications, consideration of multicomponent reactive transport is still not common. The implementation of these processes requires coupling transport at the surface‐subsurface interface and efficient solution of the non‐linear geochemical model that is consistent with the integrated hydrology solution. The Advanced Terrestrial Simulator provides a flexible multiphysics framework that facilitated this process. In this work, the integrated reactive transport process kernel (PK) was weakly coupled to the integrated hydrology PK. In turn, integrated transport and reactions were coupled using an operator splitting approach. This splitting enabled an explicit solution of the integrated transport problem, including a novel algorithm to calculate exchange fluxes across the surface‐subsurface interface and a point‐by‐point solution of the geochemical problem. Geochemical capabilities were added using well‐established external codes, but rather than using a custom interface to each, a generic interface was used that clearly specifies the variables and operations used by the chemistry PK. The implementation is demonstrated with two example simulations: transport of a tracer in a soil column as it saturates over time and water ponds on the surface and reactive transport in a hillslope driven by successive wet‐dry cycles that result in infiltration, runoff and exfiltration processes.

54 ENVIRONMENTAL SCIENCES↗

Continuous conditional generative adversarial networks for data-driven solutions of poroelasticity with heterogeneous material properties

Machine learning-based data-driven modeling can allow computationally efficient time-dependent solutions of PDEs, such as those that describe subsurface multiphysical problems. In this work, our previous approach (Kadeethum et al., 2021d) of conditional generative adversarial networks (cGAN) developed for the solution of steady-state problems involving highly heterogeneous material properties is extended to time-dependent problems by adopting the concept of continuous cGAN (CcGAN). The CcGAN that can condition continuous variables is developed to incorporate the time domain through either element-wise addition or conditional batch normalization. Moreover, this framework can handle training data that contain different timestamps and then predict timestamps that do not exist in the training data. As a numerical example, the transient response of the coupled poroelastic process is studied in two different permeability fields: Zinn & Harvey transformation and a bimodal transformation. The proposed CcGAN uses heterogeneous permeability fields as input parameters while pressure and displacement fields over time are model output. Our results show that the model provides sufficient accuracy with computational speed-up. This robust framework will enable us to perform real-time reservoir management and robust uncertainty quantification in poroelastic problems.

97 MATHEMATICS AND COMPUTING↗

Q3 Report for FY25 Theory and Simulation Performance Target: Development of an integrated modeling framework for fusion reactor design and assessment

This report describes the work and activities carried out towards the completion of each of the following milestones in FY25 Q3: 1. Demonstrate workflow for generating self-consistent CESOL plasma profiles + first wall and divertor loading prediction and generate the CAT plasma and neutron loading needed for further engineering analysis. • Benchmark between two first wall heat flux mapping methods, identify importance of various heat flux sources and physics impact of using fully coupled CESOL vs post-analysis evaluation. 2. Generate medium fidelity parametrized CAD. • Generate parametrized CAD components for the CAT example case via either user-defined modules called within the geometry generation or by defeatured/parametrized CAD, including DCLL blanket matched to divertor boundary and magnets. Define materials, labels, and boundary conditions for passing the mesh to CFD tools. 3. Demonstrate multiphysics magnet analysis. • Demonstrate magnet analysis workflow called from the FREDA workflow, and 4. Demonstrate nuclear analysis. • Add model to OpenFOAM and/or other codes possibly including Diablo to account for tritium diffusion in solids.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

High-to-Low Activities across VVI, AMA and PHI

This milestone report documents work performed in CASL that was broadly focused on the use of validated high-fidelity simulation codes to inform, improve and calibrate low-fidelity codes to facilitate multiphysics coupling, design, performance analysis, and risk assessment. This high-to-low (Hi2Lo) initiative has the following objectives: employ physics encapsulated in high-fidelity codes to improve the functionality of low-fidelity codes, quantify biases in low-fidelity codes, develop and calibrate physical models and closure relations so that low-fidelity codes more accurately incorporate requisite physics, and develop low-fidelity codes that can be effectively coupled to simulate multiphysics behavior. In combination, these efforts are designed to impact both CIPS and DNB. For example, the development and experimental design-based Bayesian calibration of dynamic gas conductance models in CTFFuel will provide the capability to better model crud-induced power shifts. Whereas the present emphasis has focused on the high-fidelity codes STAR-CCM+ and BISON and low-fidelity codes CTF and CTFFuel, future emphasis will be expanded to include MAMBA. During the reporting period, Hi2Lo efforts focused on two broad components: demonstration of the experimental design-based Bayesian calibration framework to calibrate CTF using both STAR-generated synthetic data and experimental data, and the use of STAR to inform and improve the turbulence modeling and grid heat transfer modeling capabilities in CTF. A complementary milestone focused on the implementation and verification of fuel models in CTFFuel, for which we demonstrated initial calibration results. Future Hi2Lo work will focus on the use of the experimental design-based Bayesian calibration framework, demonstrated this year for STAR and CTF, to calibrate parameters in the fuel swelling, fuel irradiation-induced densification, and fuel relocation models in CTFFuel using both experimental and BISON-generated data. A second component will focus on the continued use of STAR to improve grid modeling capabilities in CTF.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗

Advancements in Multiphysics Microdepletion Analysis of an eVinci TM -like Microreactor Leveraging OpenMC-CRAB Workflow

Nuclear microreactors (MRs) are a class of nuclear reactor technology, characterized by reduced dimensions, modular design, and reduced power output in contrast to conventional Light Water Reactors (LWRs). MRs are proposed for supplying electricity and eventual process heat to remote locations, such as military installations and disaster-affected areas. Current research work sponsored by the US Department of Energy Microreactor Program (MRP) is devoted to the development of novel modeling and simulation tools to better support MR vendors and regulatory bodies. Notably, the NRC is projected to utilize the CRAB multiphysics software driver for executing both design and beyond-design-basis accident analyses. Furthermore, the NRC has been utilizing the MELCOR code to calculate mechanistic source terms during accidents. Since MELCOR relies on isotopic inventory and reactor temperature/power profiles under accident conditions, which theoretically can be derived from CRAB, the goal is to establish a comprehensive CRAB-MELCOR computational framework. Past work was focused on testing and demonstrating CRAB's capability to generate results that can be used to inform mechanistic source term calculations in MELCOR. In particular, a computational workflow leveraging OpenMC-generated microscopic cross sections and CRAB was first applied to perform multiphysics microscopic depletion calculation followed by an accident scenario for a stylized microreactor problem. In fiscal year 2024, the research work has been focused on applying the OpenMC-CRAB workflow, which was first tested in fiscal year 2023, to a realistic 3D heat-pipe cooled MR problem representative of the eVinci TM design. The latter computational problem was developed with inputs from WEC to conserve selected neutronic and thermal characteristics of the eVinci TM design without releasing proprietary data. The results of this simulation, encompassing isotopic inventory, power density distribution, and kinetic parameters, will inform both MELCOR and the WEC-developed FATE code for mechanistic source terms calculations. The results from the two codes will then be compared for code verification purposes. This report contains the design characteristics of the realist heat pipe cooled microreactor developed as a use-case for the verification exercise, and the current results for the multiphysics microscopic depletion performed with the OpenMC-CRAB workflow. The results include eigenvalue as a function of time, power distribution at EOL, in addition to nuclides inventory's time evolution and spatial distribution. Finally, we report improvements to the workflow efficiency achieved through a collaboration with the NEAMS programs. Through this collaborative effort, we were able to strongly decrease the computational time for the multiphysics microdepletion calculation (i.e., from 17.4 hours to 5.7 hours on 280 processors) in addition to simplifying the interface to generate isotopics spatial distribution utilizable by FATE and MELCOR. Future work, including the improvement of the current microscopic cross-sections' library and the simulation of an accident scenario at EOL, is also discussed.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗

Prediction of the Lorentz Force Detuning and pressure sensitivity for a Pillbox cavity

The Lorentz Force Detuning (LFD) and the pressure sensitivity are two critical concerns during the design of a Superconducting Radio Frequency (SRF) cavity resonator. The mechanical deformation of the bare Niobium cavity walls, due to the electromagnetic fields and fluctuation of the external pressure in the Helium bath, can dynamically and statically detune the frequency of the cavity and can cause beam phase errors. The frequency shift can be compensated by additional RF power, that is required to maintain the accelerating gradient, or by sophisticated tuning mechanisms and control-compensation algorithms. Passive stiffening is one of the simplest and most effective tools that can be used during the early design phase, capable of satisfying the Radio Frequency (RF) requisites. This approach requires several multiphysics simulations as well as a deep mechanical and RF knowledge of the phenomena involved. In this paper, is presented a new numerical model for a pillbox cavity that can predict the frequency shifts caused by the LFD and external pressure. This method allows to greatly reduce the computational effort, which is necessary to meet the RF requirements and to keep track of the frequency shifts without using the time consuming multiphysics simulations.

43 PARTICLE ACCELERATORS↗

Solution of the Hot Zero Power BEAVRS Using a Pin-Cell Homogenized Model

The validation of a reactor physics code system is an essential part of the code development process. It allows for the evaluation of the code's ability to predict and model the complex physical processes that take place in a nuclear reactor under various operational conditions. Idaho National Laboratory (INL) has been developing a set of reactor physics analysis tools within the MOOSE framework. MAMMOTH is the main reactor physics application, which integrates other MOOSE applications to solve complex multiphysics problems. Results from this study demonstrate that the MOOSE based reactor physics tools, currently under development at INL, have the flexibility to model with high accuracy the HZP condition for a standard PWR. The DRAGON5-Rattlesnake full core pin-cell homogenized simulation produces eigenvalues and detector signals that are comparable to those generated with OpenMC and nTRACER. However, many improvements are necessary to achieve higher fidelity results.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗

Dimming the Lights: 2D Simulations of Deflagrations of Hybrid C/O/Ne White Dwarfs Using FLASH

The dimmest and most numerous outlier of the Type Ia supernova population, the Type Iax event, is increasingly being found in the results of observational campaigns. There is currently no single accepted model to describe these events. This 2D study explores the viability of modeling Type Iax events as a hybrid C/O/Ne white dwarf progenitor undergoing a deflagration using the multiphysics software FLASH. This hybrid was created using the stellar evolution code MESA, and its C-depleted core and mixed structure have demonstrated lower yields than traditional C/O progenitors in previous deflagration-to-detonation studies. To generate a sample, 30 “realizations” of this simulation were performed, the only difference being the shape of the initial match head used to start the deflagration. Consistent with earlier work, these realizations produce the familiar hot dense bound remnant surrounded by sparse ejecta. Our results indicate that the majority of the star remains unburned (∼70%) and bound (>90%). Our realizations produce total ejecta yields on the order of 10 −2 –10 −1 M ☉ , ejected 56 Ni yields on the order of 10 −4 –10 −2 M ☉ , and ejecta kinetic energies on the order of 10 48 –10 49 erg. Compared to yields inferred from recent observations of the dimmest Type Iax events—SN 2007qd, SN 2008ha, SN 2010ae, SN 2019gsc, SN 2019muj, SN 2020kyg, and SN 2021fcg—our simulation produces comparable 56 Ni yields but too-small total yields and kinetic energies. Reignition of the remnant is also seen in some realizations.

Hydrodynamical simulations↗

MOOSE framework enhancements for meshing reactor geometries

MOOSE is an open-source, parallel finite element framework designed to permit rapid development of robust multi-physics modeling capabilities. Under the DOE-NEAMS program, numerous solvers have been developed utilizing the open-source MOOSE framework for multiphysics reactor analysis. These solvers require input finite element meshes representing the discretized geometry. Typically, reactor analysts turn to licensed external tools for creation of reactor geometry meshes. Recently, enhancements have been added to the MOOSE framework to mesh common reactor geometries and improve MOOSE-based application user workflows. Support for hexagonal pins, assemblies, and cores has been added, and Cartesian support has been extended. Options for modeling static and rotating control drums within a hexagonal assembly are now available. Pin, assembly, and plane regions can be identified through automatically applied tags on the mesh called 'reporting IDs' for easier post- processing of physics results. An external open-source triangle routine has been leveraged within MOOSE to mesh core periphery zones. A set of reactor geometry builder routines further streamline the construction of hexagonal and Cartesian cores and include the ability to assign materials to regions during mesh generation. The new meshing routines are available through the MOOSE framework in the open-source 'Reactor' module, and the resulting directly within MOOSE-based applications or exported as Exodus II files for use in other finite element solvers. (authors)

22 GENERAL STUDIES OF NUCLEAR REACTORS↗

Molten Salt Reactor Experiment Simulation using Shift/Griffin

The Department of Energy (DOE)’s NEAMS focuses its efforts on the development of advanced modeling and simulation (M&S) tools for light-water reactors (LWRs) and non–LWRs (i.e., molten salt reactors, high-temperature gas reactors, microreactors, and fast reactors). In the previous fiscal year, the Multiphysics Applications Driver Technical Area funded molten salt reactor (MSR) M&S at Oak Ridge National Laboratory (ORNL) to generate multigroup macroscopic cross sections with Shift for a MSRE 2D lattice model in Griffin. In addition, Shift’s capability to calculate gamma dose rates from activated components in the primary exchangers in a molten salt breeder reactor was also demonstrated. In fiscal year 2023, ORNL generated multigroup macroscopic cross sections using Shift for a 3D MSRE core model. MSRE depletion calculations using Griffin were also demonstrated in this fiscal year. For the depletion calculation, one-group microscopic cross sections for the 3D MSRE core were generated using Shift, and the decay transmutation library from ORIGEN was converted to an ISOXML file, which is required as input in Griffin. Several Monte Carlo codes, such as OpenMC and Serpent, were also used to benchmark and supplement multigroup cross sections generated by Shift. Multigroup libraries were generated with 8 and 20 group structures, and the study found the 8-group structure to be more accurate when comparing Griffin results to continuous energy (CE) Monte Carlo results. The average flux from CE Shift calculations is up to ~6% higher than the CE Serpent calculations because of different values applied for the energy released per fission (κ values). The average flux in the fuel salt calculated by Griffin using cross sections generated with Shift agrees well with the reference CE Shift solution; the same is valid for the corresponding Serpent results. The maximum relative error is ~6% and ~2% compared to the CE Shift and Serpent reference solutions, respectively. Meanwhile, the average flux calculated by Griffin in the graphite moderator shows a higher difference in the thermal range when compared to both reference Monte Carlo solutions; this result suggests a need for improvement in cross section generation for the graphite moderator in the thermal range in both Monte Carlo codes. Griffin depletion calculations using cross sections from Shift and Serpent were performed and compared against ORIGEN calculations, and the nuclide densities calculated by Griffin were found to be generally in agreement with those of ORIGEN. Because a different approach was taken to calculate the energy released per fission ( κ values) in Shift and Serpent, a difference in nuclide densities from differences in the average flux was observed between Griffin using Serpent and Shift cross sections. Griffin calculations with Shift cross sections produced higher average flux in the salt than with Serpent cross sections, leading to higher consumption of 235 U and higher production of 135 Xe. For time-dependent depletion calculations, cross sections were generated with Serpent, and Griffin’s results using these cross sections were compared to CE Serpent depletion results, demonstrating good agreement. The average difference in keff between Serpent and Griffin as a function of burnup is about 155 pcm. Similarly, good agreement with small differences up to ~0.5% was also noticed in the nuclide density of 235 U and 135 Xe. More details regarding the methodologies invoked to generate the cross section to make code-to-code comparisons are discussed further in this report. User feedback on Griffin and Shift capabilities that will enhance these calculations is provided in this report for future consideration. The work performed this fiscal year can be extended further for multiphysics coupling of Griffin-Pronghorn/SAM with Mole to study precursor flow and salt chemistry.

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗

Enabling Hyper-Differential Sensitivity Analysis for Ill-Posed Inverse Problems

Inverse problems constrained by partial differential equations (PDEs) play a critical role in model development and calibration. In many applications, there are multiple uncertain parameters in a model that must be estimated. However, high dimensionality of the parameters and computational complexity of the PDE solves make such problems challenging. A common approach is to reduce the dimension by fixing some parameters (which we will call auxiliary parameters) to a best estimate and use techniques from PDE-constrained optimization to estimate the other parameters. In this article, hyper-differential sensitivity analysis (HDSA) is used to assess the sensitivity of the solution of the PDE-constrained optimization problem to changes in the auxiliary parameters. Foundational assumptions for HDSA require satisfaction of the optimality conditions which are not always practically feasible as a result of ill-posedness in the inverse problem. Here we introduce novel theoretical and computational approaches to justify and enable HDSA for ill-posed inverse problems by projecting the sensitivities on likelihood informed subspaces and defining a posteriori updates. Our proposed framework is demonstrated on a nonlinear multiphysics inverse problem motivated by estimation of spatially heterogeneous material properties in the presence of spatially distributed parametric modeling uncertainties.

97 MATHEMATICS AND COMPUTING↗

Partitioned exponential methods for coupled multiphysics systems

Multiphysics problems involving two or more coupled physical phenomena are ubiquitous in science and engineering. This work develops a new partitioned exponential approach for the time integration of multiphysics problems. After a possible semi-discretization in space, the class of problems under consideration is modeled by a system of ordinary differential equations where the right-hand side is a summation of two component functions, each corresponding to a given set of physical processes. The partitioned-exponential methods proposed herein evolve each component of the system via an exponential integrator, and information between partitions is exchanged via coupling terms. Here, the traditional approach to constructing exponential methods, based on the variation-of-constants formula, is not directly applicable to partitioned systems. Rather, our approach to developing new partitioned-exponential families is based on a general-structure additive formulation of the schemes. Two method formulations are considered, one based on a linear-nonlinear splitting of the right hand component functions, and another based on approximate Jacobians. The paper develops classical (non-stiff) order conditions theory for partitioned exponential schemes based on particular families of T-trees and B-series theory. Several practical methods of third order are constructed that extend the Rosenbrock-type and EPIRK families of exponential integrators. Several implementation optimizations specific to the application of these methods to reaction-diffusion systems are also discussed. Numerical experiments reveal that the new partitioned-exponential methods can perform better than traditional unpartitioned exponential methods on some problems.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗