Tech-X Corporation Federal Contracts
Tech-X Corporation uniquely combines object-oriented software, grid-based middleware, simulation and modeling, and massively parallel computing expertise to assist customers in solving the most difficult scientific problems.
Projects
Active Federal Contract Projects are listed alphabetically by title.
Analysis of RF Heating of Fusion Plasmas Using the Delta-f Particle-in-Cell (DFPIC) Method (DE-FG02-07ER84722) |
Critical to the US fusion program is the ability to assess the efficacy of radiofrequency heating and current drive in fusion plasmas. At present, the primary means of assessment is through use of full wave codes that assume linearity and quasi-local plasma response. Needed are computational methods for addressing the effects caused by the fact that the curvature of particle orbits can be large. The proposed project includes the development and application of delta-f particle-in-cell methods to toroidal plasmas, with the goal of developing analysis methods that apply to low-frequency (ion time scale) phenomena. The project will implement linear delta-f for the ions along with a reduced model for the electrons in R-Z geometry that will eliminate the fast time scales from the problem. This new simulation tool will be used to study physics at the ion time scale and to examine the effect of heating in such tokamaks as NSTX and ITER. |
PI: Travis Austin |
Analyzing and Visualizing Next Generation Climate Data (DE-FG02-08ER85153) |
The continuous effort to improve global warming predictions has led numerous computer climate models to abandon longitude-latitude grids in favor of assemblies of structured meshes. These novel “Mosaic” grids exhibit complex folding patterns, which are posing severe problems to postprocessing and visualization applications (e.g. CDAT). We are developing the Mosaic Data Analysis and Visualization Extension (MoDAVE) software package in order to allow postprocessing applications to correctly interpret data on large Mosaic grids. We use the Gridspec format developed at the Geophysical Fluid Dynamics Laboratory to encode the connectivity between equi-resolution sub-grids (tiles), and flux conserving expressions for other cases. |
PI: Alex Pletzer |
Center for Extended Magnetohydrodynamic Modeling (DE-FC02-08ER54972) |
The proposed work aims at further developing the world’s most powerful simulation codes for studying the macroscopic dynamics of MHD-like phenomena in fusion plasmas, and using them on the most advanced computers to address critical issues facing burning plasma experiments such as ITER. Our 8-institution Center has existed since June 2001. It has already pioneered simulations that have shed new insight in many different application areas and has demonstrated that the algorithms in its major codes can scale to 1000s of processors. We have developed excellent working relationships with the SciDAC applied math and computer science centers TOPS, ITAPS, and APDEC; these collaborations have substantially improved the algorithms and efficiency of our leading tokamak global codes, NIMROD and M3D, and have facilitated the development of a new code, AMRMHD, utilizing adaptive mesh refinement. The physical problems we propose to focus on are: sawteeth, sawteeth in the presence of energetic particles, excitation and control of neoclassical tearing modes, resistive wall modes, disruptions, edge localized modes, error field studies, and pellet fueling. |
PI: Scott Kruger |
Center for Simulation of Wave-Plasma Interactions with Magnetohydrodynamics (DE-FC02-06ER54864) |
The proposed work will build upon the successes of the DOE SciDAC programs by taking several of the most advanced fusion computer codes, combining these to provide a unique tool in the worldwide fusion program, and implementing them on the Leadership-Class and other computing facilities. This will be done in concert with the computer science and math communities in such a way that the software framework we form will lay a foundation for even more powerful and comprehensive simulations in the future. Intense radio-frequency waves provide an essential control technique for fusion grade tokamak plasmas. The frequency and phasing of these waves can be adjusted using precision control so as to keep the plasma in a quiescent state, significantly improving its quality as a confinement device and greatly reducing the likelihood of a catastrophic disruption. The work described here is aimed at providing an essential design tool for such a control system. The framework will allow for the treatment of two regimes of potential plasma global instabilities: fast phenomena and slow phenomena. In the first, the wave fields act only indirectly on the instabilities by modifying the background equilibrium, while in the second they act directly by competing with the instability drive terms. We will work with the community to develop standards and documented APIs for fusion application modules which will allow the incorporation of many of today’s fusion application codes into this framework. |
PI: Scott Kruger |
Common Component Architecture for Electron Cloud Accelerator Simulations (DE-FG02-08ER85152) |
Electron cloud generation is one of the most scientifically important effects relating to the dynamics of particle beams. The build up of an electron cloud is a potentially limiting factor in the performance of both high-intensity electron and proton machines. Electron cloud simulation presents computational challenges in that the number of particles required in simulation is large and because efficient code coupling of beam physics with electron cloud generation code is required. The modeling of electron cloud effects is important for high energy accelerators and the Fermilab main injector. In the case of the Fermilab main injector, electron cloud effects are expected to play a significant role when the main injector operates in the regime of a high-intensity proton source for the neutrino program. Computational accelerator physicists would like to be able model electron cloud generation and beam dynamics in a single self-consistent simulation. We will provide accelerator components that can be used to compose electron cloud simulations on leadership-class supercomputing hardware. Our approach of applying performance analysis and modeling will enable the most efficient use of available resources while addressing portability and usability will ensuring a broad use of the developed accelerator components. |
PI: Nanbor Wang |
Community Petascale Project for Accelerator Science and Simulation (DE-FC02-07ER41499) |
The DOE program of scientific discovery relies heavily on particle accelerators, which comprise 14 of the 28 facilities in the DOE twenty-year outlook on Facilities for the Future of Science. This proposed project, submitted to the Offices of High Energy Physics (HEP), Nuclear Physics (NP), Basic Energy Sciences (BES) and the Office of Advanced Scientific Computing Research (ASCR), will develop a comprehensive computational infrastructure for accelerator modeling and optimization. This project will advance accelerator computational capabilities from the terascale to the petascale to support DOE priorities for the next decade and beyond. The SciDAC1 accelerator project, a partnership of accelerator computationalists, applied mathematicians, and computer scientists, generated a suite of parallel accelerator simulation tools. These were applied to important accelerator projects of the DOE. Under SciDAC2, these tools will be enhanced to contain new capabilities as needed by HEP projects, such as the ILC, the LHC, the Tevatron, and PEP-II, and for Advanced Acceleration research; NP projects, such as CEBAF and RHIC, the CEBAF and RHIC upgrades, RIA, and an NP electron collider, including ELIC and eRHIC; and BES projects, such as LCLS, NSLS-II, SNS, and upgrades to the APS. This simulation suite will contain a comprehensive set of interoperable components for beam dynamics, electromagnetics, electron cooling, and advanced accelerator modeling. Beam dynamics studies will include developing an understanding of the lifetime limits from beam collisions in colliders. Electromagnetics modeling will be used to optimize cavity shapes for increased accelerating gradient and beam current. Electron cooling computations will determine the configuration of cooling systems needed for mitigating beam-beam effects. Advanced accelerator modeling is needed to develop concepts for HEP accelerators beyond the ILC and to develop tabletop electron accelerators for BES and NP projects. In each of these areas, the modeling tools require petascale supercomputers and advanced software for making effective use of these large, parallel platforms. Computational infrastructure in the areas of shape determination and optimization; advanced adaptive meshing; dynamic load balancing; embedded boundaries; component methodologies; performance measurement, assessment and improvement; linear and nonlinear solvers; and visualization will be used and advanced. Consequently, critical to this effort will be the embedded collaborations with the applied mathematics and computer science communities. In the end, high-quality computational tools developed with the best computational physics, applied mathematics, and computer science will be made available to the US particle accelerator community through installation at government laboratories, universities, and industry. |
PI: John Cary |
Customizable Web Service for Efficient Access to Distributed Nuclear Physics Relational Databases (DE-FG02-07ER84757) |
An increasing fraction of the data generated in Nuclear and High-Energy Physics (HENP) is managed in distributed and relational databases. As the size of this data grows and the collaborative nature of HENP experiments increases, the ability to access differently organized relational databases remotely, efficiently and yet in a user-friendly and interoperable manner is becoming very important. HENP community lack tools addressing this need and accommodating related challenges. Tech-X proposes to develop a customizable Web Service for efficient access to distributed Nuclear Physics (NP) databases. The proposed system will consist of a generic Web Service for accessing arbitrary distributed relational databases, a reference client implemented for the Solenoidal Tracker at the RHIC (STAR) experiment at Brookhaven National Laboratory (BNL), and a tool for creation of the high-level and domain-specific clients required by particular applications. |
PI: Mark Green |
Design and Fabrication of Three-Dimensional Photonic Crystal Accelerator Structures (DE-SC0000839) |
Optical-scale photonic crystal waveguides have great potential for future high-energy colliders due to their ability to sustain high field gradients with low loss, using as sources near- infrared lasers which are powerful, efficient, and commercially available. A design exists which has been shown through simulation to meet several requirements of an accelerator, including an accelerating gradient an order of magnitude greater than existing technology, high optical-to-beam efficiency, and a method of stable transverse particle beam confinement. In addition, the structure was designed from the outset to be amenable to lithographic fabrication, in order to take advantage of microfabrication technology from the integrated circuit industry. This structure therefore has the potential to be the basis for an accelerator system in which the various components are integrated in a single monolithic device—an “accelerator on a chip.” Several critical improvements to the structure design are necessary to achieve that goal, including development of compact power couplers and accounting for fabrication error. We propose to address these issues through high-performance simulation together with a systematic study of fabrication effects. For the coupler design problem, we will iteratively run multiple computationally intensive simulations on candidate geometries to optimize coupling efficiency. We will conduct fabrication studies in cooperation with SLAC by fabricating a number of test structures at the Stanford Nanofabrication Facility (SNF). Measurements of these structures will provide data on the fabrication errors of the available equipment. We will incorporate these errors into simulations of both the accelerating waveguide and couplers, to achieve a design for a scalable, integrated photonic accelerator structure. |
PI: Ben Cowan |
Design of Meter-Scale Laser Wakefield Accelerators (DE-SC0000840) |
Time-explicit particle-in-cell (PIC) and fluid simulations provide the most complete description of the laser-plasma interaction and electron acceleration, but the enormous ratio of interaction time to laser oscillation period makes this approach unacceptably slow. Modeling the laser field by its envelope rather than the explicit fields eliminates the need to resolve the laser wavelength, allowing lower resolution and larger time steps. Orders of magnitude speedup have been demonstrated with this technique. We will implement the additional features necessary to extend these envelope model simulations to full meter-scale experiments and to resolve injected electron beams. Simulation results will be compared to those for explicit PIC and validated via comparison with experimental data. |
PI: Ben Cowan |
Designing a Coherent Electron Cooling System for High-Energy Hadron Colliders (DE-FG02-08ER85182) |
The orders-of-magnitude higher ion luminosity required for a future electron-ion collider (EIC) requires cooling of the ion beam. For polarized protons, the only known approach that could work is electron cooling. An exciting concept known as “coherent” electron cooling (CEC) combines the best features of electron cooling and stochastic cooling, via free-electron laser (FEL) technology, to enable the cooling of >250 GeV proton bunches and other high-energy hadron beams with order-of-magnitude shorter cooling times. The CEC concept is unproven and requires detailed simulation of its key components. |
PI: David Bruhwiler |
Efficient Multiscale Algorithms for Modeling Coherent Synchrotron Radiation (DE-SC0000843) |
Collective phenomena such as coherent synchrotron radiation (CSR) and CSR-driven microbunching instability in beam transport and manipulation systems of accelerator-based light sources, if not correctly accounted for, can cause a rapid and irreversible degradation of the electron beam quality. For that reason, there is a need for faster and more accurate CSR modeling algorithms that can be used in designing the next-generation light sources, as well as in exploring upgrade options for existing ones. Straightforward approaches to computing the CSR wakefields in 2D and 3D, at the level of resolution needed in practice, are prohibitively expensive in terms of computational time and memory requirements. As a result, currently available beam dynamics codes rely on algorithms that are based on simplified models and can fail to resolve essential physics in many cases of practical importance. We propose to develop fast and accurate numerical tools for modeling coherent synchrotron radiation that can be used in the range of model parameters relevant to the design, optimization, and commissioning of the BES accelerator facilities such as an ERL-based upgrade to the APS, the X-FELO project, and the LCLS. We will use a novel approach based on recent developments in harmonic and numerical analysis and computational linear algebra. Our approach will simultaneously address the extreme speed and the memory requirements of high-accuracy numerical CSR wakefield computations. We plan to deliver high-fidelity, fast, memory-efficient algorithms for 3D CSR modeling that will directly benefit researchers working on the design of ERL upgrade to APS, design of X-FELO, and simulations in support of LCLS commissioning. |
PI: Ilya Pogorelov |
Examination and Significance of Sparse Preconditioners for High-Order Finite Element Systems (DE-FG02-08ER85154) |
Hundreds of millions of dollars have been committed towards the study of complex natural phenomena on today's massively parallel computers. Access to such computing power is enabling scientists to employ highly-accurate high-order finite element methods to solve previously intractable problems. However high-order finite element methods present new challenges to existing solution methods since the corresponding matrices have fundamental differences and also yield higher memory consumption. A common approach to solving large matrix problems is to use an algebraic mulltigrid preconditioner to improve the effectiveness of the iterative method. Using algebraic multigrid preconditioners for high-order finite element problems yields a preconditioner that requires a large amount of memory. In this work, algebraic multigrid preconditioners generated from sparser matrices than the dense matrices generated from high-order finite elements will be considered as cheaper alternatives to the algebraic multigrid preconditioners generated from high-order finite element matrices. The sparser matrices will have sparsity patterns equivalent to trilinear finite elements on a mesh constructed from the Gauss nodes of the high-order finite element basis functions and yield preconditioners with robust performance for many problems. |
PI: Travis Austin |
Exploration of GPU Computing for Advanced Bird Migration Modeling (NNG09HD00P) |
NASA is using individually-based bird simulation models driven by remotely sensed land surface data, near real-time or retrospective climate and hydrologic data assimilation models, and biological field data to forecast the impact of drought, anthropogenic wetland loss, and climate change on the routes of migratory bird populations. These models are compute intensive, requiring multiple population simulations. The purpose of this project is to explore the use of Graphical Processing Unit tools developed by Tech-X under previous SBIR awards to improve efficiency and throughput of model simulations. |
PI: Peter Messmer |
Extending Chombo with PETSc (DE-SC0000837) |
Among the most challenging problems in computational physics are elliptic equations with vastly different spatial and temporal scales. The block structured Adaptive Mesh Refinement (AMR) technique is particularly amenable to such problems, combining the benefits of “classical” finite difference/finite volume discretization in terms of efficiency, with the advantage of locally increased resolution where needed. The Chombo software infrastructure has been developed for users in need to apply AMR to problems ranging from magnetohydrodynamics to combustion, and has proven to be 1-2 orders of magnitude more efficient than standard finite differences. However, there are problems where Chombo’s built-in linear solvers encounter convergence difficulties, particularly in regimes where time steps are large and heat conductivity, electric resistivity, and/or viscosity have strong spatial variation, as in fusion devices. We propose to extend Chombo so as to allow users to invoke linear matrix solves implemented by the Portable, Extensible Toolkit for Scientific (PETSc) library. PETSc is the leading library for sparse matrix solves, offering a large choice of direct and iterative solver algorithms. |
PI: Alex Pletzer |
Framework Application for Core-Edge Transport Simulations (DE-FC02-07ER54907) |
This project will provide a multiphysics, parallel framework application, FACETS, that will enable whole-device modeling for the U.S. fusion program, and provide the modeling infrastructure needed for ITER, the next step fusion confinement device. FACETS will be highly flexible, through use of modern computational methods, including component technology and object oriented design, to facilitate switching from one model to another for a given aspect of the physics. This will enable use of simplified models for rapid turnaround or high-fidelity models that will take advantage of the largest supercomputer hardware. FACETS will do so in a heterogeneous parallel context, where different parts of the application will take advantage through parallelism based on task farming, domain decomposition, and/or pipelining as needed and applicable. An integral part of the FACETS development will be the coupling of existing core and edge simulations, with the transport and wall interactions described by reduced models. Adding more detailed wall interactions will follow. In the out years, we will bring in greater coupling complexity, including the addition of near-first-principles computations of turbulent transport in the core and the edge. This development plan was created to provide early delivery of capability, with continual delivery of new capability throughout the project. A strong, interdisciplinary team of computational physicists, applied mathematicians, and computer scientists has been assembled for this project in order to address the issues of physics model construction, dynamical systems coupling and solving, and software methodologies and application performance. |
PI: John Cary |
Fully Implicit, Jacobian-free, Newton-Krylov Methods in Production Level MHD Fusion Codes (DE-SC0000838) |
Fully implicit, fully nonlinear algorithms have been shown to yield several orders of magnitude speed-up for magnetohydronamics (MHD) modelling in many academic codes. The implicit time- step has no stability limitations and can step over the fast time scales associated with, for example, the Alfen waves. Solving the nonlinear system of equations presents a challenge, but we rely on current work that has shown the existence of a preconditioner that has the potential to scale remarkably on petascale machines. We focus initially on the momentum equations within the MHD system and develop an implicit, nonlinear solver for this single set of equations. We then extend this work to the full set of equations for 2D problems, illustrating the tremendous potential the method offers. |
PI: Travis Austin |
Fusion Simulation Program (FSP) (DE-SC0001290) |
This project will provide a multiphysics, parallel framework application, FACETS, that will enable whole-device modeling for the U.S. fusion program, and provide the modeling infrastructure needed for ITER, the next step fusion confinement device. FACETS will be highly flexible, through use of modern computational methods, including component technology and object oriented design, to facilitate switching from one model to another for a given aspect of the physics. This will enable use of simplified models for rapid turnaround or high-fidelity models that will take advantage of the largest supercomputer hardware. FACETS will do so in a heterogeneous parallel context, where different parts of the application will take advantage through parallelism based on task farming, domain decomposition, and/or pipelining as needed and applicable. An integral part of the FACETS development will be the coupling of existing core and edge simulations, with the transport and wall interactions described by reduced models. Adding more detailed wall interactions will follow. In the out years, we will bring in greater coupling complexity, including the addition of near-first-principles computations of turbulent transport in the core and the edge. This development plan was created to provide early delivery of capability, with continual delivery of new capability throughout the project. A strong, interdisciplinary team of computational physicists, applied mathematicians, and computer scientists has been assembled for this project in order to address the issues of physics model construction, dynamical systems coupling and solving, and software methodologies and application performance. |
PI: John Cary |
High Fidelity Simulation of Low-Energy Ion Beam Chopping for the Spallation Neutron Source (DE-SC0000844) |
The parallel simulation framework VORPAL will be used to simulate H- ion beam transport experiments being planned as part of the long-term SNS power upgrade project at Oak Ridge National Lab (ORNL). The Phase I project will use data from the planned experiments to validate the utility of the simulations, including complicated electrode geometries and impact ionization physics, and perhaps provide useful insights to the experimental program at ORNL. At the end of Phase II, we will have developed state-of-the-art software capable of designing future low-emittance LEBT configurations. Phase I tasks will resolve the key technical questions and set the stage for Phase II work. |
PI: David Bruhwiler |
High-Fidelity Modulator Simulations of Coherent Electron Cooling Systems (DE-SC0000835) |
The modulator of a CEC system is similar to a conventional electron cooling system, with copropagating ion and electron beams. The key physics to be simulated is the coherent aspect of the electron density and velocity wake created by each ion. This is an extremely subtle and difficult problem to solve – previous simulation efforts have proved only marginally successful. The parallel VORPAL framework, which has successfully simulated the underlying physics of conventional electron coolers, will be used to test the utility of novel algorithms, such as full 3D Vlasov-Poisson and delta-f particle-in-cell. The Phase I project will be devoted entirely to the prototype implementation, testing and comparison of algorithms. The Phase II project will fully implement and optimize the most likely algorithm(s) identified during Phase I. |
PI: David Bruhwiler |
High-Fidelity Simulations of Fixed-Field Alternating Gradient Accelerators (DE-FG02-96ER84508) |
The next-generation particle accelerator for nuclear physics research, necessary for making fundamental advances in this important field, will likely involve high-energy electron-ion collisions. A promising candidate for the cost-efficient acceleration of high-charge electron bunches is the non-scaling fixed-field alternating gradient (FFAG) synchrotron. Existing codes include only some of the key effects required for the accurate design and evaluation of non-scaling FFAG accelerators. |
PI: Dan Abell |
Integrated Multiple Effects Software for Nuclear Physics Applications (DE-FG02-08ER85184) |
Design, analysis, and computer modeling of hardware for nuclear physics accelerators is a complex process, usually involving different software for different physical aspects of the same hardware. In one of the most important examples of this, electromagnetic modeling of RF sources, waveguides, couplers, and cavities is handled by one piece of software, while heat load analysis of the same components is handled by completely different software. The modeling and analysis of this hardware is a critical task since cooling requirements are a main driver of the overall cost of a superconducting rf cavity system. Using separate software tools for modeling of the same hardware has several disadvantages, including: a) duplicative efforts in terms of specifying the same hardware in different software packages, b) difficulties in moving data from one package to another, c) the complexity of the design and analysis iteration cycle, d) and the overall speed of the design cycle. Furthermore, there is strong desire to have the design cycle functioning on modern parallel computing architectures, which can be difficult if using legacy software. Tech-X has been developing multiple effects software, VORPAL, which is designed for modern parallel computing platforms. It is a leading electromagnetic particle simulation code, with demonstrated capabilities in accelerator component modeling, charge-particle simulation, fluid simulation, and plasma modeling. However, missing from its capabilities is thermal and temperature analysis. New multi-purpose field analysis techniques recently introduced into the software permits rapid development and deployment of the desired thermal modeling capability. |
PI: David Smithe |
Inverse Cyclotrons for Intense Muon Beams (DE-FG02-08ER85044) |
Statement of the problem or situation that is being addressed: Intense muon beams are sought for their role in the future of both the high-energy and medium-energy physics programs at national labs, such as BNL and Fermilab. Inverse cyclotrons are a promising alternative to more expensive methods of cooling muon beams, but their ability to cool intense muon beams is untested. We will provide to the muon beams community a software suite capable of providing a complete, end-to-end simulation of an inverse cyclotron. It is uncertain what the acceptance of such a device is for intense muon beams or what kind of losses can be expected. We will develop the particle-in-cell (PIC) code VORPAL to perform detailed simulations of the entire inverse cyclotron, from injection to extraction and initial re-acceleration. |
PI: Kevin Paul |
Library-Based Tuning of Mathematical Kernals on Petascale Systems (DE-FG02-07ER84731) |
Petascale supercomputers offer the computational power to make significant progress to DoE’s most complex electromagnetic problems. Existing simulation tools have to be carefully tuned to take advantage of the high degree of concurrency offered by these systems. In this project, we will optimize a plasma physics modeling tool widely used within DoE's scientific community for leadership-class supercomputing systems. General statement of how this problem is being addressed Good performance on largest scale systems is achieved by addressing algorithm scalability, single processor performance and code portability. By separating the physics algorithms form the numerical kernel routines we will be able to take advantage of highly tuned numerical libraries, thus getting good performance while maintaining a high level of portability. |
PI: Peter Messmer |
Magnetic Insulation and the Effects of External Magnetic Fields on RF Cavity Operation in Muon Accelerators (DE-SC0000841) |
High-gradient RF cavities operating in strong external magnetic fields are required for efficient and cost-effective muon ionization cooling for neutrino factories and muon colliders. It has been observed, however, that strong external magnetic fields degrade the maximum achievable gradient of typical RF cavities. One possible explanation for this is that the external magnetic fields focus accelerated surface-emitted electrons within the cavity to localized regions on the cavity surface, increasing the potential for surface damage and breakdown. If a means can be found for designing high-gradient RF cavities that can operate in such strong external magnetic fields without degraded performance, more cost-effective muon cooling channels can be achieved. We propose to develop numerical models that can be used to help engineers design RF cavities that can operate in external magnetic fields without the potentially adverse effects of surface damage caused by surface-emitted electron bombardment. We plan to develop the parallel 3D PIC plasma simulation code VORPAL to simulate both the microphysics of thermal-dependent surface emission and the macrophysical dynamics of the electrons within the cavity after emission, in the presence of RF ends and external magnetic fields. We will then develop a means of coupling the two scales together, yielding a framework for accurate, full-scale simulations of RF cavity operation to study the likelihood of breakdown with a given cavity design. |
PI: Kevin Paul |
Modeling Accelerator Beam Dynamics Including Superconducting RF Cavities (DE-FG02-96ER84485) |
Successful design and operation of accelerating structures, such as at the heart of the Continuous Electron Beam Accelerator Facility (CEBAF) at Thomas Jefferson National Accelerator Facility (TJNAF), depends on knowledge of how the passage of the particles through the structures excite extraneous signals (wakefields) within them. Wakefields are at the origin of many operational constraints, since they can lead to undesirable emittance growth, instability within the subsequent bunches, undesirable high-order modes (HOMs), and undesirably large power loading to cavity, tube, and coupler structures. The wakefields are a complicated superposition of hundreds to sometimes thousands of oscillating electromagnetic modes, which in turn can depend critically on detailed aspects of the cavity and coupler shapes. The use of computational models to predict wakefield behavior is quite challenging, as existing tools and techniques are insufficient for the task, or require impractically long run times. We are proposing to address the difficult problem of speeding up the computation of wakefields for nuclear physics accelerator cavities by combining two approaches: a) augmenting existing parallel processing software with the Time-Domain Extrapolation Method developed and demonstrated at TJNAF, and b) applying and evaluating a recently demonstrated and validated technique borrowed from another field, the Filter Diagonalization mode-analysis technique. Both approaches will be used to compute a quantity called the Beam Impedance, which characterizes the wakefields seen by a bunch due to the passage through the cavity of the preceding bunches. Finally, the results of the Filter Diagonalization Technique will be used to determine the power loading to cavities, drift tube, and couplers. |
PI: Peter Stoltz |
Modeling of Signal Generation in Gamma-Ray Detectors (DE-FG02-07ER84758) |
Gamma ray detectors are widely used devices in nuclear physics research. The GRETINA gamma ray detector significantly improves upon existing technologies like Gammasphere due to its novel segmentation design, however much work is needed in order to maximize its full capabilities. In particular, there are problematic regions on the detector that could strongly benefit from high fidelity computational modeling. We will develop accurate bulk- and surface-transport models of charged carriers as well as codes for computing the electric fields in the GRETINA crystals. Moreover, we will focus our efforts on poorly understood regions of the devices such as the passivated surface where the measured signals are highly pathological. These developments will yield a more accurate set of basis signals from which one can efficiently determine the location of gamma ray events. Ultimately, this will vastly improve the performance of these detectors. |
PI: Paul Mullowney |
New Boundary Algorithms for Next Generation Simulation and Design of High-Power Microwave Devices (FA9451-07-C-0025) |
We propose to identify, prototype, and test a conformal (curve-matching) boundary algorithm for use in EM-PIC (Electromagnetic Particle-in-Cell) codes. The next generation of weapons for national defense will use directed electromagnetic energy, such as HPM (high-power microwaves). Successfully developing these new weapons systems requires advanced computational modeling, especially modeling that (i) provides sufficient accuracy to predict realistically the behavior of the weapons systems, and (ii) scales well enough to take advantage of large supercomputers. Modeling approaches like finite-element algorithms provide excellent accuracy for field calculations, but do not handle particles accurately or scale well in parallel. Modeling approaches like finite-difference algorithms handle particles well and scale well in parallel, but have difficulty with accurately modeling fields at curved boundaries. However, recent improvements in finite-difference approaches offer the possibility of improving the modeling of fields at curved surfaces. The details of how these new approaches scale with cell size and how one would implement particles are still undetermined, and are the subject of this proposal. |
PI: Chet Nieter |
NIMROD Development and Application for Advanced Simulations of Tokamak Plasmas (DE-FC02-04ER54875) |
Low-frequency, long-wavelength instabilities in tokamak plasmas remain a challenge for theoretical understanding. The most challenging problems; for example, plasma shaping effects; are difficult to study analytically and require a numerical treatment. Because of the temporal and spatial scales, fluid models remain the most effect approach for computational modeling of these instabilities. The efficacy of the fluid model for modeling these nearlycollisionless plasmas depends on how one closes the plasmas. This work plans on improving heuristic “local closures” for computational efficiency. With improved closures, the NIMROD code will be able to improve the understanding of several problems of importance to ITER: disruptions, neoclassical tearing modes, and ELMs. With this improved understanding, numerical scans will be able to explore the extent to which shaping and profile modifications can ameliorate these problems. |
PI: Scott Kruger |
Non-Axisymmetric Modeling of RF in Tokamaks (DE-FG02-09ER55006) |
This project will perform highly detailed, fully 3-D investigation of the ITER RF coupling scenarios, including non-axisymmetric equilibrium and vessel wall. These investigations will use the nascent capability of the software recently developed in the CSWPI SciDAC project, and provide a means of evaluating the performance of these tools on a challenging and important topic to magnetic fusion energy. Specifically, these studies will use: a) integrated time-domain, general geometry, plasma simulation capability, and b) the multi-scale-basis methodology to handle non-axi-symmetric modeling with an accuracy unavailable in frequency-domain methods. The ITER experiment is planning to use resonant magnetic perturbation (RMP) to help suppress edge-localized modes (ELMs). Such perturbation creates an apparent toroidal asymmetry in the plasma equilibrium, when viewed at the time-scale of RF (ICRH, LHCD, ECRH). Yet, in the case of the state-of-the-art RF full-wave modeling tools, such as TORIC and AORSA-2D for example, toroidal symmetry is the well-worn approximation which permits practical solution of the wave fields, by virtue of superposition of independent toroidal modes. ITER presents an additional toroidal asymmetry in its recessed antenna structure, such that the asymmetry of the vacuum vessel wall, and its effect on the implied long distance coupling is also more of a concern than in previous tokamak configurations. Again, existing full-wave modeling tools are not readily capable of treating walls with asymmetry. Finally, the possibility of magnetic islands, even at nominal volume and small field variation, generate a non-axisymmetric plasma volume comparable to that of today’s large tokamaks. Analysis of power deposition into such volumes is not presently modeled, nor is it widely known how it should be characterized, as the usual assignment of power to axi-symmetric flux surfaces is no longer valid. |
PI: David Smithe |
Parallel Validation Tools for Fusion Simulations (DE-SC0000832) |
Numerical simulations play a critical role in exploration of new experimental regimes needed for design and operation of new fusion devices such as ITER. Hence, simulations must be carefully validated against existing experiments. Synthetic diagnostics, the generation of data that can be directly compared to experimental diagnostics, is an excellent test for simulations. The development and use of synthetic diagnostics is often problematic due to incompatibilities of data representation used in simulations and experiments. There is a need for tools to standardize and streamline generation of synthetic diagnostics and a means to run them in a scalable manner. Tech-X therefore proposes to develop a standard to describe diagnostics data, implement a set of synthetic diagnostics modules, and facilitate the workflow related to running diagnostics and sensitivity analysis in high-performance environment as a post-processing step or concurrently, as a co-process, thus making efficient use of high-performance resources and without degrading the performance of the main simulation. The synthetic diagnostics modules will apply transformations to simulation data and produce data directly comparable with experiments. |
PI: Sveta Shasharina |
Plasma Jet Modeling for MIF (DE-SC0000833) |
Los Alamos National Lab is currently developing an experiment to investigate the use of plasma jets to implode magnetized targets in what is called magneto-inertial fusion (MIF). In a typical MIF scheme a solid liner is imploded around a magnetized target, however this poses the problem of placing the liner for each shot and rebuilding mechanical connections - this is know as the stando problem. Plasma jets, on the other hand, can be located far from the target and do no require mechanical connections, therefore a high rep rate reactor becomes much more feasible. Theoretical analysis of implosion using plasma jets suggests that this may be an economical approach to a commercial fusion reactor however a number of issues need to be resolved. These issues will be resolved through experiment in combination with device modeling. We propose to validate and enhance Tech-X's plasma fluids code, TxFluids, for the particular application of plasma jets for MIF. The code is based on shock capturing finite volume and discontinuous Galerkin techniques for MHD, Hall MHD, Two-Fluid plasma models as well as neutral fluids. These fluid models will be used to analyze the effect of instabilities during jet merging and magnetized target implosion in order to learn of potential issues with the concept and approaches to mitigating those issues. |
PI: John Loverich |
QuAI: A Quality Assurance Infrastructure for Data-Centric Applications (DE-SC0000842) |
Data processing is becoming an integral part of modern experiments and should operate with the timeliness and quality necessary to each project. There is a need for a customizable, resource-aware and dynamic mechanism allowing for automatic dissemination of the quality assurance data to the right parties at the right time. The Data Distribution Service (DDS) is a promising new technology that is expected to deliver these features but is not readily available for the quality assurance applications. Tech-X therefore proposes to develop QuAI, a Quality Assurance Infrastructure for timely distribution of quality assurance data between concurrent data processing applications. The infrastructure will consist of reusable core DDS entities and application-specific extensions. In particular, we will target quality assurance for data processing in the upcoming Join Dark Energy Mission (JDEM) cosponsored by HEP/DOE and NASA. QuAI then will be available to the DOE teams implementing the production data processing system of JDEM. |
PI: Sveta Shasharina |
Rapid Prediction of Long Range Wakefields for Beam Impedance and Power Loading in Complex Accelerator Structures (DE-SC0000836) |
Successful design and operation of accelerating structures, such as at the heart of the Continuous Electron Beam Accelerator Facility (CEBAF) at Thomas Jefferson National Accelerator Facility (TJNAF), depends on knowledge of how the passage of the particles through the structures excite extraneous signals (wakefields) within them. Wakefields are at the origin of many operational constraints, since they can lead to undesirable emittance growth, instability within the subsequent bunches, undesirable high-order modes (HOMs), and undesirably large power loading to cavity, tube, and coupler structures. The wakefields are a complicated superposition of hundreds to sometimes thousands of oscillating electromagnetic modes, which in turn can depend critically on detailed aspects of the cavity and coupler shapes. The use of computational models to predict wakefield behavior is quite challenging, as existing tools and techniques are insufficient for the task, or require impractically long run times. We are proposing to address the difficult problem of speeding up the computation of wakefields for nuclear physics accelerator cavities by combining two approaches: a) augmenting existing parallel processing software with the Time-Domain Extrapolation Method developed and demonstrated at TJNAF, and b) applying and evaluating a recently demonstrated and validated technique borrowed from another field, the Filter Diagonalization mode-analysis technique. Both approaches will be used to compute a quantity called the Beam Impedance, which characterizes the wakefields seen by a bunch due to the passage through the cavity of the preceding bunches. Finally, the results of the Filter Diagonalization Technique will be used to determine the power loading to cavities, drift tube, and couplers. |
PI: David Smithe |
Remote Data Exploration with the Interactive Data Language (NNX09CA72C) |
We propose to develop a tool for NASA researchers based on IDL (Interactive Data Language) and DAP (Data Access Protocol) for user-friendly remote data access. IDL is a popular data analysis tool in the NASA research community. For instance, two examples of important NASA missions that rely on IDL are the Solar and Heliospheric Observatory (SOHO) and the Solar TErrestrial RElations Observatory (STEREO). However, a difficulty for many NASA researchers using IDL is that often the data to analyze is located remotely from the scientist and the data is too large to transfer for local analysis (the recent report1 on data standards in the Space and Solar Physics community exemplifies how data access needs are a main concern currently). Researchers have developed a protocol for accessing remote data, DAP, which is used for both SOHO and STEREO datasets. Presently one can use DAP from within IDL, but the DAP-IDL interface is both limited and cumbersome. We propose to develop a more powerful, user friendly interface to DAP for IDL. A final main goal of the Phase II will be to enable user-friendly interaction with multiprocessor and multi-core resources, allowing users to seamlessly retrieve remote data with DAP and take advantage of parallel computing resources for analyzing that data. Where possible, we plan to make these features available in the open-source GDL (Gnu Data Language) version of IDL. Figure 1 shows the prototype GUI interface developed in Phase I. This figure shows the DAP browser with an open NASA data set, including the list of data available for retrieval and meta-information about that data. The Phase II tasks listed above will build on this successful work. We indicate the tremendous potential for this project by including letters of support from the research community (Princeton Plasma Physics Lab and NCAR), the commercial community (Tokyo Electron), and the company that develops IDL (ITT Visual Information Solutions). |
PI: Michael Galloy |
Schema-Based Environment for Configuring, Analyzing and Documenting Integrated Fusion Simulations (DE-FG02-08ER85155) |
The FACETS and SWIM SciDAC efforts, with the upcoming Fusion Simulation Project, form a foundation for a comprehensive, full-device modeling needed for successful operations of ITER and designing the steps beyond. For the software developed in these projects to become a productive tool for a wide range of users, there is a need for a user-friendly interface to facilitate configuring components, setting up integrated simulations and performing simulation analysis such as sensitivity studies, parameter scans and optimization – all efficiently working on a variety of platforms, in particular on the Leadership Computing Facilities (LCFs). Tech-X therefore proposes to develop SECAD - a user-friendly environment for setting up integrated fusion simulations and their analysis, and performing visualization and performance improvement. In Phase I we have developed a prototype of SECAD capable of editing partial configuration files for the SWIM framework and running DAKOTA parameter studies for the IPS simulations using IPS model components. In Phase II we will fully develop SECAD consisting of: (1) infrastructure for integrating FACETS and SWIM with the DAKOTA analysis tool, (2) a Graphical User Interface (GUI) allowing users to edit and validate FACETS and SWIM inputs, visualize their simulation data via VisIt, and set up DAKOTA runs for parameter studies and optimization runs and visualize the output, and (3) performance instrumentation enhancements for all the components developed for a variety of platforms including the LCFs. To promote adoption and foster the scientific goals of FACETS and SWIM, we will develop several DAKOTA scenarios for particular optimization and parameter studies relevant for ITER modeling. This project will be performed in collaboration with the FACETS, SWIM and ParaTools teams and strongly leveraged by their synergistic activities and several ongoing Tech-X efforts. |
PI: Sveta Shasharina |
SciDAC Center for Simulation of Wave-Plasma Interactions (DE-FG02-08ER54953) |
The Massachusetts Institute of Technology (MIT) is submitting this grant application as part of a combined proposal for the SciDAC Center for Simulation of Wave – Plasma Interactions (CSWPI). The principal investigator for MIT is Paul T. Bonoli who is also the principal investigator for the combined project. The world community has joined together to construct ITER, the next scientific and engineering step on the path towards the development of a safe and economically attractive controlled fusion energy source. This device is designed to produce burning plasma conditions for the first time ever in a man-made laboratory plasma. The power to drive the ITER plasma to the burning regime will be supplied primarily with a combination of externally supplied power from radio frequency waves in the ion cyclotron range of frequencies (ICRF) and the injection of energetic ions from either negative ion or neutral beam injection (NINB/NBI) sources, in addition to internally-generated Ohmic heating from the induced plasma current that also serves to create the magnetic equilibrium for the discharge. Since the success of the ITER project depends critically on the ability to create and maintain burning plasma conditions, it is absolutely necessary to have physics-based models that can accurately simulate the RF processes that affect the dynamical evolution of the ITER discharge. In this new proposal MIT will work with other collaborating institutions to take advantage of high performance, massively parallel computing platforms to develop a predictive simulation capability in three distinct areas: First for ICRF antenna-edge coupling processes in order to understand the linear coupling between a 3D antenna structure and plasma, nonlinear parasitic loss mechanisms such as RF sheath formation and parametric decay instability (PDI), and the excitation of surface waves and coaxial modes and their interaction with the 3D geometry and structure of the SOL and vacuum vessel. Second we will work on the development of a predictive simulation capability for core heating and current drive processes in burning plasma, including a complete microscopic description of the interaction of ICRF and lower hybrid (LH) waves with energetic particles produced by the RF itself and with energetic ions produced by fusion reactions and neutral beam injection. Third, we will work on combing the antenna – edge and the core heating / current drive simulation capabilities into a fully integrated model. This activity will require an extensive program of code verification and validation (V&V) and code algorithm refinement and improvement. This core to edge RF description will be ideal for providing macroscopic and microscopic source information needed by MHD and transport codes in fully integrated modeling activities such as the Fusion Simulation Project (FSP) and Focused Integration Initiatives (FII’s). The work proposed here, dealing with multi-physics issues, strong non-linearity, and involving much more intense verification and validation will require even greater use of both capacity and capability high performance computing. |
PI: David Smithe |
Service-Oriented Architecture for Next Generation, Large-Scale Accelerator Control Systems (DE-FG02-08ER85043) |
Traditional frameworks for accelerator control system development do not scale for control systems of next-generation large- scale accelerators that consist of many sub-accelerators and operation teams. A new development paradigm is needed to encourage the robust integration of heterogeneous control subsystems. Most importantly, we need to be able to manage the inherent and accidental complexity of developing and running such large-scale control systems. Tech-X proposes to develop a reference Service-Oriented Architecture (SOA) that promotes multiple levels of loose coupling to increase the robustness and adaptability of overall control applications. Various control subsystems and activities can be implemented separately and packaged into “services” with well-defined “interfaces” for to enable composition of applications. |
PI: Nanbor Wang |
Simulation of Direct-drive Magneto-Inertial Fusion (DE-SC0000831) |
In the magneto-inertial fusion approach based on laser-driven magnetic- ux compression, an imploding target traps and amplifies a pre-seeded magnetic ux. The extremely high magnetic field created by the implosion reduces thermal-conduction losses in the hot spot and enhances alpha energy deposition, leading to increasing hot-spot temperature at lower implosion velocities than required in conventional inertial confinement fusion approach. However,the maximum magnetic field intensity that can be achieved by the direct-driven laser compression is still unclear. The threshold of a compressed magnetic field required to suppress the thermal transport and lower the ignition requirement is unknown. To help answer these questions, Magnetohydrodynamic (MHD) simulations with thermal transport and laser deposition in one and two dimensions is necessary. Benchmarking the simulation results against experiments is important for researchers to improve the design of MIF targets. We propose to develop detailed numerical models of laser-driven magnetic ux compression in two dimensions. We will build on an existing MHD numerical code to assess the feasibility to carry out simulations for laser-driven implosions of large convergence ratio. We plan to use the massively parallel MHD simulation code TxFluids to simulate the propagation of shocks and waves driven by absorbed laser energy in the dense shell that compresses the magnetic field \frozen-in\ the plasma. We will design the simulations to be relevant to the geometries of cylindrical and spherical targets. |
PI: Sean Chuandong Zhou |
Simulation of Short-Range Wakefields in Accelerating Structures for X-Ray Sources (DE-SC0000845) |
Several independent features exist for the well-established particle-in-cell technique which can be used to address this problem. These include the moving window, to reduce the simulation domain and hence the computational e ort; conformal boundary algorithms for accurate representation of curved cavity shapes; and perfect dispersion algorithms, to reduce numerical errors. In this project, we will combine these features so they can be used together to e ciently run accurate simulations of short-range wakefields. We will benchmark the new code features for test cases with known results. |
PI: Ben Cowan |
Simulation Package for Parallel 3D Modeling of an Electron Gun with a Diamond Amplifier (DE-FG02-96ER84509) |
The Relativistic Heavy Ion Collider (RHIC) contributes fundamental advances to nuclear physics by colliding a wide range of ions. A novel electron cooling section, which is a key component of the proposed luminosity upgrade for RHIC, requires the acceleration of high-charge electron bunches with low emittance and energy spread. A promising candidate for the electron source is the recently developed concept of a high quantum efficiency photoinjector with a diamond amplifier. However, there is no available code to use for investigating such a photoinjector via computer simulations. Numerical algorithms will be developed for the generation and transport of secondary electrons through diamond amplifiers in strong electric fields. These algorithms will be implemented in an existing parallel 3D particle-in-cell (PIC) code and then tested against experimental data. The resulting simulation package will be specialized for the purpose of evaluating and designing highcurrent electron sources with diamond amplifiers |
PI: Dimitre Dimitrov |
Simulations of Alpha Wall Load in ITER (DE-SC0000834) |
We propose to improve the workhorse ORNL code DELTA5D, a drift-orbit following Monte-Carlo code, in close collaboration with its primary developer. Two pieces of new physics will be added: full gyro-orbit dynamics and a realistic model of the ITER wall. Several new algorithms will be evaluated and implemented, if proven to work. The current drift-orbit integrator will be compared with two new integrators and the best of the three will be used in the updated version of the code. A full-orbit integrator will be added and a scheme will be implemented for splitting a drift orbit into a set of full orbits, to simultaneously make the simulations efficient and to reduce the statistical noise. Finally, retrograde Monte Carlo will be investigated to see if it can suppress statistical noise for this particular application. |
PI: Johan Carlsson |
Simulations of Waveguide Breakdown (DE-FG02-07ER84833) |
Future neutrino experiments will require neutrino beam intensities beyond the capabilities of today's sources. These experiments will require high-energy neutrinos from muon decay. The muons need to be cooled, and in order to reduce costs the number of cooling elements should minimized. However, breakdown of the accelerating cavities is expected to limit the performance of any proposed beam system. We propose to help researchers use simulation to understand the breakdown of metallic structures planned for muon beam systems, enabling them to reduce the length and thereby the cost of future accelerators. We will implement new physics algorithms relevant to breakdown and add them to an existing library of routines developed to model plasma/material interactions. We will also make existing codes easier to use for non-experts. |
PI: Seth Veitzer |
SVOPME: A Scalable Virtual Organization Privilege Management Environment (DE-FG02-07ER84733) |
Although modern Grid middleware is beginning to support role-based authorization, there is an information disconnect in existing mechanisms between VO and site authorization control. This disconnect prevents privilege policies defined by VOs from propagating to Grid sites automatically. As more Virtual Organizations (VO)s are joining the Grid, manually maintaining and administrating VOs and grid sites becomes very costly. Tech-X Corporation is proposing to develop SVOPME for automating the propagating of VO privileges to Grid sites. We utilize the eXtensible Access Control Markup Language (XACML) for specifying VO privilege policies will develop tools and services to facilitate the functionality. |
PI: Nanbor Wang |
Virtual Cavity Prototyping with VORPAL (DE-SC0000846) |
We propose to improve the workhorse ORNL code DELTA5D, a drift-orbit following Monte-Carlo code, in close collaboration with its primary developer. Two pieces of new physics will be added: full gyro-orbit dynamics and a realistic model of the ITER wall. Several new algorithms will be evaluated and implemented, if proven to work. The current drift-orbit integrator will be compared with two new integrators and the best of the three will be used in the updated version of the code. A full-orbit integrator will be added and a scheme will be implemented for splitting a drift orbit into a set of full orbits, to simultaneously make the simulations efficient and to reduce the statistical noise. Finally, retrograde Monte Carlo will be investigated to see if it can suppress statistical noise for this particular application. |
PI: Chet Nieter |
Virtual Instrumentation Experiment Optimization for High Throughput Scientific Analysis (DE-FG02-08ER85000) |
Large-scale neutron facilities such as the Spallation Neutron Source (SNS) located at Oak Ridge National Laboratory need easy-to-use Portal access to the Department of Energy Leadership Computing Facilities. Demand for neutron beams at the SNS have exceeded expectations resulting in a strong incentive to rapidly proceed with a planned facility expansion. Neutron scientists are very serious about maximizing the usage of these neutron beams and want to be prepared when they arrive at the facility but the virtual experiment capabilities are not sufficient for proper planning. General statement of how this problem is being addressed. Tech-X proposes to extend the Orbiter Virtual File System, developed in collaboration with the Spallation Neutron Source, for providing virtual instrumentation experiment optimization capabilities to neutron scientists. In achieving the project goals not only will the neutron scientist have the ability to perform extensive pre-experiment virtual simulations, thus arriving at the facility better prepared, but during experiment the Orbiter software and tools will provide a high Quality of Service yielding more efficient use of the facility. |
PI: Mark Green |