SEARCH · Search NASA
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.
There are no author-identified significant results in this report.
Particle Accelerators are complex machine with many pieces of equipment running in synchronization to deliver required beam. However, faults in particle accelerators reduce the availability of the beam for experiments affecting the overall science output. To avoid these faults, we apply anomaly detection techniques to predict any unusual behavior and perform preemptive actions to improve the total availability. Many researchers have adopted semi-supervised Machine Learning (ML) methods such as auto-encoders and variational auto-encoders for such tasks. However, supervised ML techniques designed for similarity learning such as Siamese Neural Network (SNN) can outperform semi-supervised or unsupervised methods for anomaly prediction. One of the challenges associated with application of ML models to particle accelerators is the variability in observed data over time due to system configuration changes. We employ conditional models such as Conditional Siamese Neural Networks (CSNN), and Conditional-VAE (CVAE) to learn the variability in the data by using beam configuration parameters as conditional input. We apply these models for errant beam prediction at Spallation Neutron Source accelerator under different system configurations and compare their performance. We demonstrate that CSNN outperforms CVAE in our application. This talk will present the data source, collection, analysis, data-preparation, model development, hyper-parameter studies and the results.
Many present and future particle accelerators are limited by the maximum electric gradient and peak surface fields that can be realized in RF cavities. Despite considerable effort, a comprehensive theory of RF breakdown has not been achieved and mitigation techniques to improve practical maximum accelerating gradients have had only limited success. Part of the problem is that RF breakdown in an evacuated cavity involves a complex mixture of effects, which include the geometry, metallurgy, and surface preparation of the accelerating structures and the make-up and pressure of the residual gas in which plasmas form. Studies showed that high gradients can be achieved quickly in 805 MHz RF cavities pressurized with dense hydrogen gas, as needed for muon cooling channels, without the need for long conditioning times, even in the presence of strong external magnetic fields. This positive result was expected because the dense gas can practically eliminate dark currents and multipacting. In this project we used this high pressure technique to suppress effects of residual vacuum and geometry that are found in evacuated cavities in order to isolate and study the role of the metallic surfaces in RF cavity breakdown as a function of magnetic field, frequency, and surface preparation. One of the interesting and useful outcomes of this project was the unanticipated collaborations with LANL and Fermilab that led to new insights as to the operation of evacuated normal-conducting RF cavities in high external magnetic fields. Other accomplishments included: (1) RF breakdown experiments to test the effects of SF6 dopant in H2 and He gases with Sn, Al, and Cu electrodes were carried out in an 805 MHz cavity and compared to calculations and computer simulations. The heavy corrosion caused by the SF6 components led to the suggestion that a small admixture of oxygen, instead of SF6, to the hydrogen would allow the same advantages without the corrosion in a practical muon beam line. (2) A 1.3 GHz RF test cell capable of operating both at high pressure and in vacuum with replaceable electrodes was designed, built, and power tested in preparation for testing the frequency and geometry effects of RF breakdown at Argonne National Lab. At the time of this report this cavity is still waiting for the 1.3 GHz klystron to be available at the Wakefield Test Facility. (3) Under a contract with Los Alamos National Lab, an 805 MHz RF test cavity, known as the All-Seasons Cavity (ASC), was designed and built by Muons, Inc. to operate either at high pressure or under vacuum. The LANL project to use the (ASC) was cancelled and the testing of the cavity has been continued under the grant reported on here using the Fermilab Mucool Test Area (MTA). The ASC is a true pillbox cavity that has performed under vacuum in high external magnetic field better than any other and has demonstrated that the high required accelerating gradients for many muon cooling beam line designs are possible. (4) Under ongoing support from the Muon Acceleration Program, microscopic surface analysis and computer simulations have been used to develop models of RF breakdown that apply to both pressurized and vacuum cavities. The understanding of RF breakdown will lead to better designs of RF cavities for many applications. An increase in the operating accelerating gradient, improved reliability and shorter conditioning times can generate very significant cost savings in many accelerator projects.
Male, Long-Evans hooded rats were instrumented for monitoring core and hypothalamic temperatures as well as shivering and nonshivering thermogenesis in response to decreased ambient temperature in order to characterize the nature of the neural controller of temperature in rats at 1G and evaluate chronic implantation techniques for the monitoring of appropriate parameters at hypergravic fields. The thermoregulatory responses of cold-exposed rats at 2G were compared to those at 1G. A computer model was developed to simulate the thermoregulatory system in the rat. Observations at 1 and 2G were extended to acceleration fields of 1.5, 3.0 and 4.0G and the computer model was modified for application to altered gravity conditions. Changes in the acceleration field resulted in inadequate heat generation rather than increased heat loss. Acceleration appears to impair the ability of the neurocontroller to appropriately integrate input signals for body temperature maintenance.
Using NVIDIA graphics processing units (GPUs) equipped with Tensor Cores has enabled the significant acceleration of general matrix multiplication (GEMM) for applications in machine learning (ML) and artificial intelligence (AI) and in high-performance computing (HPC) generally. The use of such power-efficient, specialized accelerators can provide a performance increase between 8 × and 20 ×, albeit with a loss in precision. However, a high level of precision is required in many large scientific and HPC applications, and computing in single or double precision is still necessary for many of these applications to maintain accuracy. Fortunately, mixed-precision methods can be employed to maintain a higher level of numerical precision while also taking advantage of the performance increases from computing with lower-precision AI cores. With this in mind, we extend the state of the art by using NVIDIA’s new TF32 framework. This new framework not only burdens some constraints of the previous frameworks, such as costly 32 16-bit castings but also provides an equivalent precision and performance by using a much simpler approach. We also propose a new framework called TF64 that attempts double-precision arithmetic with low-precision Tensor Cores. Although this framework does not exist yet, we validated the correctness of this idea and achieved an equivalent of 64-bit precision on 32-bit hardware.
Graphics Processing Units (GPUs), the dominantly adopted accelerators in HPC systems, are susceptible to transient hardware fault. New generation of GPUs feature mixed-precision architectures such as NVIDIA Tensor Cores to accelerate matrix multiplications. While being widely adapted, how would they behave under transient hardware faults remain unclear. In this study, we conduct a large-scale fault injection experiments on GEMM kernels implemented with different floating-point data types on the V100 and A100 Tensor Cores, and show distinct error resilience characteristics for the GEMMS with different formats. In the future, we plan to explore this space by building precision-aware floating-point fault tolerance techniques for applications such as DNNs that exercise low-precision computations.
The FLUKA Monte Carlo Radiation Transport and Interaction code package is widely used to simulate the interaction of particles with matter in a variety of fields, including high energy physics, space radiation, medical applications, radiation protection and shielding assessments, accelerator studies, astrophysical studies and well logging. This paper gives a brief overview of the FLUKA program and describes recent developments, in particular, improvements in the modelling of particle interactions and transport are described in detail. In addition, an overview of selected applications is given.
Modern particle accelerators and high-energy physics experiments that deployed up to several hundred of accelerating superconducting RF cavities require accurate frequency control. This is achieved by using cavity tuners typically actuated with the piezoelectric ceramic actuators. Piezoelectric ceramic actuators have become "standard" components of the SRF cavity tuner and depending on the application could be operated in different environments: in air, at cryogenic temperature, in vacuum, and submerged in liquid helium. Different applications place different requirements on the piezo actuators, but the important parameters common to all applications are the lifetime and reliability of the actuators. Several programs targeting the development of reliable piezo actuators are presented in this contribution.
In the 1870s, English photographer Eadweard Muybridge captured motion pictures within one cycle of a horse’s gallop, which settled a hotly debated question of his time by showing that the horse became temporarily airborne. In the 1940s, Manhattan project photographer Berlin Brixner captured a nuclear blast at a million frames per second, and resolved a dispute about the explosion’s shape and speed. In this project, we developed methods to capture detailed motion pictures of evolving, light-velocity objects created by a laser pulse propagating through matter. These objects include electron density waves used to accelerate charged particles, laser-induced refractive index changes used for micromachining, and ionization tracks used for atmospheric chemical analysis, guide star creation and ranging. Our “movies”, like Muybridge’s and Brixner’s, are obtained in one shot, since the laser-created objects of interest are insufficiently repeatable for accurate stroboscopic imaging. Our high-speed photographs have begun to resolve controversies about how laser-created objects form and evolve, questions that previously could be addressed only by intensive computer simulations based on estimated initial conditions. Resolving such questions helps develop better tabletop particle accelerators, atmospheric ranging devices and many other applications of laser-matter interactions. Our photographic methods all begin by splitting one or more “probe” pulses from the laser pulse that creates the light-speed object. A probe illuminates the object and obtains information about its structure without altering it. We developed three single-shot visualization methods that differ in how the probes interact with the object of interest or are recorded. (1) Frequency-Domain Holography (FDH). In FDH, there are 2 probes, like “object” and “reference” beams in conventional holography. Our “object” probe surrounds the light-speed object, like a fleas swarming around a sprinting animal. The object modifies the probe, imprinting information about its structure. Meanwhile, our “reference” probe co-propagates ahead of the object, free of its influence. After the interaction, object and reference combine to record a hologram. For technical reasons, our recording device is a spectrometer (a frequency-measuring device), hence the name “frequency-domain” holography. We read the hologram electronically to obtain a “snapshot” of the object’s average structure as it transits the medium. Our published work shows numerous snapshots of electron density waves (“laser wakes”) in ionized gas (“plasma”), analogous to a water wake behind a boat. Such waves are the basis of tabletop particle accelerators, in which charged particles surf on the light-speed wave, gaining energy. Comparing our snapshots to computer simulations deepens understanding of laser wakes. FDH takes snapshots of objects that are quasi-static --- i.e. like Muybridge’s horse standing still on a treadmill. If the object changes shape, FDH images blur, as when a subject moves while a camera shutter is open. Many laser-generated objects of interest do evolve as they propagate. To overcome this limit of FDH, we developed .... (2) Frequency-Domain Tomography (FDT). In FDT, 5 to 10 probe pulses are fired simultaneously across the object’s path at different angles, like a crossfire of bullets. The object imprints a “streaked” record of its evolution on each probe, which we record as in FDH, then recover a multi-frame “movie” of the object’s evolving structure using algorithms of computerized tomography. When propagation distance exceeds a few millimeters, reconstructed FDT images distort. This is because the lenses that image probes to detector have limited depth of field, like cameras that cannot focus simultaneously on both nearby and distant objects. But some laser-generated objects of interest propagate over meters. For these applications we developed … (3) Multi-Object-Plane Phase-Contrast Imaging (MOP-PCI). In MOP-PCI, we image FDT-like probes to the detector from multiple “object planes” --- like recording an event simultaneously with several cameras, some focused on nearby, others on distant, objects. To increase sensitivity, we exploit a phase-contrast imaging technique developed by Dutch Nobel laureate Fritz Zernike in the 1930s. Using MOP-PCI we recorded single-shot movies of laser pulse tracks through more than 10 cm of air. We plan to record images of meter-long tracks of electron bunches propagating through plasma in an experiment at the Stanford Linear Accelerator Center (SLAC). This will help SLAC scientists understand, optimize and scale small plasma-based particle accelerators that have applications in medicine, industry, materials science and high-energy physics.
Tabletop reflex discharges in a Penning geometry have many applications including ion sources and eXtreme Ultra-Violet (XUV) sources. The presence of primary electrons accelerated across the cathode sheaths is responsible for the distribution of ion charge states and of the unusually high XUV brightness of these plasmas. Absolutely calibrated space resolved XUV spectra from a table top reflex discharge operating with Al cathodes and Ne gas are presented. The spectra are analyzed with a new and complete model for ion charge distribution in similar reflex discharges. The plasma in the discharge was found to have a density of ∼10{sup 18 }m{sup −3} with a significant fraction >0.01 of fast primary electrons. The implications of the new model on the ion states achievable in a tabletop reflex plasma discharge are also discussed.
There are significant gaps between the observations, models, and decision support tools that make use of new data. These challenges include: 1) Decreasing the time to incorporate new satellite data into operational forecast assimilation systems, 2) Blending in-situ and satellite observing systems to produce the most accurate and comprehensive data products and assessments, 3) Accelerating the transition from research to applications through national test beds, field campaigns, and pilot demonstrations, and 4) Developing the partnerships and organizational structures to effectively transition new technology into operations. At the Short-term Prediction Research and Transition (SPORT) Center in Huntsville, Alabama, a NASA-NOAA-University collaboration has been developed to accelerate the infusion of NASA Earth science observations, data assimilation and modeling research into NWS forecast operations and decision-making. The SPoRT Center research focus is to improve forecasts through new observation capability and the regional prediction objectives of the US Weather Research Program dealing with 0-1 day forecast issues such as convective initiation and 24-hr quantitative precipitation forecasting. The near real-time availability of high-resolution experimental products of the atmosphere, land, and ocean from the Moderate Resolution Imaging Spectroradiometer (MODIS), the Advanced Infrared Spectroradiometer (AIRS), and lightning mapping systems provide an opportunity for science and algorithm risk reduction, and for application assessment prior to planned observations from the next generation of operational low Earth orbiting and geostationary Earth orbiting satellites. This paper describes the process for the transition of experimental products into forecast operations, current products undergoing assessment by forecasters, and plans for the future. The SPoRT Web page is at (http://www.ghcc.msfc.nasa.gov/sport).
With the dawn of the exascale era, computer scientists and engineers are faced with tremendous challenges across all facets of the HPC system - scalability, performance, reliability, and power consumption. In particular, the power-performance benefit from one processor generation to the next is seeing ever-diminishing returns and will require fundamental changes in the way we approach computation. In fact, it is likely that different applications will require different types of accelerators in order to meet power, performance, and reliability requirements at scale. One potential type of accelerator, a dataflow architecture, diverges from the traditional sequentially executed instruction model into one that reflects the inherent instruction-level parallelism in a program. This work presents the initial steps toward a tool that can extract the control-dataflow graph from an application.
Superconducting radiofrequency (SRF) cavities are critical components of modern particle accelerators. The industrialization of SRF technology in the U.S. has a long history, with several companies becoming interested in the technology in the past 30 years. However, there is currently no U.S. manufacturer with the capability to produce fully assembled, ready-to-test SRF cavities. A majority of these companies have either left the field or shut their doors. As a result, government-sponsored SRF accelerator projects in the U.S. largely rely on European vendors for the cavity supply. This presents not only a supply chain risk for scientific use cases, but also severely limits the commercialization of accelerator-based technologies. SRF accelerators have a range of strategic applications, including environmental remediation, semiconductor lithography, medical isotope production, and irradiation/sterilization. The nurturing of a domestic SRF industry is critical to maintaining U.S. economic competitiveness.
In the framework of the project “Novel high-energy physics studies using intense lasers and plasmas” we conducted the study of ion acceleration and “flying mirrors” with high intensity lasers in order to develop sources of ion beams and high frequency radiation for different applications. Since some schemes of laser ion acceleration are also considered a good source of “flying mirrors”, we proposed to investigate the mechanisms of “mirror” formation. As a result we were able to study the laser ion acceleration from thin foils and near critical density targets. We identified several fundamental factors limiting the acceleration in the RPA regime and proposed the target design to compensate these limitations. In the case of near critical density targets, we developed a concept for the laser driven ion source for the hadron therapy. Also we studied the mechanism of “flying mirror” generation during the intense laser interaction with thin solid density targets. As for the laser-based positron creation and capture we initially proposed to study different regimes of positron beam generation and positron beam cooling. Since the for some of these schemes a good quality electron beam is required, we studied the generation of ultra-low emittance electron beams. In order to understand the fundamental physics of high energy electron beam interaction with high intensity laser pulses, which may affect the efficient generation of positron beams, we studied the radiation reaction effects.
The Space Acceleration Measurement System (SAMS) is discussed in viewgraph form. Applications of the SAMS are as follows: (1) measurement of low-g accelerations; (2) monitoring of low-g environment; (3) monitoring of experiment-induced vibrations; and (4) validation of vibration isolation techniques. Additionally, typical locations for the SAMS are given.
Electron-Ion colliders such as the Medium energy Electron Ion Collider (MEIC) being developed by JLAB require high current electrons with low energy spread for electron cooling of the collider ring. Accelerator techniques for improving bunch charge, average current, emittance, and energy spread are required for Energy Recovery Linacs (ERLs) and Circulator Rings (CR) for next generation colliders for nuclear physics experiments. Example candidates include thermionic-cathode electron guns with RF accelerating structures. Thermionic cathodes are known to produce high currents and have excellent lifetime. The success of the IR and THz Free-Electron Laser (FEL) designed and installed by Advanced Energy Systems at the Fritz Haber Institute (FHI) of the Max Planck Society in Berlin [1,2] demonstrates that gridded thermionic cathodes and rf systems be considered for next generation collider technology. In Phase 1 Advanced Energy Systems (AES) developed and analyzed a design concept using a superconducting cavity pair and gridded thermionic cathode. Analysis included Beam Dynamics and thermal analysis to show that a design of this type is feasible. The latest design goals for the MEIC electron cooler were for electron bunches of 420 pC at a frequency of 952.6 MHz with a magnetic field on the cathode of 2kG. This field magnetizes the beam imparting angular momentum that provides for helical motion of the electrons in the cooling solenoid. The helical motion increases the interaction time and improves the cooling efficiency. A coil positioned around the cathode providing 2kG field was developed. Beam dynamics simulations were run to develop the particle dynamics near the cathode and grid. Lloyd Young added capability to Tstep to include space charge effects between two plates and include image charge effects from the grid. He also added new pepper-pot geometry capability to account for honeycomb grids. These additions were used to develop the beam dynamics for this gun. The general design is a modified ballistic compression cavity pair with two independently powered cells [3]. The first is a cathode cell that includes the thermionic cathode and grid to provide for beam bunching. The second is a full cell with independent phasing and field levels designed to minimize energy spread. The primary goal for Phase II is to manufacture a superconducting gun with a thermionic cathode and imbedded coil. The system developed here is applicable to many high current electron accelerators. The analysis and design constraints imposed by the magnetized cathode make the cathode system developed here more complicated and limited than one without the magnetized beam constraints. High power ERLs would benefit by a gun with the capabilities shown here, 400 mA or more of current. ERLs hold great promise for electron cooling experiments, advanced light sources and Free Electron Lasers. This high current electron injector is a technological advance that will place the requirements for an ERL capable of providing quality bunches needed for cooling within the MEIC circulator ring within reach. This injector would have application to future ERLs around the world.
Big data processing systems such as Spark are employed in an increasing number of diverse applications-such as machine learning, graph computation, and scientific computing-each with dynamic and different resource needs. These applications increasingly run on heterogeneous hardware, e.g., with out-of-core accelerators. However, big data platforms do not factor in the multi-dimensional heterogeneity of applications and hardware. This leads to a fundamental mismatch between the application and hardware characteristics, and the resource scheduling adopted in big data platforms. For example, Hadoop and Spark consider only data locality when assigning tasks to nodes, and typically disregard the hardware capabilities and suitability to specific application requirements. In this paper, we present RUPAM, a heterogeneity-aware task scheduling system for big data platforms, which considers both task-level resource characteristics and underlying hardware characteristics, as well as preserves data locality. RUPAM adopts a simple yet effective heuristic to decide the dominant scheduling factor (e.g., CPU, memory, or I/O), given a task in a particular stage. Our experiments show that RUPAM is able to improve the performance of representative applications by up to 62.3% compared to the standard Spark scheduler.
Combustion-based sources of shaft power tend to significantly penalize distributed propulsion concepts, but electric motors represent an opportunity to advance the use of integrated distributed propulsion on an aircraft. This enables use of propellers in nontraditional, non-thrust-centric applications, including wing lift augmentation, through propeller slipstream acceleration from distributed leading edge propellers, as well as wingtip cruise propulsors. Developing propellers for these applications challenges long-held constraints within propeller design, such as the notion of optimizing for maximum propulsive efficiency, or the use of constant-speed propellers for high-performance aircraft. This paper explores the design space of fixed-pitch propellers for use as (1) lift augmentation when distributed about a wing's leading edge, and (2) as fixed-pitch cruise propellers with significant thrust at reduced tip speeds for takeoff. A methodology is developed for evaluating the high-level trades for these types of propellers and is applied to the exploration of a NASA Distributed Electric Propulsion concept. The results show that the leading edge propellers have very high solidity and pitch well outside of the empirical database, and that the cruise propellers can be operated over a wide RPM range to ensure that thrust can still be produced at takeoff without the need for a pitch change mechanism. To minimize noise exposure to observers on the ground, both the leading edge and cruise propellers are designed for low tip-speed operation during takeoff, climb, and approach.