SEARCH · Engineering Papers
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.
The goal is to develop an architecture for parallel processors enabling optimal handling of multi-disciplinary computation of fluid-solid simulations employing finite element and difference schemes. The goals, philosphical and modeling directions, static and dynamic poly trees, example problems, interpolative reduction, the impact on solvers are shown in viewgraph form.
The applicability of a parallel architecture to the solution of large electromagnetic scattering problems is demonstrated. Two techniques, finite difference and the method of moments, are used to provide insight into the comparative speedups which can be attained for several algorithms. The flexibility of the hypercube architecture for different analysis algorithms is illustrated.
An attempt is made to determine whether or not the structure of supernova remnant shells is dependent on the angle between the line of sight and the Galactic magnetic field and if this structure can be explained either by quasi-parallel or quasi-perpendicular acceleration. The method used to determine the line of sight angles to the magnetic field is described. A model for the emissivity of a supernova remnant is developed.
An explicit flow solver, applicable to the hierarchy of model equations ranging from Euler to full Navier-Stokes, is combined with several techniques designed to reduce computational expense. The computational domain consists of local grid refinements embedded in a global coarse mesh, where the locations of these refinements are defined by the physics of the flow. Flow characteristics are also used to determine which set of model equations is appropriate for solution in each region, thereby reducing not only the number of grid points at which the solution must be obtained, but also the computational effort required to get that solution. Acceleration to steady-state is achieved by applying multigrid on each of the subgrids, regardless of the particular model equations being solved. Since each of these components is explicit, advantage can readily be taken of the vector- and parallel-processing capabilities of machines such as the Cray X-MP and Cray-2.
ISEE-1 and-2 plasma and magnetic field observations are used to examine several encounters with brief magnetic pulsations in the quasi-parallel region of the earth's bow shock. The two-spacecraft timing of the magnetic field signature is examined to see if the pulsation encounters are nested (as for back and forth motion of the shock over the spacecraft) or simply time-shifted (as for a structure convected anti-sunward across the spacecraft). Examples of both types of signatures are found, some within minutes of each other, suggesting that at least some pulsations probably originate from the growth and steepening of upstream waves but may eventually be able to stand in the upstream flow, perhaps becoming part of the shock surface itself. The possibility that some of the pulsations may simply be brief encounters with the moving shock is also not ruled out.
The execution overhead inherent in the conversation scheme, which is a scheme for realizing fault-tolerant cooperating processes free of the domino effect, is analyzed. Multiprocessor/multicomputer systems capable of parallel execution of conversation components are considered and a queuing network model of such systems is adopted. Based on the queuing model, various performance indicators, including system throughput, average number of processors idling inside a conversation due to the synchronization required, and average time spent in the conversation, have been evaluated numerically for several application environments. The numeric results are discussed and several essential performance characteristics of the conversation scheme are derived. For example, when the number of participant processes is not large, say less than six, the system performance is highly affected by the synchronization required on the processes in a conversation, and not so much by the probability of acceptance-test failure.
Large scale one-dimensional hybrid simulations with resistive electrons have been carried out of a quasi-parallel high-Mach-number collisionless shock. The shock initially appears stable, but then exhibits cyclic behavior. For the magnetic field, the cycle consists of a period when the transition from upstream to downstream is steep and well defined, followed by a period when the shock transition is extended and perturbed. This cyclic shock solution results from upstream perturbations caused by backstreaming gyrating ions convecting into the shock. The cyclic reformation of a sharp shock transition can allow ions, at one time upstream because of reflection or leakage, to contribute to the shock thermalization.
The present examination of radiative heat transfer through fibrous materials gives attention to the interactions between radiation scattered from individual parallel fibers for the case of a normally-incident plane EM wave and different representative geometries. Experimental results are obtained for the specific case corresponding to fibers in one plane. Analytical models are developed for the radiative scattering behavior of fibrous media when the interference mechanism is accounted for; the results obtained indicate that interference decreases the scattering efficiency of fibrous media containing a large number of randomly positioned fibers.
Techniques for the enhancement of mixing in a scramjet combustor are presently investigated experimentally with wall-mounted parallel injector ramps that are speculated to prove useful at high speeds in the extraction of thrust from engine- or airframe-coolant hydrogen. The injector ramps were designed to generate a reflected shock wave from the duct top wall which will pass just downstream of the fuel injectors' barrel shock; fuel injector Mach number is 1.7, in virtue of which an underexpanded fuel flow is generated. Shadowgraph and UV-TV flow visualizations were conducted for three duct configurations.
Consideration is given to the suggestion by Basu and Coppi (1988, 1989) that the DE-2 observations of broadband turbulence associated with sheared ion flows can be explained by electrostatic waves driven by a sheared ion flow along the magnetic field. It is pointed out that such a theory ignores the stronger shear in the ion flow transverse to the magnetic field, and that, when this shear is taken into account, the modes described by Basu and Coppi are easily destabilized. The theory of Basu and Coppi is shown to break down even when the shear in the parallel flow exceeds the shear in the transverse flow.
High-frequency-link power conversion and distribution based on a resonant inverter (RI) has been recently proposed. The design of several topologies is reviewed, and a simple approximate design procedure is developed for the phase-controlled parallel-loaded RI. This design procedure seeks to ensure the benefits of resonant conversion and is verified by data from a laboratory 2.5 kVA, 20-kHz converter. A simple phasor analysis is introduced as a useful approximation for design purposes. The load is considered to be a linear impedance (or an ac current sink). The design procedure is verified using a 2.5-kVA 20-kHz RI. Also obtained are predictable worst-case ratings for each component of the resonant tank circuit and the inverter switches. For a given load VA requirement, below-resonance operation is found to result in a significantly lower tank VA requirement. Under transient conditions such as load short-circuit, a reversal of the expected commutation sequence is possible.
In the present automated system for the organization of the data and computational operations entailed by parallel problems, in ways that optimize multiprocessor performance, general heuristics for partitioning program data and control are implemented by capturing and manipulating representations of a computation at run time. These heuristics are directed toward the dynamic identification and allocation of concurrent work in computations with irregular computational patterns. An optimized static-workload partitioning is computed for such repetitive-computation pattern problems as the iterative ones employed in scientific computation.
A programming methodology is introduced that utilizes computational synchronization and avoids tight control flow synchronization in parallel programs. In this methodology, each phase of the computation is assigned a status that can be ready, blocked, or completed, and tasks in each computational phase are self-scheduled to ensure computational progress by the available executing processes. Results indicate that this methodology avoids the catastrophic performance losses resulting from the swapping of processes in multiprogrammed multiprocessors.
Parallel systolic algorithms for dynamic programming(DP) and their respective hardware implementations are presented for a problem in on-line trajectory optimization. The method is applied to a model for helicopter flight path optimization through a complex constraint region. This problem has application to an air traffic control problem and also to a terrain following/threat avoidance problem.
Momentarily-best compression algorithm selected. Proposed competitive-parallel-processing system compresses data for transmission in channel of limited band-width. Likely application for compression lies in high-resolution, stereoscopic color-television broadcasting. Data from information-rich source like color-television camera compressed by several processors, each operating with different algorithm. Referee processor selects momentarily-best compressed output.
Elements assigned approximately optimally to subdomains. Mapping algorithm based on simulated-annealing concept used to minimize approximate time required to perform finite-element computation on hypercube computer or other network of parallel data processors. Mapping algorithm needed when shape of domain complicated or otherwise not obvious what allocation of elements to subdomains minimizes cost of computation.
Heuristic algorithm minimizes amount of memory used by multiprocessor system. Distributes load of many identical, short computations among multiple parallel digital data processors, each of which has its own (local) memory. Each processor operates on distinct and independent set of data in larger shared memory. As integral part of load-balancing scheme, total amount of space used in shared memory minimized. Possible applications include artificial neural networks or image processors for which "pipeline" and vector methods of load balancing inappropriate.
Simple configuration for wiring of multiple parallel-connected switching transistors minimizes stray wiring inductance while providing for use of balancing transformers, which equalize currents in transistors. Currents balanced on twisted pairs of wires. Because twisted pairs carry both "hot-side" and return currents, this configuration has relatively low inductance.