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.
An apparatus is disclosed for reducing acoustic transmission from mechanical or acoustic sources by means of a double wall partition, within which an acoustic pressure field is generated by at least one secondary acoustic source. The secondary acoustic source is advantageously placed within the partition, around its edges, or it may be an integral part of a wall of the partition.
An important problem facing numerous research projects on parallelizing compilers for distributed memory machines is that of automatically determining a suitable data partitioning scheme for a program. Any strategy for automatic data partitioning needs a mechanism for estimating the performance of a program under a given partitioning scheme, the most crucial part of which involves determining the communication costs incurred by the program. A methodology is described for estimating the communication costs at compile-time as functions of the numbers of processors over which various arrays are distributed. A strategy is described along with its theoretical basis, for making program transformations that expose opportunities for combining of messages, leading to considerable savings in the communication costs. For certain loops with regular dependences, the compiler can detect the possibility of pipelining, and thus estimate communication costs more accurately than it could otherwise. These results are of great significance to any parallelization system supporting numeric applications on multicomputers. In particular, they lay down a framework for effective synthesis of communication on multicomputers from sequential program references.
Some new codes are presented which have good performance on Rician fading channel with small decoding complexities. A new M-way partition chain is proposed for the L x MPSK (L less than or equal to M) signal set which maximizes the intra-set distance of each subset at each partition level. Based on this partition chain, a class of asymptotical optimum codes was found. For M = 4, these codes have both large symbol distances and product distances. Multi-level coding scheme allows to construct a code by hand such that the code meets some desired parameters, e.g., symbol distance, product distance, etc. In design of a multi-level code, all factors were considered which affect the performance and complexity of the code, such as, the decoding scheme, decoding complexity, and performance under the decoding scheme, e.g., if the multi-stage decoding scheme is used, the performance degradation due to the suboptimum decoding is taken into consideration. The performance for most of the presented codes was simulated on Rayleigh fading channel, and the results show that these codes have good performance with small decoding complexities.
The present study uses plagioclase compositions and partition coefficients for several elements to estimate melts from which lunar anorthosites formed and the partition liquids from which these melts were derived. The Fe and Mg contents of the plagioclases are examined for their potential use as predictors of Fe and Mg in associated melts. Usable trace elements are employed with selected partition coefficients to predict the equilibrium melts for lunar anorthosites, and melts for anorthosites are used in various models to calculate the composition of the initial melt from which the lunar mantle and eventually the anorthosites evolved. The initial melts are compared with bulk moon compositions to evaluate the nature of the melting that could have produced the anorthosites. All the models indicate that extensive melting of all or nearly all of the lunar mantle is required to produce a lunar crust that constitutes 13-15 percent of the moon's volume.
The problem of Cholesky factorization of a sparse matrix has been very well investigated on sequential machines. A number of efficient codes exist for factorizing large unstructured sparse matrices. However, there is a lack of such efficient codes on parallel machines in general, and distributed machines in particular. Some of the issues that are critical to the implementation of sparse Cholesky factorization on a distributed memory parallel machine are ordering, partitioning and mapping, load balancing, and ordering of various tasks within a processor. Here, we focus on the effect of various partitioning schemes on the performance of sparse Cholesky factorization on the Intel iPSC/860. Also, a new partitioning heuristic for structured as well as unstructured sparse matrices is proposed, and its performance is compared with other schemes.
The studies of the proposed Ada 9X constructs for distribution, now referred to as AdaPT are reported. The goals for this time period were to revise the chosen example scenario and to begin studying about how the proposed constructs might be implemented. The example scenario chosen is the Submarine Combat Information Center (CIC) developed by IBM for the Navy. The specification provided by IBM was preliminary and had several deficiencies. To address these problems, some changes to the scenario specification were made. Some of the more important changes include: (1) addition of a system database management function; (2) addition of a fourth processing unit to the standard resources; (3) addition of an operator console interface function; and (4) removal of the time synchronization function. To implement the CIC scenario in AdaPT, the decided strategy were publics, partitions, and nodes. The principle purpose for implementing the CIC scenario was to demonstrate how the AdaPT constructs interact with the program structure. While considering ways that the AdaPt constructs might be translated to Ada 83, it was observed that the partition construct could reasonably be modeled as an abstract data type. Although this gives a useful method of modeling partitions, it does not at all address the configuration aspects on the node construct.
Clustering methods have been used extensively in computer vision and pattern recognition. Fuzzy clustering has been shown to be advantageous over crisp (or traditional) clustering in that total commitment of a vector to a given class is not required at each iteration. Recently fuzzy clustering methods have shown spectacular ability to detect not only hypervolume clusters, but also clusters which are actually 'thin shells', i.e., curves and surfaces. Most analytic fuzzy clustering approaches are derived from Bezdek's Fuzzy C-Means (FCM) algorithm. The FCM uses the probabilistic constraint that the memberships of a data point across classes sum to one. This constraint was used to generate the membership update equations for an iterative algorithm. Unfortunately, the memberships resulting from FCM and its derivatives do not correspond to the intuitive concept of degree of belonging, and moreover, the algorithms have considerable trouble in noisy environments. Recently, we cast the clustering problem into the framework of possibility theory. Our approach was radically different from the existing clustering methods in that the resulting partition of the data can be interpreted as a possibilistic partition, and the membership values may be interpreted as degrees of possibility of the points belonging to the classes. We constructed an appropriate objective function whose minimum will characterize a good possibilistic partition of the data, and we derived the membership and prototype update equations from necessary conditions for minimization of our criterion function. In this paper, we show the ability of this approach to detect linear and quartic curves in the presence of considerable noise.
Clustering methods have been used extensively in computer vision and pattern recognition. Fuzzy clustering has been shown to be advantageous over crisp (or traditional) clustering in that total commitment of a vector to a given class is not required at each iteration. Recently fuzzy clustering methods have shown spectacular ability to detect not only hypervolume clusters, but also clusters which are actually 'thin shells', i.e., curves and surfaces. Most analytic fuzzy clustering approaches are derived from Bezdek's Fuzzy C-Means (FCM) algorithm. The FCM uses the probabilistic constraint that the memberships of a data point across classes sum to one. This constraint was used to generate the membership update equations for an iterative algorithm. Unfortunately, the memberships resulting from FCM and its derivatives do not correspond to the intuitive concept of degree of belonging, and moreover, the algorithms have considerable trouble in noisy environments. Recently, the clustering problem was cast into the framework of possibility theory. Our approach was radically different from the existing clustering methods in that the resulting partition of the data can be interpreted as a possibilistic partition, and the membership values may be interpreted as degrees of possibility of the points belonging to the classes. An appropriate objective function whose minimum will characterize a good possibilistic partition of the data was constructed, and the membership and prototype update equations from necessary conditions for minimization of our criterion function were derived. The ability of this approach to detect linear and quartic curves in the presence of considerable noise is shown.
Binary dissection is widely used to partition non-uniform domains over parallel computers. This algorithm does not consider the perimeter, surface area, or aspect ratio of the regions being generated and can yield decompositions that have poor communication to computation ratio. Parametric Binary Dissection (PBD) is a new algorithm in which each cut is chosen to minimize load + lambda x(shape). In a 2 (or 3) dimensional problem, load is the amount of computation to be performed in a subregion and shape could refer to the perimeter (respectively surface) of that subregion. Shape is a measure of communication overhead and the parameter permits us to trade off load imbalance against communication overhead. When A is zero, the algorithm reduces to plain binary dissection. This algorithm can be used to partition graphs embedded in 2 or 3-d. Load is the number of nodes in a subregion, shape the number of edges that leave that subregion, and lambda the ratio of time to communicate over an edge to the time to compute at a node. An algorithm is presented that finds the depth d parametric dissection of an embedded graph with n vertices and e edges in O(max(n log n, de)) time, which is an improvement over the O(dn log n) time of plain binary dissection. Parallel versions of this algorithm are also presented; the best of these requires O((n/p) log(sup 3)p) time on a p processor hypercube, assuming graphs of bounded degree. How PBD is applied to 3-d unstructured meshes and yields partitions that are better than those obtained by plain dissection is described. Its application to the color image quantization problem is also discussed, in which samples in a high-resolution color space are mapped onto a lower resolution space in a way that minimizes the color error.
A mesh-vertex finite volume scheme for solving the Euler equations on triangular unstructured meshes is implemented on an MIMD (multiple instruction/multiple data stream) parallel computer. An explicit four-stage Runge-Kutta scheme is used to solve two-dimensional flow problems. A family of implicit schemes is also developed to solve these problems, where the linear system that arises at each time step is solved by a preconditioned GMRES algorithm. Two partitioning strategies are employed, one that partitions triangles and the other that partitions vertices. The choice of the preconditioner in a distributed memory setting is discussed. All the methods are compared both in terms of elapsed times and convergence rates. It is shown that the implicit schemes offer adequate parallelism at the expense of minimal sequential overhead. The use of a global coarse grid to further minimize this overhead is also investigated. The schemes are implemented on a distributed memory parallel computer, the iPSC/860.
Applications are described of high-performance parallel, computation for the analysis of complete jet engines, considering its multi-discipline coupled problem. The coupled problem involves interaction of structures with gas dynamics, heat conduction and heat transfer in aircraft engines. The methodology issues addressed include: consistent discrete formulation of coupled problems with emphasis on coupling phenomena; effect of partitioning strategies, augmentation and temporal solution procedures; sensitivity of response to problem parameters; and methods for interfacing multiscale discretizations in different single fields. The computer implementation issues addressed include: parallel treatment of coupled systems; domain decomposition and mesh partitioning strategies; data representation in object-oriented form and mapping to hardware driven representation, and tradeoff studies between partitioning schemes and fully coupled treatment.
Current object-oriented analysis and design methodologies fall short in their use of mechanisms for identifying threads of control for the system being developed. The scenarios which typically describe a system are more global than looking at the individual objects and representing their behavior. Unlike conventional methodologies that use data flow and process-dependency diagrams, object-oriented methodologies do not provide a model for representing these global threads end-to-end. Tracing through threads of control is key to ensuring that a system is complete and timing constraints are addressed. The existence of multiple threads of control in a system necessitates a partitioning of the system into processes. This paper describes the application and representation of end-to-end threads of control to the object-oriented analysis and design process using object-oriented constructs. The issue of representation is viewed as a grouping problem, that is, how to group classes/objects at a higher level of abstraction so that the system may be viewed as a whole with both classes/objects and their associated dynamic behavior. Existing object-oriented development methodology techniques are extended by adding design-level constructs termed logical composite classes and process composite classes. Logical composite classes are design-level classes which group classes/objects both logically and by thread of control information. Process composite classes further refine the logical composite class groupings by using process partitioning criteria to produce optimum concurrent execution results. The goal of these design-level constructs is to ultimately provide the basis for a mechanism that can support the creation of process composite classes in an automated way. Using an automated mechanism makes it easier to partition a system into concurrently executing elements that can be run in parallel on multiple processors.
This paper presents a taxonomy of problems that must be solved in order to achieve on-line upgradability of long-lived programs, and presents a solution to the fundamental problems in the taxonomy. The solutions are based upon AdaPT, a set of language extensions designed to aid in the distribution of a single Ada program. AdaPT introduces three major units, the public, the partition, and the node. Publics are primarily used to share type information. Partitions are the basic units of distribution while nodes are used to control the configuration of the program. Nodes and partitions can be created dynamically via the allocator. A node-level routine controls the replacement process. The controlling node creates a new instance of the routine being replaced; the run-time system must ensure that the new instance is the updated one. Once access to the new version has been established, all further calls to the module are redirected to the new version and the caller is informed of the change so that it may make subsequent calls directly. When a module is being upgraded, there is a transition period during which both the old and new versions are present. We require that clients of a potentially replaceable module 'check in' with the controlling node. The controlling node keeps track of the number of clients that have been redirected to the replacement. When all have been redirected, the old version can be deallocated.
The work done on this grant primarily concerns the measurement of aerosol in the stratosphere from NASA ER-2 aircraft in studies of stratospheric ozone depletion in the northern and southern hemispheres. The ER-2 Condensation Nucleus Counter (CNC) measures the number concentration of particles in the diameter range of approximately 0.01 to 1 micron. The Passive Cavity Aerosol Spectrometer measures size distributions in the 0.17 to 3 micron diameter range. This instrument was upgraded during this grant period to a Focused Cavity Aerosol Spectrometer (FCAS). This upgrade permitted the instrument to measure particles as small as 0.05 micron in diameter. The inlet for the PCAS and FCAS was modified and characterized under this grant so that the modifications to the aerosol due to anisokinetic sampling and heating upon sampling and in transport to the measurement location were accounted for in the data analysis. These measurements permitted observations of particle production in the southern hemisphere winter polar vortex and observation of the impact of denitrification on the number concentration of the aerosol in the denitrified air. In the northern polar vortex, the measurements provided a characterization of the sulfate aerosol. Following the eruption of Mount Pinatubo in 1991, the measurements permitted an accurate characterization of the sulfate aerosol enhancements resulting from the eruption. This led to studies of the impact of heterogeneous chemistry on the partitioning of the partitioning of the reactive nitrogen species and the partitioning of the chlorine reservoir.
The numerical solution procedure for multibody dynamics (MBD) systems is termed a staggered MBD solution procedure that solves the generalized coordinates in a separate module from that for the constraint force. This requires a reformulation of the constraint conditions so that the constraint forces can also be integrated in time. A major advantage of such a partitioned solution procedure is that additional analysis capabilities such as active controller and design optimization modules can be easily interfaced without embedding them into a monolithic program. After introducing the basic equations of motion for MBD system in the second section, Section 3 briefly reviews some constraint handling techniques and introduces the staggered stabilized technique for the solution of the constraint forces as independent variables. The numerical direct time integration of the equations of motion is described in Section 4. As accurate damping treatment is important for the dynamics of space structures, we have employed the central difference method and the mid-point form of the trapezoidal rule since they engender no numerical damping. This is in contrast to the current practice in dynamic simulations of ground vehicles by employing a set of backward difference formulas. First, the equations of motion are partitioned according to the translational and the rotational coordinates. This sets the stage for an efficient treatment of the rotational motions via the singularity-free Euler parameters. The resulting partitioned equations of motion are then integrated via a two-stage explicit stabilized algorithm for updating both the translational coordinates and angular velocities. Once the angular velocities are obtained, the angular orientations are updated via the mid-point implicit formula employing the Euler parameters. When the two algorithms, namely, the two-stage explicit algorithm for the generalized coordinates and the implicit staggered procedure for the constraint Lagrange multipliers, are brought together in a staggered manner, they constitute a staggered explicit-implicit procedure which is summarized in Section 5. Section 6 presents some example problems and discussions concerning several salient features of the staggered MBD solution procedure are offered in Section 7.
A mesh-vertex finite volume scheme for solving the Euler equations on triangular unstructured meshes is implemented on a multiple-instruction/multiple-data stream parallel computer. An explicit four-stage Runge-Kutta scheme is used to solve two-dimensional flow problems. A family of implicit schemes is also developed to solve these problems, where the linear system that arises at each time step is solved by a preconditioned GMRES algorithm. Two partitioning strategies are employed: one that partitions triangles and the other that partitions vertices. The choice of the preconditioner in a distributed memory setting is discussed. All of the methods are compared both in terms of elapsed times and convergence rates. It is shown that the implicit schemes offer adequate parallelism at the expense of minimal sequential overhead. The use of a global coarse grid to further minimize this overhead is also investigated. The schemes are implemented on a distributed memory parallel computer, the Intel iPSC/860.
Distributed-memory parallel computers dominate today's parallel computing arena. These machines, such as Intel Paragon, IBM SP2, and Cray Origin2OO, have successfully delivered high performance computing power for solving some of the so-called "grand-challenge" problems. Despite initial success, parallel machines have not been widely accepted in production engineering environments due to the complexity of parallel programming. On a parallel computing system, a task has to be partitioned and distributed appropriately among processors to reduce communication cost and to attain load balance. More importantly, even with careful partitioning and mapping, the performance of an algorithm may still be unsatisfactory, since conventional sequential algorithms may be serial in nature and may not be implemented efficiently on parallel machines. In many cases, new algorithms have to be introduced to increase parallel performance. In order to achieve optimal performance, in addition to partitioning and mapping, a careful performance study should be conducted for a given application to find a good algorithm-machine combination. This process, however, is usually painful and elusive. The goal of this project is to design and develop efficient parallel algorithms for highly accurate Computational Fluid Dynamics (CFD) simulations and other engineering applications. The work plan is 1) developing highly accurate parallel numerical algorithms, 2) conduct preliminary testing to verify the effectiveness and potential of these algorithms, 3) incorporate newly developed algorithms into actual simulation packages. The work plan has well achieved. Two highly accurate, efficient Poisson solvers have been developed and tested based on two different approaches: (1) Adopting a mathematical geometry which has a better capacity to describe the fluid, (2) Using compact scheme to gain high order accuracy in numerical discretization. The previously developed Parallel Diagonal Dominant (PDD) algorithm and Reduced Parallel Diagonal Dominant (RPDD) algorithm have been carefully studied on different parallel platforms for different applications, and a NASA simulation code developed by Man M. Rai and his colleagues has been parallelized and implemented based on data dependency analysis. These achievements are addressed in detail in the paper.
The Tropospheric Trace Gas and Airborne Measurements Group's (TTGAMG) efforts on NASA GTE (Global Tropospheric Experiment) PEM (Pacific Exploratory Mission) West A & B field campaign primarily involved the acquisition of NO, NO2 and NO(y) measurements, as well as the subsequent analysis and interpretation of the data base obtained during the PEM West field campaign. These investigations focused on the distribution of trace gases, sources and sinks of ozone, ozone producing precursors with a heavy emphasize on ozone's photochemical state, and the partitioning of the molecules within the NO(y) family over the north western Pacific Ocean. The two components of PEM West were focused on observing air masses as they reached the Asian Continent (PEM West A) or as the air mass departed the Asian Continent (PEM West B). NO(x) concentrations play a pivotal role in controlling the photochemical lifetime of ozone in these environments, and understanding the NO(x) species partitioning is paramount. The transport of NO(x) into the regions, in the form of longer lived NO(y) family members, was examined in relation to the comparison of natural occurring sources of NO(x) (i.e., lightning and stratosphere/troposphere exchange) to those produced as a result of anthropogenic activity (i.e., biomass burning and aircraft emissions). The TTGAMG's measurements of NOx and NO(y), in conjunction with other investigators' measurements of PAN (H. B. Singh's group) and HNO3 (R. W. Talbot's group), have been used to assess the total reactive odd nitrogen levels over the study regions, the partitioning of the reactive odd nitrogen species in their various forms, and the usefulness of the NO, measurement and its measurement technique. The TTGAMG's primary PEM West objectives were the characterization of the factors controlling the distribution and fate of reactive odd nitrogen compounds over the western Pacific Ocean and an analysis of the concentration of various trace gases in the troposphere as the air mass aged by both dynamical mixing and photochemical processes in the troposphere.