Computing Systems in Engineering最新文献

Large scale high speed design and analysis of next generation aerospace systems is achieved by focusing on the computational integration and synchronization of probabilistic mathematics, structural/material mechanics, and parallel computing. Design costs have been driven upward by mathematical models that require multiple levels of interactive analysis and utilize time consuming convergence criteria that further drive computing and design costs upward. To reduce CPU time and memory limitations, an effective real time parallelization (RTP) of the solution is introduced. Recursive Internal Partitioning (RIP) is used to partition the entire domain into subdomains with one or more subdomains being assigned to each independent processor. The Alpha STAR multifrontal algorithm (AMF) is implemented to assemble, condense, and solve for the unknowns at all finite element nodes. A multi level optimization technique is utilized to speed up the simulation processing time of the diversified field of specialized analysis techniques and mathematical models. These models require hierarchical multiple levels of interactive analysis utilizing time consuming convergence. The generic high speed civil transport (hsct) model is used to demonstrate the large scale computing capability. Results of multicriteria optimization indicate an order of magnitude reduction in computing time. Numerical solutions as well as physical phenomena are discussed and recommendations are provided for future solutions.

Amongst the established advantages of CAD is the ease of making changes. However, making one change usually has further implications but most existing systems are unable to assist with these changes unless an explicit link is established during the development of the design. This paper identifies different classes of change and proposes a data structure which incorporates relational operators to capture the relationships between the model components. These spatial relationships are acquired directly as the model is constructed and are used to identify the implications of a change and to suggest a possible implementation of them, based on design intent.

This paper describes the development of a knowledge-based system to assist Southern California Edison (SCE) in the design of electric service facilities for residential developments. The system supports SCE planners both by performing spatial reasoning in support of the design task and by performing more routine parts of the design. The paper also outlines the technical issues underlying how this knowledge-based designer's assistant is being integrated into the CADD environment currently being used throughout the SCE grid.

In this paper, we describe Spartan, a simple environment for developing engineering knowledge systems. The design of Spartan is motivated by the observation that many development environments incur computational overhead that buy us largely unneeded sophistication. Spartan provides only a simple set of building blocks but offers a convenient way to extend its functionalities. We offer both anecdotal and analytical support for some example applications constructed using Spartan. We conclude that Spartan can be used to construct some fairly straightforward knowledge systems. A more surprising result is that even some moderately large knowledge systems can be constructed using Spartan.

The principal objective of this work is to develop portable and extensible programming tools for the development of object-oriented parallel finite element codes for structural engineering applications. An object-oriented parallel portability interface for message-passing operations has been designed and implemented. An existing object-oriented matrix library is currently being extended to support the management of distributed matrix data and parallel solution of linear systems of algebraic equations. By taking advantage of C++ object-oriented programming, both the class libraries provide clean and consistent user interfaces, which not only help to improve the clarity and expressiveness of the client parallel codes, but also hide implementation details and complexity from the user to ease parallel programming tasks. In this paper, the object-oriented design and implementation of the class libraries are discussed. The libraries were first developed and tested using a network of Sun SPARC 10 workstations. Application examples were then studied on two commercial parallel computers: the IBM SP1 and the Intel Paragon XP/S 10, for evaluation of the portability and efficiency of the present class libraries.

A KBS for the preliminary costing of bridges is described. This system is based on the principle of heuristic substitution and provides the designer with a quick and easy way of obtaining a preliminary costing for a bridge design. The system can then be used to explore the implications of varying this design and allows easy comparison with alternative designs. By doing this, the system successfully fills a gap in the costing mechanisms currently used in practice. Preliminary trials with the system show that the speed and ease of use of the software enables designers to search for optimal solutions. It is thought that systems of this type will have a substantial impact on present conceptual design practices.

A coarse/fine mesh preconditioner is presented which can be successfully applied in a parallel finite element context. The proposed method involves the reconstruction of the stiffness equations using a coarse/fine mesh idealisation with relative degrees-of-freedom derived directly from the element shape functions. This approach leads naturally to an effective preconditioner which only requires a direct solution on coarse mesh variables and which is implemented sequentially. On the other hand, the preconditioning of the fine mesh variables involves a perfectly parallelizable diagonal scaling. The proposed derivation of the coarse/fine mesh discretization via the use of transformation matrices can be very efficient and is directly applicable to existing finite elememt solution procedures.

The simulation of instabilities and laminar-trubulent transition in high-speed boundary-layer flows represents one of the major computational challenges of the decade. By taking advantage of recent advances in computational science and instability theory for compressible flows, we have formulated an approach to this problem that combines parabolized stability equation (PSE) methodology with spatial direct numerical simulation (DNS). The relatively inexpensive PSE method is used to explore the parameter space, to compute the early (weakly and moderately nonlinear) stages of laminar breakdown, and to provide inflow conditions for the spatial DNS, which is then used to compute the highly nonlinear laminar-breakdown stage. The approach is made feasible by an accurate and efficient DNS algorithm, which has been implemented in parallel on a CRAY C90 supercomputer. The design of the DNS algorithm is discussed in detail, with emphasis on factors that affect both accuracy and efficiency. The method is applied to the investigation of the laminar breakdown of the boundary layer on an axisymmetric sharp cone in Mach 8 flow. Techniques for analysis of the resulting data are also addressed, including novel computational flow imaging procedures.

As the speed of microprocessors increases at a breath-taking rate, the gap between processor and memory system performance is getting worse. To alleviate this problem, all modern processors contain caches, but even using caches, processors cannot achieve their peak performance. We propose a mechanism, smart caching, which extends the power of conventional memory subsystems by including a prefetch unit. This prefetch unit is responsible for efficiently using the available memory bandwidth by fetching memory data before they are actually needed. Prefetching allows high-level application knowledge to increase memory performance, which is currently constraining the performance of most systems. While prefetching does not reduce the latency of memory accesses, it hides this latency by overlapping memory access and instruction execution.