Computer Execution Time Research Articles

Comparison of traffic assignments in evacuation modeling

The Nuclear Regulatory Commission (NRC) requires that all electric utilities develop and update evacuation plans for the areas surrounding their nuclear power plants. The evacuation of people around these plants has certain peculiar transportation-related characteristics, such as the location of the incident is fixed, the evacuation area is predetermined by NRC, and all the inhabitants must be evacuated from the affected area. These characteristics pose a challenge to evacuation planners and modelers. The developed evacuation models should incorporate these characteristics and address the peculiarities of the evacuation process. Hobeika et al. have developed a mass evacuation computer program (MASSVAC 3.0) that models the evacuation process and utilizes the traffic assignment of all-or-nothing and Dial's algorithm to simulate the traffic movements during evacuation. The objective of this paper is to describe the upgrade of MASS-VAC 3.0 to MASSVAC 4.0, which incorporates new modeling features such as the user equilibrium (UE) assignment algorithm, which was not available in MASSVAC 3.0. The focus of this paper, however, is on comparing the assignment results of Dial's algorithm with the UE algorithm using performance measures such as the evacuation time, the number of congested links, and the computer execution time. The results show that the evacuation performance measures are largely dependent on the highway network structure and the number of vehicles produced in an emergency planning zone, Generally, it was found that the UE algorithm was more efficient than the Dial's algorithm in the three performance measures mentioned above.

Evaluation of existing optimal VAr planning tools on utility systems

Findings of a recently completed EPRI project conducting a comprehensive evaluation of optimal VAr planning/OPF tools are reported. Three selected tools were adapted to address VAr allocation studies on four large-scale utility systems. The results obtained indicate that the use of such tools offers significant savings in cost of added MVAr requirements and the number of allocated sites with reasonable and manageable computer execution times. Based on the adaptation requirements for the different programs and studies conducted, recommendations on enhancement areas for the selected tools are provided. One of the major concerns in these studies should be the implementability of the results. This requires careful design including assignment of controls and constraints that can be implemented in the field together with the available tools to provide such requirements.

ON A MONOTONIC CONVECTION-DIFFUSION SCHEME IN ADAPTIVE MESHES

In this article we apply our recently proposed upwind model to solve the two-dimensional steady convection-diffusion equation in adaptive meshes. In an attempt to resolve high-gradient solutions in the flow, we construct finite-element spaces through use of Legendre polynomials. According to the fundamental analysis conducted in this article, we confirm that this finite-element model accommodates the monotonicity property. According to M-matrix theory, we know within what range of Peclet numbers the Petrov-Galerkin method can perform well in a sense that oscillatory solutions are not present in the flow, This monotonic region is fairly restricted, however, and limits the finite-element practioner's choices of a fairly small grid size. This limitation forbids application to practical flow simulations because monotonic solutions are prohibitively expensive to compute. Circumvention of this shortcoming is accomplished by remeshing the domain in an adaptive way. To alleviate the excessive memory requirement, our implementation incorporates a reverse Cuthill-McKee ( RCM) renumbering technique. Numerical results are presented in support of the ability of the finite-element model developed herein to resolve sharp gradients in the solution. Also shown from these numerical exercises is that considerable savings in computer storage and execution time are achieved in adaptive meshes through use of the RCM element-reordering technique.

A Simulation Approach to Evaluate the Time Cost of Parallel Computations

In this paper, the execution time cost of a parallel computation in a shared memory environment is defined as a function of the input, the algorithm, the data structure, the processor speed, the number of processors, the processing power allocation and the communication between processors. The computation structure model is modified to describe the impact of these seven factors on the computation's execution time cost. A simulation technique is developed to analyze and evaluate the computation's execution time. A software tool, TCSS (Time Cost Simulation System) that implements the simulation technique is also described.

A computer based heuristic methodology for the development of salary administration guidelines

This paper discusses the development of a heuristic methodology to generate salary merit increases and to schedule these salary actions along a specified planning horizon. The heuristic approach consists of two procedures, referred to as Heuristic Procedure 1 and Heuristic Procedure 2. For a group of executive personnel, Heuristic Procedure 1 yields salary increase percentages, in such a way that external consistency is maximized without violating internal consistency conditions. Heuristic Procedure 2 establishes the time intervals between salary increases for each eligible employee. A comparison is performed, in terms of computer execution time and optimality of solutions, between the heuristic methodology and a mathematical programming approach developed for formulating salary administration guidelines. Computational results are reported for a number of sample problems.

TRANSIENT HEAT CONDUCTION DURING QUENCHING OF DOWNWARD FACING COPPER AND STAINLESS STEEL CONVEX SURFACES

Abstract Quenching experiments were performed to determine the effect of watt material properties on pool boiling heat transfer for downward facing convex surfaces( radius of curvature 218.5 mm) in saturated water. Experiments employed 303e-type stainless steel and copper test sections having identical dimensions ( 75 mm in diameter and 20 mm thick). Pool boiling curves were constructed on the basis of two-dimensional numerical solution of transient heat conduction in spherical coordinates( r, 9) in test sections during quenching. The measured temperature histories at nine interior locations near the boiling surface ( ∼0.5 mm) provided a time-dependent boundary condition for the numerical solution. To ensure stability and reduce both computer storage and execution time, the numerical solution used the alternating direction implicit ( ADD method with control volume representations. A sensitivity analysis was conducted to assess the effect of grid size on computation time as well as the accuracy of calculat...

Solving linear diffusion‐reaction networks in porous catalyst particles using BEM

AbstractMultiple diffusion reactions are frequently encountered in the modeling of heterogeneous catalytic reactors. Obtaining an accurate estimate of the yield and selectivity in such reactions is crucial for an optimal design of reactors. Due to the inadequacy of analytical techniques in handling nonuniform catalyst shapes and mixed boundary conditions, numerical techniques are often employed to compute these design parameters. Among other numerical techniques, the boundary element method (BEM) is a superior method to solve linear diffusion reaction problems. The integral nature of the BEM formulation allows for boundary‐only discretization of the particle, thus reducing the computer execution time and the data preparation effort. A boundary element algorithm is developed to solve a network of linear diffusion reactions in porous catalyst particles in two dimensions. For this purpose, a matrix of fundamental solutions is defined and derived. The developed algorithm is applied to complex reaction networks to obtain the yield of intermediates for nonregular catalyst shapes and nonuniform boundary conditions. The method can be used as a design tool to study particle scale modeling in detail and can be incorporated into an overall reactor model.

Vibratory Behavior of Doubly Curved Shallow Shells of Curvilinear Planform

A continuum approach based on the principle of minimum total energy is presented to study the free flexural vibration of doubly curved shallow shells of curvilinear planform. The Ritz procedure supplemented with the pb -2 shape functions is employed to seek for an effective numerical algorithm, which is capable of providing computational solutions to a great accuracy. This p -version Ritz method is easily implemented numerically, is versatile in accounting for various boundary conditions, and requires less computing memory and execution time because no domain discretization is involved. The doubly curved shallow shells under consideration spread over a wide class of positive and negative Gaussian curvatures (spherical and hyperbolic paraboloidal shells) and the boundary could be free, simply supported, or clamped. The accuracy of numerical results has been established through a convergence and comparison study of eigenvalues with available published data. Extensive benchmark frequency parameters covering ...

Space-Time Spectral Element Methods for One-Dimensional Nonlinear Advection-Diffusion Problems

The following space-time Galerkin spectral element methods are developed and applied to solve the Burgers equation with small viscosity: (a) coupled methods, consisting of an explicit method for hyperbolic dominated equations and an implicit method for parabolic dominated equations; (b) two splitting methods which solve the hyperbolic substep explicitly and the parabolic one implicitly (one uses spectral elements in the explicit part and the other uses the Adams-Bashforth multistep method). A subcycling technique, in which several convective steps are taken for each implicit viscous step was also investigated for the two splitting methods. A stability analysis of the four methods is performed and subsequent results are debated. A convergence study and a comparison of computer execution time for the four methods is made and the results are discussed. Comparative study leads to the conclusion that the space-time spectral element splitting method with subcycling is superior to the other methods presented in terms of robustness and computer execution time. The number of subcycles should be kept low (2-3) in order to avoid significant loss of accuracy. The coupled explicit method is also applied to the solution of the one-dimensional coupled continuity, momentum, and energy equations for non-isothermal flow of an ideal gas with temperature dependent properties in a cylindrical duct of variable radius.

The practicability of using Richards' equation for general purpose soil-water dynamics models

Use of finite difference solutions of Richards' equation is generally considered to be impracticable for general purpose models of soil-water dynamics, because numerical performance is not predictable with unfamiliar parameter values, and because of excessive computer execution times. The Broadbridge-White model of soil hydraulic properties yields finite soil-water diffusivity, so that solutions of the finite difference equations always exist. Further, it permits the flow equation to be scaled to realistic soils in terms of three independent variables. Searching a practical three-dimensional parameter space yields criteria for numerical stability and for guaranteed convergence of the iterative procedure for obtaining the solution at each time step. An efficient numerical scheme yields soil-water models with practical execution times. Comparison shows higher computational speed and greater model simplicity, relative to an alternative, less rigorous, model based on generalising the sharp wetting front infiltration model of Green and Ampt.

Efficient algorithm for linear problems using packed data structure with a set of parallel pointers

Nowadays the algorithms developed impose the same restrictions as many years before such as storage and execution time. In this paper an algorithm is presented which is efficient as regards to computer execution time and storage requirements.

Description and Evaluation of Program USM

USM is a hydraulic‐simulation computer program developed by the U.S. Bureau of Reclamation to analyze unsteady‐flow conditions in canals. The program USM uses the method of characteristics to solve the complete Saint Venant equations for unsteady flow in open channels. This solution method yields accurate numerical results but sometimes requires more computer execution time than other methods. USM analyzes canal systems with a variety of structural boundary conditions, including check gates, weirs, siphons, and pumps. Check gates can have either manual or automatic gate control. Branching canals, supercritical flow, and channel dewatering cannot be analyzed. This paper describes the model and compares its strengths and weaknesses with those of other models. Several test data sets compiled by the ASCE Task Committee on Irrigation Canal System Hydraulic Modeling were used in the evaluation. USM simulation results are compared to field data collected from the Central Arizona Project canal system.

Sparse matrices in astronomical data reduction

Sparse matrices occur frequently in astronomical data adjustment problems. Use of special techniques that take advantage of the sparcity structure can result in substantial saving of computer memory and execution time. In astrometry, for example, when one uses minor planets, or other objects, to determine parameters common to all of the observations, memory and execution time are saved by storing only the nonzero elements of the matrix in a vector and deriving an index function to locate them uniquely. Although the indexing of elements and the solution of the linear system become complicated, the saving of memory compared with an upper triangular matrix is 1-36p(p−1)/n(n+1), where p is the number of minor planets and n the number of unknowns

KALMOD, a stochastic-deterministic model for simulating groundwater flow with Kalman filtering

This paper presents the program KALMOD which has been developed to enable the execution of the integration of the Kalman filtering and the numerical groundwater flow model MODFLOW on microcomputers. The program can be applied to quantify and reduce the uncertainty of the groundwater flow model, and to analyse and design groundwater monitoring networks. KALMOD consists of a preprocessor, a processor and a postprocessor. The preprocessor acts as an interface between the user and the processor. The processor manipulates the measurement processes and carries out the filtering tasks. The filtering algorithm is implemented so that it is relatively efficient with respect to computer memory and execution time. The postprocessor was designed to present the model results in graphics. The program is suitable for small scale problems and for educational purposes.

Induction motor model for the study of transient stability in both balanced and unbalanced multi-machine networks

An induction motor model is presented that represents the motor as a constant equivalent impedance in series with a variable voltage source that is a function of slip. The model is based on that presented by M.A. Lahoud et al. (see Electr. Power Syst. Res., vol.7, no.1, p.53-63, 1984) for balanced systems using the full nonlinear two-axis equations, and it is adapted for use in unbalanced networks with phase-frame network representation. This model overcomes the need to update the (Y/sub bus/) matrix at every integration step and significantly reduces the computer execution time. In addition, the model takes into account the time rate of change of rotor flux linkages. Results show a reduction in computation time of between 31% for a 10-bus system and 73% for a 15-bus system.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>

Numerical Model of Longshore Current for Bar and Trough Beaches

This paper describes a numerical modeling system called NMLONG that was developed to calculate the wave height, wave direction, mean water level, and steady‐state time‐averaged longshore current velocity across a multiple bar and trough beach profile. The modeling system includes the longshore current and change in mean water surface elevation produced by waves and wind, lateral mixing, wave driving by monochromatic or random waves as options, linear or quadratic bottom friction law as options, and wave‐current interaction as an option. The principal requirement for use of the model is longshore uniformity in bathymetry and waves. Verification against laboratory and field measurements of the wave height, longshore current, and mean water surface elevation (laboratory only) indicates that default values of the model calibration parameters should be acceptable for most engineering applications. Techniques are employed to minimize computer execution time, including efficient time averaging of the nonlinear bottom friction stress and a skillful first approximation in the iterative solution of the longshore current using the nonlinear bottom friction law, allowing the model to run effectively on a desktop computer.

放射性物質輸送容器の衝突解析用簡易計算コードの開発，（Ｉ） 緩衝体付き輸送容器衝突計算コードＣＲＵＳＨ

In impact analysis for radioactive materials transport casks, it has become possible to perform them in detail by using finite element and finite difference computer codes, such as HONDO-II, DYNA2D, DYNA3D and PISCES. However, the considerable cost and computer execution time are necessary to perform analyses by these computer codes. Therefore, simplified computer codes which are capable of reducing cost and computer execution time are needed. To meet the above requirements, CRUSH code has been developed. CRUSH is a static calculation computer code capable evaluating the maximum acceleration of the cask body and the maximum displacement of the shock absorber.Analytical results obtained by CRUSH code are compared with experimental results and the detailed computer code DYNA3D. Good agreements were obtained between simplified results and experimental and detailed analysis ones.

Comparison of computer codes for seismic analysis of dams

An extensive set of computer studies has been made using the EAGD, SAP IV, and ADINA computer codes. Dam-reservoir interaction has been incorporated by two different methods: fluid element and discretized lumped water mass. Parametric studies have been made to investigate the effect of (a) water compressibility vs incompressibility, (b) flexible vs rigid foundation, (c) empty vs full reservoir, and (d) numerical technique on central processor execution time. It has been found that the alpha factor (reservoir absorption coefficient) used in the EAGD program has an influence on the individual modal response of the dam. However, the overall deflections and stresses are not very sensitive to the variation in the value of this alpha parameter. The deflections and stresses obtained from EAGD, which uses finite element simulation for the water reservoir, are compared with SAP IV and ADINA results with discretized lumped water masses. Seismic behavior of the Pine Flat Dam subjected to the Taft earthquake shows that the results of the compressible water model are almost the same as those of the incompressible water model. Lumped added water mass results are about twice as large as results obtained using the fluid element model. A summary of computer execution time is also provided for comparison of relative merits of various numerical techniques. It is noted that EAGD uses a Fast Fourier Transform algorithm whereas SAP IV and ADINA use time domain methods. Execution time for ADINA is three times longer than that for the SAP IV program.

A new solution for mechanisms including Coulomb friction

This paper presents a new solution method for the analysis of multibody systems with Coulomb friction. Complete equations of motion and reaction forces are derived by using Lagrangian formulation and the traditional friction circle concepts for the analysis of Coulomb friction. The numerical solutions by the new method, as well as the conventional method for comparision, are illustrated. The new method saves considerably the computer execution time to solve equations of motion and reaction forces compared with the conventional method for same accuracy. The higher the coefficient of friction is, the more the computation time of the conventional method is needed but the computation time of the new method is nearly independent on the coefficient of friction.

Predicting performance of parallel computations

An accurate and computationally efficient method for predicting the performance of a class of parallel computations running on concurrent systems is described. A parallel computation is modeled as a task system with precedence relationships expressed as a series-parallel directed acyclic graph. Resources in a concurrent system are modeled as service centers in a queuing network model. Using these two models as inputs, the method outputs predictions of expected execution time of the parallel computation and the concurrent system utilization. The method is validated against both detailed simulation and actual execution on a commercial multiprocessor. Using 100 test cases, the average error of the prediction when compared to simulation statistics is 1.7%, with a standard deviation of 1.5%; the maximum error is about 10%. >