Simulation of Large and Complex Structures With Multiple Contacts and Wear

This paper deals with the simulation methodology of large and complex structures with multiple contacts and wear. The methodology developed is used to evaluate the dynamics and wear of gas turbine combustors. A unified approach of representing multiple rigid/elastic bodies with numerous contacts is developed. Representation is made, too, of the changing nature of these contacts — both geometric and material. The entire methodology is implemented in a generic and easy-to-use simulation code which serves as a useful generic design/analysis evaluation tool MAP (Mechanism Analysis Program). Appropriate analytical models for inter-material constitutive laws — both incremental (contact friction, pressure, damping, etc.) and cumulative (wear theories) — are incorporated in the tool. As applications of this approach, dynamic simulations of two different gas turbine combustor designs are run, and comparisons are made with real systems. Excellent correlations have been obtained, both with respect to laboratory test (accelerometer) data, and wear patterns at various contacts and junctions on field samples.

Using Automatic Feature Recognition to Interface CAD to CAPP

Interfacing CAD to CAPP (computer-aided process planning) is crucial to the eventual success of a fully-automated computer-integrated manufacturing (CIM) environment. Current CAD and CAPP systems are separated by a “semantic gap” that represents a fundamental difference in the ways in which they represent information. This semantic gap makes the interfacing of CAD to CAPP a non-trivial task. This paper argues that automatic feature recognition is an indispensable technique in interfacing CAD to CAPP. It then surveys the current literature on automatic feature recognition methods and systems, and analyzes their suitability as CAD/CAPP interfaces. It also describes a relatively recent automatic feature recognition method based on volumetric decomposition, using Kim’s alternating sum of volumes with partitioning (ASVP) algorithm. The paper’s main theses are: (1) that most previous automatic feature recognition approaches are ultimately based on pattern-matching; (2) that pattern-matching approaches are unlikely to scale up to the real world; and (3) that volumetric decomposition is an alternative to pattern-matching that avoids its shortcomings. The paper concludes that automatic feature recognition by volumetric decomposition is a promising approach to the interfacing of CAD to CAPP.

The Development of an Automatic Three-Dimensional Mesh Generator via Modified Ray Casting

A fully automatic three-dimensional mesh generation method is developed by modifying the well-known ray casting technique. The method is capable of meshing objects modeled using the CSG representation scheme. The input to the method consists of solid geometry information, and mesh attributes such as element size. The method starts by casting rays in 3D space to classify the empty and full parts of the solid. This information is then used to create a cell structure that closely models the solid object. The next step is to further process the cell structure to make it more succinct, so that the cells close to the boundary of the solid object can model the topology with enough fidelity. Moreover, neighborhood relations between cells in the structure are developed and implemented. These relations help produce better conforming meshes. Each cell in the cell structure is identified with respect to a set of pre-defined types of cells. After the identification process, a normalization process is developed and applied to the cell structure in order to ensure that the finite elements generated from each cell conform to each other and to other elements produced from neighboring cells. The last step is to mesh each cell in the structure with valid finite elements.

Development and Application of Graphics Displays for a Mini-Computer Based Vibration Analysis System

Radian Corporation has developed extensive data display capabilities to analyze vibration and acoustic data from structures and rotating equipment. The Machinery Interactive Display and Analysis System (MIDAS) displays data collected through the acquisition functions of MIDAS. The graphics capabilities include displaying spectra in three-dimensional waterfall and in X-Y formats. Both types of plots can relate vibrations to time, equipment speed, or process parameters. Using menu-driven parameter selection, data can be displayed in formats that are the most useful for analysis. The system runs on a popular mini-computer, and it can be used with a great variety of graphics terminals, workstations, and printer/plotters. The software was designed and written for interactive display and plotting. Automatic plotting of large data files is facilitated by a batch plotting mode. The user can define display formats for the analysis of noise and vibration problems in the electric utility, chemical processing, paper, and automotive industries. This paper describes the history and development of graphics capabilities of the MIDAS system. The system, as illustrated in the examples, has proven efficient and economical for displaying large quantities of data.

Effect of Engine Operating Parameters on Combustion and Emission Characteristics

Numerical simulation of flow, combustion, heat release rate and pollutants emission characteristics have been obtained using a single cylinder internal combustion engine operating with propane as the fuel. The data are compared with experimental results and show excellent agreement for peak pressure and the rate of pressure rise as a function of crank angle. The results obtained for NO and CO are also found to be in good agreement and are similar to those reported in the literature for the chosen combustion chamber geometry. The results have shown that both the combustion chamber geometry and engine operating parameters affects the flame growth within the combustion chamber which subsequently affects the pollutants emission levels. The code employed the time marching procedure and solves the governing partial differential equations of multi-component chemically reacting fluid flow by finite difference method. The numerical results provide a cost effective means of developing advanced internal combustion engine chamber geometry design that provides high efficiency and low pollution levels. It is expected that increased computational tools will be used in the future for enhancing our understanding of the detailed combustion process in internal combustion engines and all other energy conversion systems. Such detailed information is critical for the development of advanced methods for energy conservation and environmental pollution control.

A Comparative Study of Numerical Techniques for Nonequilibrium Method of Characteristics

Three techniques which employ different approaches for obtaining a method of characteristics solution for chemical non-equilibrium flows are reviewed and compared. Two features of the solution process are evaluated to determine their effect on the accuracy of the solution. The first aspect to be considered is the integration of the stiff conservation equations in a unit process. A new fifth-order accurate, multi-step integration routine is contrasted with a first-order accurate, single-step forward differencing scheme. The second comparison is designed to determine if a solution of the flowfield along continuous streamlines is superior to one along discontinuous segments of the streamlines. Tests are performed, using a chemical model describing the supersonic combustion of H2-air. Calculations of single unit processes are used to validate the techniques and to determine suitable grid sizes. Solutions for constant area duct flow show that the techniques which use the multi-step integration routine are more accurate. Results from the constant area duct test, for an initial pressure of 3.685 atm, show that a method of characteristics technique which utilizes continuous streamlines is able to converge at a grid size two orders of magnitude larger than that needed by a technique which uses discontinuous segments of streamlines.

A Study of a Sharp 90 Degree Curved Flow

An investigation of air flow along a 90 degree elbow-like tube is conducted to determine the velocity and temperature distributions of the flow. The tube has a sharp 90 degree turn with a radius of curvature of almost zero. The flow is assumed to be a steady two-dimensional turbulent flow satisfying the ideal gas relation. The flow will be analyzed using a finite difference technique with the K-ε turbulence model, and the algebraic stress model (ASM). The FLUENT code was used to determine the parameter distributions in the passage. There are certain conditions for which the K-ε model does not describe the fluid phenomenon properly. For these conditions, an alternative turbulence model, the ASM with or without QUICK was employed. FLUENT has these models among its features. The results are compared with the result computed by using elementary one-dimensional theory including the kinetic energy loss along the passage of the sharp 90 degree curved tube.

Analysis of Pressure Suppression Pool Swell Using Microcomputers

The present paper discusses the development of a computer software or code for a best-estimate analysis of Pressure Suppression Pool Hydrodynamics in a Pressurized Heavy Water Reactor (PHWR) system during a Loss-of-Coolant Accident (LOCA) at the primary heat transport system. The software has been developed on Microcomputers, namely, PC-XT or AT (286) under MS-DOS operating system.

Natural Convection-Radiation Induced Flows of Air Over a Horizontal Plate

Combined effects of natural convection and radiation on a laminar boundary-layer flow over a semi-infinite horizontal flat plate are studied. Increasing the natural convection-radiation interaction in the boundary layer increases the shearing stress on the wall, the boundary-layer thickness, the maximum velocity attainable by fluid, and the buoyancy force in the boundary layer. However, the temperature gradient at the plate decreases with increasing interaction. They generally increase with increasing temperature ratio. Automated computations of the dimensionless velocity, temperature, and buoyancy force in the boundary layer yield convergent solutions which are substantially different from those available in the literature for high natural convection-radiation interactions.

Design Optimization of a Packing Material Using Finite Element Techniques

This paper presents size and shape optimization of a recyclable packing material. Shell thickness and shape was optimized to prevent contact of surfaces within the packing material. Thus optimization was used to improve cushioning characteristics of the packing material.