Showing papers by "Terrence W. Simon published in 2004"

A 2-D CFD model of a Free Piston Stirling Engine for Space Applications with Annular Heat Exchangers

Abstract: The CFD-ACE commercial code has been utilized for a 2 -D model of a Free Piston Stirling Engine (FPSE). Several code validations were conducted including laminar flow in oscillatory pipe and parallel plate flows. The CFD results showed good agreement with available experimental data as well as analytical solutions. The 2-D model consisted of an Expansion Space (ES), Heater (HR), Regenerator (RG), Cooler (CR) and Compression Space (CS). The HR and CR were modeled as concentric fins, while the RG utilizes the CFD-ACE porous media model. CFD data were obtained for the PV power from the ES and CS as well as heat in and out of the heat exchangers. The model for the FPSE was conducted for two grids, coarse (64,823 cells) and fine (133,078 cells) and includes the CS, CR, RG, HR and ES. Both the power piston and displacer were also modeled. Results were obtained for energy in and out from each component including enthalpy flux at both sides of the regenerator. The CFD-ACE porous media model was utilized (which is known not to accurately represent the unsteady heat transfer process in the regenerator due to the assumption of gas-solid temperature equilibrium). The results obtained for 110 cycles (coarse grids) and 100 cycles (fine grids) were compared with Sage results. The codes for both cases seem to head towards the right direction of energy balance. The similarities and differences among the different cases were examined and discussed in the paper. Two approaches have been proposed to accelerate the convergence process: 1) Replace the solid walls (as much as possible, e.g. in the heater and cooler) by a uniform temperature surface. This might accelerate the convergence of the CFD process but will alter the B.C., which would result in different heat transfer rates, 2) Model computationally each component separately and merge them one at a time using user subroutine in CFD-ACE. Both approaches have been attempted and the results are encouraging.

CFD Modeling of Surface Roughness in Laminar Flow

Abstract: Surface roughness is known to have significant effect on turbulent flow depending on the magnitude of the roughness. Also, a pipe is considered hydrodynamically smooth (for a given surface roughness) below certain value of Re. As the flow becomes laminar, there is little known about the surface roughness effect, particularly for small channel diameters (micro-channels). Data available to date from the literature, on micro-channels, indicate that the friction-Re relationship deviates from the known Moody Chart, for laminar flow. These differences are attributed, sometimes, to the surface roughness. Micro-channels are currently under consideration as a candidate for Stirling Engine Regenerators. In this paper, the effect of surface roughness on the flow and heat transfer characteristics has been studied using the CFD-ACE commercial software. A CFD model is created with small tooth like structures on the inner walls of a pipe with equal height and spacing between them. CFD analysis is done for different roughness values and the variation of heat transfer and fluid characteristics with the increase in roughness is studied. The study is done on a pipe with a hydraulic diameter of 0.62 mm by considering it first as smooth and then with average roughness values of 1.0 and 2.2 microns. The results were compared with the analytical solution (for smooth pipes) and experimental results available from the literature.

On Experimental Evaluation of Eddy Transport and Thermal Dispersion in Stirling Regenerators

Abstract: Thermal losses from the hot end to the cold end of a Stirling cycle regenerator due to thermal dispersion through the regenerator matrix may significantly degrade the performance of the machine. Because of poor access to the void spaces within the porous medium, no direct measurements of thermal dispersion have been made and dispersion models have been derived indirectly by measuring the overall thermal performance of the regenerator while subtracting off the energy transfer caused by molecular conduction and advected enthalpy flows as computed from volume-averaged fluid velocity and temperature. In the current program, a large-scale porous matrix consisting of stacked screens with a porosity of 90% is installed in a flow rig which is operated in a dynamically similar fashion to that of a Stirling engine regenerator flow. Experiments are conducted to measure turbulent transport of momentum at the exit plane using hot-wires. Future measurements with resistance thermometry will add turbulent transport of thermal energy. Such turbulent transport terms are related to the thermal dispersion term in the volumetric-averaged energy equation for the regenerator matrix. A dispersion model based upon the measurements is proposed and compared with a model deduced from previous measurements of jet spreading in the same matrix and with results of models documented in the literature. The test program is ongoing. The present paper reports on the test program and the experimental results to-date.

Evaluation of Algebraic Transition Models for Application to Unsteady Flows in Low-Pressure Turbines

Abstract: Currently, numerical simulation models for design of low-pressure turbines need submodels for determining the onset and path of transition. It is well known that such models must include the effects of passing wakes. In this paper, data taken in an experiment that simulates flow over the suction surface of a low-pressure turbine blade, as affected by passing wakes, are used for testing models from the literature. The base case data set replicates reasonable design conditions. A second case with double the wake spacing of the base case is applied to heighten the effects of between-wake calming. A third case with high background freestream turbulence intensity allows testing models for application to high disturbance level flows. Tested are (1) separated flow transition onset models such as those of Mayle and of Davis et al. and (2) attached flow transition onset models, such as those of Mayle, Abu-Ghannam and Shaw, Suzen and Huang and Drela. The dependent variables of the models, time-resolved locations of transition onset, are compared to time-resolved experimental values for transition onset. The transition onset values from the models are computed by assuming that the correlations can be applied in a quasistatic way using instantaneous values of the input parameters computed from the experimental data. Comparisons of the three cases are made to discuss the robustness of the transition onset models. Previously, the base case was presented. New in this paper are (1) modeling of an increased wake spacing data set, (2) modeling of a data set with a highly turbulent approach flow and (3) the comparison of the two with the base case results.Copyright © 2004 by ASME