Search

In this paper, we present software tools for predicting the rate and nature of observable SEU induced errors in microprocessor systems. These tools are built around a commercial microprocessor simulator and are used to analyse real satellite application systems. Results obtained from simulating the nature of SEU induced errors are shown to correlate with ground-based radiation test data.

This paper discusses large-scale, low-speed experiments that explain unexpected flow-interaction phenomena witnessed during annular cascade studies into the influence of external crossflow on film cooling hole discharge coefficients. More specifically, the experiments throw light on the crossover phenomenon, where the presence of the external crossflow can, under certain circumstances, increase the discharge coefficient. This is contrary to most situations, where the external flow results in a decrease in discharge coefficient. The large-scale testing reported helps to explain this phenomenon through an increased understanding of the interaction between the emerging coolant jet and the free-stream flow. The crossover phenomenon came to light during an investigation into the influence of external crossflow on the discharge coefficients of nozzle guide vane film cooling holes. These experiments were performed in the Cold Heat Transfer Tunnel (CHTT), an annular blowdown cascade of film cooled vanes that models the three-dimensional external flow patterns found in modern aero-engines. (Rowbury et al., 1997, 1998). The variation in static pressure around the exit of film cooling holes under different flow conditions was investigated in the large-scale tests. The study centered on three holes whose geometries were based on those found in the leading edge region of the CHTT vanes, as the crossover phenomenon was witnessed for these rows during the initial testing. The experiments were carried out in a low-speed wind tunnel, with the tunnel free-stream flow velocity set to match the free-stream Reynolds number (based on the local radius of curvature) and the 'coolant' flow velocity set to replicate the engine coolant-to-free-stream momentum flux ratio. It was found that the apparent enhancement of film cooling hole discharge coefficients with external crossflow was caused by a reduction in the static pressure around the hole exit, associated with the local acceleration of the free-stream around the emerging coolant jet. When these measured static pressures (rather than the free-stream static pressure) were used to calculate the discharge coefficient, the crossover effect was absent. The improved understanding of the crossover phenomenon and coolant-to-free-stream interactions that has been gained will be valuable in aiding the formulation of predictive discharge coefficient schemes. [DOI: 10.1115/1.1375171]

An empirical means of predicting the discharge coefficients of film cooling holes in an operating engine has been developed. The method quantifies the influence of the major dimensionless parameters, namely hole geometry, pressure ratio across the hole, coolant Reynolds number, and the freestream Mach number. The method utilizes discharge coefficient data measured on both a first-stage high-pressure nozzle guide vane from a modern aero-engine and a scale (1.4 times) replica of the vane. The vane has over 300 film cooling holes, arranged in 14 rows. Data was collected for both vanes in the absence of external flow. These noncrossflow experiments were conducted in a pressurized vessel in order to cover the wide range of pressure ratios and coolant Reynolds numbers found in the engine. Regrettably, the proprietary nature of the data collected on the engine vane prevents its publication, although its input to the derived correlation is discussed. Experiments were also conducted using the replica vanes in an annular blowdown cascade which models the external flow patterns found in the engine. The coolant system used a heavy foreign gas ([SF.sub.6]/Ar mixture) at ambient temperatures which allowed the coolant-to-mainstream density ratio and blowing parameters to be matched to engine values. These experiments matched the mainstream Reynolds and Mach numbers and the coolant Mach number to engine values, but the coolant Reynolds number was not engine representative (Rowbury, D. A., Oldfield, M. L. G., and Lock, G. D., 1997, 'Engine-Representative Discharge Coefficients Measured in an Annular Nozzle Guide Vane Cascade,' ASME Paper No. 97-GT-99, International Gas Turbine and Aero-Engine Congress & Exhibition, Orlando, Florida, June 1997; Rowbury, D. A., Oldfield, M. L. G., Lock, G. D., and Dancer, S. N., 1998, 'Scaling of Film Cooling Discharge Coefficient Measurements to Engine Conditions,' ASME Paper No. 98-GT-79, International Gas Turbine and Aero-Engine Congress & Exhibition, Stockholm, Sweden, June 1998). A correlation for discharge coefficients in the absence of external crossflow has been derived from this data and other published data. An additive loss coefficient method is subsequently applied to the cascade data in order to assess the effect of the external crossflow. The correlation is used successfully to reconstruct the experimental data. It is further validated by successfully predicting data published by other researchers. The work presented is of considerable value to gas turbine design engineers as it provides an improved means of predicting the discharge coefficients of engine film cooling holes. [DOI: 10.1115/1.1354137]

The influence of surface roughness on heat transfer coefficient and cooling effectiveness for a fully film cooled three-dimensional nozzle guide vane (NGV) has been measured in a transonic annular cascade using wide band liquid crystal and direct heat flux gages (DHFGs). The liquid crystal methods were used for rough surface measurements and the DHFGs were used for the smooth surfaces. The measurements have been made at engine representative Mach and Reynolds numbers and inlet free-stream turbulence intensity. The aerodynamic and thermodynamic characteristics of the coolant flow have been modeled to represent engine conditions by using a heavy 'foreign gas' (30.2 percent [SF.sub.6] and 69.8 percent Ar by weight). Two cooling geometries (cylindrical and fan-shaped holes) have been tested. The strategies of obtaining accurate heat transfer data using a variety of transient heat transfer measurement techniques under the extreme conditions of transonic flow and high heat transfer coefficient are presented. The surfaces of interest are coated with wide-band thermochromic liquid crystals, which cover the range of NGV surface temperature variation encountered in the test. The liquid crystal has a natural peak-to-peak roughness height of 25 [micro]m creating a transitionally rough surface on the NGV. The time variation of color is processed to give distributions of both heat transfer coefficient and film cooling effectiveness over the NGV surface. The NGV was first instrumented with the DHFGs and smooth surface tests preformed. Subsequently the surface was coated with liquid crystals for the rough surface tests. The DHFGs were then employed as the means of calibrating the liquid crystal layer. The roughness of 25 [micro]m, which is the typical order of roughness for the in-service turbine blades and vanes, increases the heat transfer coefficient by up to 50 percent over the smooth surface level. The film cooling effectiveness is influenced less by the roughness. [S0889-504X(00)00804-7]

An ion microdosimetry extension to the Monte Carlo High Energy Transport Code (HETC) has been developed to allow tracking of all the reaction products and has been applied to model charge-deposition spectra in pin diodes caused by atmospheric neutron spectra, as well as upsets in DRAMs from ground and space irradiation by protons. These cases cover sensitive zone sizes ranging from hundreds of microns to sub-micron. Angular distributions of both incident particles and reaction products are found to be important, particularly for the prediction of multiple-bit upsets in devices of small feature size.

This paper presents the results of two test programmes using novel instrumentation to characterise the pressure and turbulent velocity fields in gas-turbine combustor exit flows. The probes are uncooled, therefore a fast-insertion traverse system is employed to prevent thermal degradation of the instrumentation in these severely hostile high-temperature environments. High-bandwidth ultra-miniature pressure transducers are used to measure unsteady total pressure, whilst a Pitot tube is employed to measure time-averaged total pressure. The probes are 4 mm in diameter with a measurement bandwidth of the order of 100 kHz. In the first test programme, the probes are used to characterise the streamwise turbulent velocity field approximately two axial chords downstream of an uncooled single-stage turbine in a turbojet engine. Established data reduction methods and calibration against a hot-wire are used to obtain turbulent velocity fluctuations from unsteady total pressure measurements. Comprehensive turbulence results are presented including time-histories, power spectra, intensities, and lengthscales obtained at four-engine conditions and at two radial and two circumferential measurement locations. In the second test programme the probes are demonstrated in an industrial combustor rig, featuring a can combustor with swirler nozzle and no dilution holes, at temperatures up to 1500 K. Static pressure fluctuations are obtained up to 100 kHz, and some typical combustor spectral features are identified. [ABSTRACT FROM AUTHOR]