Engineering topics
Publications and source records attributed to Marek, C. J..
The turbulence entering the turbine is produced in the combustor. High turbulence levels from the combustor can alter the location of the transition point on the turbine vane. The dynamics of turbulence and the progress being made in computing the flow are discussed.
The random vortex method (RVM) of Chorin was developed by the recirculating flows. The RVM method models turbulence from first principles, tracking the vorticity and obtaining the interaction of vorticity with the bulk flow field. A computer program was produced called MIMOC, Modeling the Interface Motion of Combustion, which can be used to calculate the reacting flow field behind a rearward facing step. Several comparisons between experimental data and calculations were made. The RVM methood computes qualitatively good results, but the quantitative agreement as yet is not completely satisfactory. Much of the difficulty may be caused by the treatment of boundary conditions and the techniques used for obtaining statistical averages of velocities and turbulence quantities. For the rearward facing step the computed reattachment length equals the experimental value. However, the reverse velocity in the recirculation zone is over predicted by 300 percent. In the calculations, a uniform entrance velocity was assumed with no boundary layer at the step lip. This high velocity may be overdriving the reverse flow region.
A study was undertaken to determine the effect of boundary layers on autoignition and flashback for premixed Jet-A fuel in a unique high-pressure windowed test facility. A plate was placed in the center of the fuel-air stream to establish a boundary layer. Four experimental configurations were tested: a 24.5-cm-long plate with either a pointed leading edge, a rounded edge or an edge with a 0.317-cm step, or the duct without the plate. Experiments at an equivalence ratio ranging from 0.4 to 0.9 were performed at pressures to 2500 kPa (25 atm.) at temperatures of 600, 645, and 700 K and velocities to 115 meters per second. Flame shapes were observed during flashback and autoignition using high speed cinematography. Flashback and autoignition limits were determined.
A computer code for calculating the flow field and flame propagation in a turbulent combustion tunnel is described. The model used in the analysis is the random vortex model, which allows the turbulent field to evolve as a fundamental solution to the Navier-Stokes equations without averaging or closure modeling. The program was used to study the flow field in a model combustor, formed by a rearward-facing step in a channel, in terms of the vorticity field, the turbulent shear stresses, the flame contours, and the concentration field. Results for the vorticity field reveal the formation of large-scale eddy structures in the turbulent flow downstream from the step. The concentration field contours indicate that most burning occurred around the outer edges of the large eddies of the shear layer.
A study of the effects of secondary air addition on the stability and emissions of a gas turbine combustor has been performed. Tests were conducted with two types of flameholders and varying amounts of dilution air addition. Results indicate that NO(x) decreases with increasing dilution air injection, whereas CO is independent of the amount of dilution air and is related to the gas temperature near the walls. The axial location of the dilution air addition has no effect on the performance or stability. Results also indicate that the amount of secondary air entrained by the flameholder recirculation zone is dependent on the amount of dilution air and flameholder geometry.
a fuel injection system for gas turbines is described including a pair of high pressure pumps. The pumps provide fuel and a carrier fluid such as air at pressures above the critical pressure of the fuel. A supercritical mixing chamber mixes the fuel and carrier fluid and the mixture is sprayed into a combustion chamber. The use of fuel and a carrier fluid at supercritical pressures promotes rapid mixing of the fuel in the combustion chamber so as to reduce the formation of pollutants and promote cleaner burning.
Lean blowout limits were reported for a premixed prevaporized propane jet issuing into a cylindrical combustor. A single hole in a flat plate was used as a flameholder. Flameholders with various hole diameters were used. Jet velocities were varied from 3 to 290 meters per second. The combustor cross sectional area was changed by using different quartz liners of 12.7 and 22.2 millimeters diameters. As a result the combustor Reynolds number varied from 1000 to 9000. Stability was achieved at laminar as well as turbulent conditions. Three zones of flame stability were observed. The blowout equivalence ratio varied with step size and the combustor and jet Reynolds numbers. The combustor inlet mixture temperature was 395 K, and the combustor pressure was 1 atmosphere.
A fuel combustor comprises a chamber with air and fuel inlets and a combination gas outlet. The fuel is supplied to a vaporization zone and fuel and air are mixed in a pair of mixing chambers, each exemplified by a swirl can. The resultant mixture is directed into a combustion zone within the combustor. Heat pipes are arranged with one end portion substantially in the combustion zone and the other end in the vaporization zone of its appropriate mixing chamber. Some of the heat of combustion is thus carried back upstream into the swirl cans, to vaporize the fuel as it enters the vaporization zone in the swirl can, thereby improving vaporization and fuel mixing.
A recently implemented NASA effort specifically aimed at reducing cruise oxides of nitrogen from high-altitude aircraft is discussed. The desired emission levels and the combustor technology required to achieve them are discussed. A brief overview of the SCERP operating plan is given. Lean premixed-prevaporized combustion and some of the potential difficulties that are associated with applying this technique to gas turbine combustors are examined. Base technology was developed in several key areas. These fundamental studies are viewed as a requirement for successful implementation of the lean premixed combustion technique.
Lean equivalence ratios from 0.3 to 0.7 were observed. Combustor inlet air pressures were varied from 0.54 to 2.5 MPa, combustor inlet air temperatures from 550 to 700 K, and reference velocities from 8 to 35 meters per second. Autoignition delay times ranged from 15 to 100 milliseconds and varied inversely with pressure. The Arrhenius activation energy was 41,840 joules per mole. Temperature rise data were obtained in a long premixing-prevaporizing tube at a pressure of 0.56 MPa. Preflame temperature rise data were a function of equivalence ratio, inlet air temperature, and tube residence time. Significant temperature rise occurred above temperatures of 760 K, with autoignition occurring at 775 K for equivalence ratios greater than 0.47. The reactions were similar to cool-flame phenomena. Flashback velocities were measured at temperatures of 610 and 700 K, pressure of 0.56 MPa, and equivalence ratios from 0.6 to 1. Flashback velocities varied from 30 to 65 meters per second.
An experimental study of two coaxial gas streams, which react very rapidly, was performed to investigate the mixing characteristics of turbulent flow fields. The center stream consisted of a CO-N2 mixture and the outer annular stream consisted of air vitiated by H2 combustion. The streams were at equal velocity (50 m/sec) and temperature (1280 K). Turbulence measurements were obtained using hot film anemometry. A sampling probe was used to obtain time averaged gas compositions. Six different turbulence generators were placed in the annular passage to alter the flow field mixing characteristics. The turbulence generators affected the bulk mixing of the streams and the extent of CO conversion to different degrees. The effects can be related to the average eddy size (integral scale) and the bulk mixing. Higher extents of conversion of CO to CO2 were found be increasing the bulk mixing and decreasing the average eddy size.
An experimental study of the turbulent mixing of two coaxial gas streams, which react very rapidly (pseudo-hypergolic chemical kinetics), was performed to investigate the mixing characteristics of turbulent flow fields. The center stream consisted of a CO-N2 mixture and the outer annular stream consisted of air vitiated by H2 combustion. The streams were at equal velocity (50 m/sec) and temperature (1280 K). Turbulence measurements were obtained using hot film anemometry. A sampling probe was used to obtain time averaged gas compositions. Six different turbulence generators were placed in the annular passage to alter the flow field mixing characteristics. The turbulence generators affected the bulk mixing of the streams and the extent of CO conversion to different degrees. The effects can be related to the average eddy size and the bulk mixing. Higher extents of conversion of CO to CO2 were found by increasing the bulk mixing and decreasing the average eddy size.
The effects of uncooled exhaust gas recirculation as an inert diluent on emissions of oxides of nitrogen (NO + NO2) and on combustion efficiency were investigated. Ratios of recirculated combustion products to inlet airflow were varied from 10 to 80 percent by using an inlet air ejector nozzle. Liquid Jet A fuel was used. The flame-tube combustor was 10.2 cm in diameter. It was operated with and without a flameholder present. The combustor pressure was maintained constant at 0.5 MPa. The equivalence ratio was varied from 0.3 to 1.0. The inlet air temperature was varied from 590 to 800 K, and the reference velocity from 10 to 30 m/sec. Increasing the percent recirculation from 10 to 25 had the following effects: (1) the peak NOx emission was decreased by 37 percent, from 8 to 5 g NO2/kg fuel, at an inlet air temperature of 590 K and a reference velocity of 15 m/sec; (2) the combustion efficiency was increased, particularly at the higher equivalence ratios; and (3) for a high combustion efficiency of greater than 99.5 percent, the range of operation of the combustor was nearly doubled in terms of equivalence ratio. Increasing the recirculation from 25 to 50 percent did not change the emissions significantly.
Emissions of nitrogen oxides, carbon monoxide, and unburned hydrocarbons were measured in a burner where liquid Jet A fuel was sprayed into the heated air stream and vaporized upstream of a perforated plate flameholder. The burner was tested at inlet air temperatures at 640, 800, and 833 K, an inlet pressure of 5.6 X 100,000 N/m squared, a reference velocity of 25 m/sec, and equivalence ratios from lean blowout to 0.7. Nitrogen oxide levels of below 1.0 g NO2/kg fuel were obtained at combustion efficiencies greater than 99 percent. The measured emission levels for the liquid fuel agreed well with previously reported premixed gaseous propane data and agreed with well stirred reactor predictions. Autoignition of the premixed fuel air mixture was a problem at inlet temperatures above 650 K with 104 msec premixing time.
Film-cooling experiments were conducted at four levels of free-stream turbulence to test the hypothesis that the film-cooling effectiveness is inversely related to the free-stream turbulence level. The hot-gas operating conditions were held constant at a temperature of 590 K, a pressure of 1 atmosphere, and a velocity of 62 m/sec. The film-cooling air was at ambient inlet temperature, and the film-cooling flow rates were 2.5, 5.0, 7.5, and 10.0 percent of the total airflow. Blockage plates with blockage areas of 0, 52, 72, and 90 percent were placed upstream of the film-cooling slot and produced axial turbulence intensities of 7, 14, 23, and 35 percent, respectively. The film-cooling effectiveness decreased as much as 50 percent as the freestream turbulence intensity was increased from 7 to 35 percent. The value of the turbulent mixing coefficient used in previous work was compared with the axial turbulence intensity. The turbulent mixing coefficient was found to be 10 to 40 percent of the axial turbulence intensity.
Data were taken with a film-cooled test plate placed in the exhaust stream of a rectangular combustor. Results showed that in a fuel-rich zone, fuel entrained into the film-cooling air would burn at the conditions tested. Test conditions were cooling-gas flow rates, 9.5 and 5 percent of total gas flow; cooling-gas velocities, 23 and 12 m/sec; ambient-temperature cooling gas; hot-gas velocity, nominally 215 m/sec; fuel-air ratio to stoichiometric fuel-air ratio values of 0.1, 0.8, 1.0, 1.2, and 1.35, resulting in hot-gas temperatures from 590 to 2100 K; and atmospheric pressure. Analytical prediction of wall temperatures agreed reasonably well with experimental results.
A tangential-injection film cooled test section was placed in the exhaust stream of a high pressure combustor. Film cooling data were taken at pressure of 1, 10, and 20 atmospheres. The film cooling effectiveness was found to be independent of pressure. The data were correlated adequately by a turbulent-mixing film cooling correlation with a turbulent-mixing coefficient of 0.05 + or - 0.02.
Data were obtained on a parallel-flow film- and convection-cooled test section placed in the exhaust stream of a rectangular-sector combustor. The combustor was operated at atmospheric pressure and at exhaust temperatures of 589 and 1033 K (600 and 1400 F). The cooling air was at ambient pressure and temperature. Test results indicate that it is better to use combined film and convection cooling rather than either film or convection cooling alone for a fixed total coolant flow. An optimum ratio of film to convection cooling flow rates was determined for the particular geometry tested. The experimental results compared well with calculated results.