Engineering topics
Publications and source records attributed to Ghorashi, Bahman.
This report includes copies of relevant publications, including conference proceedings, resulting from studies funded by NAG3-1010. Topics include flow visualization, LDV (laser doppler velocimeter), straight and curved simulations, and kinetics modeling of nitrogen oxides, including a turbulent mixing model for O2 and a maximum mixedness model.
A thermal NO(x) prediction model is developed to interface with a CFD, k-epsilon based code. A converged solution from the CFD code is the input to the postprocessing model for prediction of thermal NO(x). The model uses a decoupled analysis to estimate the equilibrium level of (NO(x))e which is the constant rate limit. This value is used to estimate the flame (NO(x)) and in turn predict the rate of formation at each node using a two-step Zeldovich mechanism. The rate is fixed on the NO(x) production rate plot by estimating the time to reach equilibrium by a differential analysis based on the reaction: O + N2 = NO + N. The rate is integrated in the nonequilibrium time space based on the residence time at each node in the computational domain. The sum of all nodal predictions yields the total NO(x) level.
Successful modeling of combustion and emissions in gas turbine engine combustors requires an adequate description of the reaction mechanism. For hydrocarbon oxidation, detailed mechanisms are only available for the simplest types of hydrocarbons such as methane, ethane, acetylene, and propane. These detailed mechanisms contain a large number of chemical species participating simultaneously in many elementary kinetic steps. Current computational fluid dynamic (CFD) models must include fuel vaporization, fuel-air mixing, chemical reactions, and complicated boundary geometries. To simulate these conditions a very sophisticated computer model is required, which requires large computer memory capacity and long run times. Therefore, gas turbine combustion modeling has frequently been simplified by using global reaction mechanisms, which can predict only the quantities of interest: heat release rates, flame temperature, and emissions. Jet fuels are wide-boiling-range hydrocarbons with ranges extending through those of gasoline and kerosene. These fuels are chemically complex, often containing more than 300 components. Jet fuel typically can be characterized as containing 70 vol pct paraffin compounds and 25 vol pct aromatic compounds. A five-step Jet-A fuel mechanism which involves pyrolysis and subsequent oxidation of paraffin and aromatic compounds is presented here. This mechanism is verified by comparing with Jet-A fuel ignition delay time experimental data, and species concentrations obtained from flametube experiments. This five-step mechanism appears to be better than the current one- and two-step mechanisms.
Successful modeling of combustion and emissions in gas turbine engine combustors requires an adequate description of the reaction mechanism. Detailed mechanisms contain a large number of chemical species participating simultaneously in many elementary kinetic steps. Current computational fluid dynamic models must include fuel vaporization, fuel-air mixing, chemical reactions, and complicated boundary geometries. A five-step Jet-A fuel mechanism which involves pyrolysis and subsequent oxidation of paraffin and aromatic compounds is presented. This mechanism is verified by comparing with Jet-A fuel ignition delay time experimental data, and species concentrations obtained from flametube experiments. This five-step mechanism appears to be better than the current one- and two-step mechanisms.
A review of chemically reactive flow analysis is presented. Key studies are reviewed. The approaches to modeling chemically reactive flows that are discussed are PDF, conserved PDF, conserved scalar, stochastic, and mixing length. Turbulent mixing mechanisms are also discussed and a model is formulated that could be used for optimizing non-premixed gas reacting systems. Consequently, a re-examination of the NO(x) reduction concepts for nonpremixed reactants are suggested together with two new designs for low NO(x) subsonic combustor research.
An experimental and analytical study was performed on a lean, premixed-prevaporized Jet A/air flame tube. The NO(x) emissions were measured in a flame tube apparatus at inlet temperatures ranging from 755 to 866 K (900 to 1100 F), pressures from 10 to 15 atm, and equivalence ratios from 0.37 to 0.62. The data were then used in regressing an equation to predict the NO(x) production levels in combustors of similar design. Through an evaluation of parameters it was found that NO(x) is dependent on adiabatic flame temperature and combustion residence time, yet independent of pressure and inlet air temperature for the range of conditions studied. This equation was then applied to experimental data that were obtained from the literature, and a good correlation was achieved.
The papers presented in this volume provide an overview of the latest advances in laser anemometry and optical flow diagnostics. Topics discussed include turbulence, jets, and chaos; novel optical techniques for velocity measurements; chemical reactions and combusting flows; and LDA/CFD interface. Attention is also given to particle image velocimetry, high speed flows and aerodynamic flows, internal flows, particle sizing, optics and signal processing, two-phase flows, and general fluid mechanics applications.
A study was conducted to determine the effect of eddy distribution on momentum and heat transfer near the wall in turbulent pipe flow. The buffer zone was of particular interest in that it is perhaps the most complicated and least understood region in the turbulent flow field. Six eddy diffusivity relationships are directly compared on their ability to predict mean velocity and temperature distributions in turbulent air flow through a cylindrical, smooth-walled pipe with uniform heat transfer. Turbulent flow theory and the development of the eddy diffusivity relationships are briefly reviewed. Velocity and temperature distributions derived from the eddy diffusivity relationships are compared to experimental data for fully-developed pipe flow in turbulent air at a Prandtl number of 0.73 and Reynolds numbers ranging from 8100 to 25 000.