Dmytro M. Voytovych

Simulation of Coupled Supersonic Inlet and a Fan

Simulations of a coupled supersonic inlet and fan were carried out for to analyze the effects of developed radial distortion on fan performance and stability. The coupling at the interface between the inlet and fan was accomplished using a mixing plane. An in-house implicit second-order solver with upwind discretization of fluxes and dual-time stepping for unsteady calculations was used. The Navier-Stokes equations were complemented by a ω κ - turbulence model. A choked nozzle with an adjustable throttle area was used for the outlet boundary to allow natural flow development and to control mass flow rate along a speed line. The inlet profiles at the aerodynamic interface plane were realistic in that they depended on the mass flow rate and the cowl lip shock location/shape. Total pressure losses varied from 4% in the free steam and up to 20% near the hub. The effect of distortion was reduced at lower mass flow rates recovering the performance. Unsteady simulations at unstable flow conditions beyond the stability limit were carried out as well. It was found that under these conditions a double loop was developed on the performance map as a result of coupling between the fan and cowl shock at the inlet.

Analysis of Transient Mixing of Streams Injected into a Chamber with a Cavity

The transient mixing between gases of dissimilar molecular weight is simulated numerically with a goal of identifying which type of initial condition most closely matches companion experimental observations. Nitrogen and helium are chosen as the working fluids. The two-dimensional geometry of interest is composed of helium channels on either side of a central nitrogen channel feeding into a mixing chamber. The rate at which a cavity in the injector face on one side of the three channels fills is of particular interest. Two pairs of initial conditions are considered. In one pair, transient flows of helium are injected into a background flow of nitrogen established by first steady, and then transient, computations. In the second, both fluid streams are started simultaneously and injected into quiescent helium or nitrogen. With all initial conditions, the mixing process exhibited asymmetries similar to those observed in experiment but considerably stronger mixing and much closer agreement with experiment occurred when the background nitrogen flow was computed in transient fashion. The results indicate the face cavity has a minor impact on flow asymmetry but its rate of fill appears to be dominated by span-wise, three-dimensional motions in the experiment that are absent in the present two-dimensional simulations.

Modeling of Transient Flow Mixing of Streams Injected into a Mixing Chamber

The transient mixing between gases of dissimilar molecular weight is simulated numerically and compared with companion experiments. The two -dimensional geometry of interest is composed of two helium channe ls on either side of a central nitrogen channel feeding into a mixing chamber. The mixing process starts by first establishing a steady flow of nitrogen and then injecting the helium streams into this nitrogen -filled chamber in transient fashion. Simulat ions of the steady nitrogen flow indicate a strong tendency for the jet to attach to the chamber walls. As the helium enters the chamber the attachment point of this established nitrogen jet moves farther downstream but the asymmetric flow pattern still p ersists. The helium flow fills the smaller recirculation region on the attachment side in 50 to 100 ms and eventually fills the larger recirculation region. Comparisons between two - dimensional simulations and quantitative PLIF images indicate reasonable qualitative agreement. Representative three -dimensional solutions show non -planar effects have a significant effect on the mixing process . I. Introduction Ignition is recognized as one the critical drivers in the reliability of multiple -start rocket engin es. Residual combustion products from previous engine operation can condense on valves and related structures thereby creating difficulties for subsequent starting procedures. Alternative ignition methods that require fewer valves can mitigate the valve reliability problem, but require improved understanding of the spatial and temporal propellant distribution in the pre -ignition chamber. Current design tools based mainly on one -dimensional analysis and empirical models cannot predict local details of the injection and ignition processes. The goal of this work is to evaluate the capability of the modern computational fluid dynamics (CFD) tools in predicting the transient flow mixing in pre -ignition environment by comparing the results with the experimental data. This study is a part of a program to improve analytical methods and methodologies to analyze reliability and durability of combustion devices. In the present paper we describe a series of detailed computational simulations of the unsteady mixing events as the cold propellants are first introduced into the chamber as a first step in providing this necessary environmental description. The present computational modeling represents a complement to companion experimental simulations 1,2 and includes co mparisons with experimental results from those efforts. A large number of rocket engine ignition studies have been previously reported. Here we limit our discussion to the works discussed in Refs. 3, 4 and 5 which are both similar to and different from t he present approach.

Effects of Radial Distortion on Performance of a Fan

κ turbulence model. A choked nozzle with adjustable throttle area was used for the outlet boundary to allow natural flow development and to control mass flow rate along a speed line. The results demonstrated that including the nozzle helped to compute solutions closer to stall margin. The solutions became unsteady and stable approaching the stall margin. For the clean inlet, the unsteadiness originated near the tip due to tip vortex oscillations and was larger at lower rotational speeds. Low momentum zone developed near the pressure side of the blade close to the leading edge due to tip vortex and oblique shock interaction. Imposing the radial distortion reduced operational range. At higher speeds led to development of separated flow near the hub at the exit of the passage.

Investigation of Effects of Radial Distortion on Transonic Fan Behavior

The flow analysis around blades of a transonic fan is presented for both clean and radially distorted inlets. Computations are shown for four-blade passages that are accomplished with a second order accurate code using a k-ω turbulence model. The mass flow rate along a speed line is controlled by varying a choked nozzle downstream of the fan. The results show good agreement with data for three speed lines. In the near-stall region, the flow first becomes unsteady and then unstable with the unsteadiness increasing at lower speeds. The four-blade simulations remained stable to lower mass flow rates than the single-blade simulations. In the near-stall vicinity, tip vortex breakdown occurred creating a low momentum zone near the blade tip on the pressure side that grew as the mass flow was decreased until it eventually blocked the passage. The presence of distortion reduced the operational range and moved the stall line to higher mass flow rates. At high speeds distortion reduced both the mass flow rate and total pressure ratio while at lower speeds, the choking mass flow rate was reduced, but the total pressure ratio was slightly improved. The flow separation near the hub on the suction side was caused by the distortion. Its size was decreasing with rotational speed.Copyright

Numerical Study of Rotating Stall of a Modern Fan

The aerodynamic instability known as rotating stall of an isolated fan is considered. Steady and unsteady 3-D simulations were carried out for flows around a high-speed transonic aero-engine fan. Several speed characteristic lines of the performance map were analyzed ranging from almost subsonic to transonic flow conditions with strong shocks. Unstructured grids were used for single- and multiple-blade passages. A choked nozzle was used for the outlet boundary to allow natural flow development and to avoid employing radial equilibrium. A bleed channel before the nozzle was used to control the mass flow rate along a particular speed line. The implicit solver is based on upwind discretization of fluxes with dual-time stepping for the unsteady calculations and a ω κ - turbulence model. The transition from steady to unsteady flow patterns was observed during calculations from the choked condition to stall on a fixed speed characteristic. The presence of strong shocks at higher rotational speeds caused more abrupt transition to stall with higher pressure drop, whereas the transition was more gradual at lower speeds.

Numerical Simulation of Mixing Between Established Gas Flow and Start-up Injected Gas Inside a Chamber

Numerical simulations of a start-up helium flow injected into an established background flow of nitrogen are carried out to identify the dominant features of the transient mixing process between these two dissimilar gases. Simulations were accomplished on both two- and three-dimensional grids and are compared with companion experiments of mixing in a confined, ‘two-dimensional’ chamber. The simulations started by first simulating the background nitrogen flow and then initiating the helium flow. Both computational and experimental results indicate the flow is asymmetric, three-dimensional and highly unsteady but two-dimensional simulations give reasonable predictions on the chamber mid-plane. Both the initial nitrogen-only jet and the compound helium/nitrogen jet diverge from the centerline and attach to one chamber wall. The recirculation region generated by this attachment location is the first to fill with helium, followed by the downstream portion with the near field region opposite this recirculation zone filling last.