Jet Flow Structure

The boundary condition encountered by an expanding nozzle flow is a constant ambient pressure. The process of pressure adjustment produces a complicated shock structure in the separated jet that alternates from overexpansion to over compression, which is shown by the repeating shock diamonds or cells (Fig. 1). The first one or two cells usually contain a Mach disk across the axis terminating at a triple point. The initial triple point has an incident shock that originates at the separation point on the nozzle wall, a reflected shock, a normal shock or Mach disk, and a slipstream separating supersonic and subsonic flows aft of the triple point. Of primary importance is the . first complete cell after the flow has separated. Subsequent cells are disrupted by shear layers at the boundary and slipstream and by vorticity gradients from the shocks; thus, their shock structure is not as sharply defined. The adverse pressure gradient necessary to separate the mainstream flow and the nozzle boundary layer from the wall is supplied by the separation shock. The shock originates at the separation point, but its strength and location are dictated by the overall adjustment process of the complete first cell, where all parts are controlled by the ambient pressure boundary condition. The proposed flow separation model is dictated by the jet mainstream adjustment to ambient pressure and not by the condition of the nozzle boundary layer.

1) Jet flow structure

Flow separation model Nozzle flow separation may occur when the ambient pressure lies between two limits, the nozzle static pressure at the lip resulting from subsonic compression after reaching sonic velocity at the throat and the nozzle static pressure at the lip from supersonic expansion. Between these two limits, nozzle flow separation is driven by the manner in which the flow adjusts to the ambient pressure.

During operation, mechanical and thermal loads affect the nozzle structure and lead to its deformation. The combustion is unsteady and turbulent. Fuel and oxidizer may exist in droplet form before they evaporate and burn in a bright flame, emitting radiation. The gas molecules may be vibrationally and rotationally excited, dissociated, or ionized. The nozzle surface can be catalytic. The flow is not axisymmetric, which is also valid for the separation line (teepee pattern1), and is violently unsteady.

In thrust optimized parabolic nozzles, an internal shock wave originating from the nozzle throat can interact with the normal shock and cause the appearance of the cap-shock pattern (Fig. 3). Depending on the flow momentum balance2 the flow can separate from the wall without reattaching to it [free-shock separation (FSS)] or with reattaching to it [restricted shock separation (RSS), Fig. 4]. If the normal shock crosses the nozzle centreline downstream of the reflection point of the internal shock wave at the axis (Fig. 5), the shock pattern resembles that of the classical Mach

disk (Fig. 2). For RSS, the reattaching boundary layer, having passed through the oblique shocks associated with boundary-layer separation and reattachment, is very hot and exposes the nozzle wall downstream of the recirculation bubble to large thermal loads. It is not fully understood why the nozzle wall is also subjected to large heat loads in case of FSS. The oblique shock associated with flow separation is responsible for the high pressure and temperature of the separated supersonic flow. The outer region of this separated flow encounters violent turbulent mixing with entrained ambient air, possibly causing post-combustion. If the flow separates inside the nozzle, and the pressure ratio is held constant, the wall downstream of the separation may be exposed to thermal loads high enough to make the nozzle material glow. The mechanisms responsible for the transport of the hot gases to the wall are not well understood. The highly turbulent, threedimensional, and unsteady nature of the flow may offer an explanation. If the flow was steady and axisymmetric, entrained cool ambient air would flow alongside the nozzle wall and cool it. Gas radiation from the post combustion zone is not strong enough to serve as an explanation for the wall heat loads observed in the experiment.

Schematic view of FSS.

Schematic view of cap-shock pattern in a parabolic nozzle.

Schematic view of RSS.

Schematic view of FSS with Mach disk in a parabolic nozzle.

The origin of the rocket nozzle side loads during startup was commonly explained in the literature as a consequence of the viscous separation of the boundary layer from the wall. In particular, other phenomena, related to the transient character of startup flows and to the generation of a recirculating flow region in the nozzle core, could explain the origin of side loads, as the following analysis of wall pressure behaviour would also suggest

Flow separation in supersonic nozzles originates at the wall as a consequence of the adverse pressure gradient experienced by the boundary layer, when the supersonic expansion is large enough to reduce wall static pressure at some value, lower than the ambient one. This phenomenon occurs every time that a rocket nozzle is operated in transient conditions (start-up or shut-down) and can originate during steady-state conditions if the nozzle expansion ratio is large enough. the flow separation can be considered as the ultimate effect of an adverse pressure gradient acting on the supersonic boundary layer: the boundary layer is progressively decelerated and the velocity profile is eventually inverted. Therefore, the boundary layer thickness increases and the flow locally separates, generating a shock wave which originates near the wall and penetrates into the supersonic jet. The pressure jump through the shock wave enhances the adverse pressure gradient, enforcing the local separation and producing a massive flow separation. Once the free shock separation (FSS) is settled, a strong recirculation stream, fed by ambient air, occupies the separated area and sustains the separated flow configuration. It is important to notice that the sustenance of the flow separation is only possible if the recirculating bubble is permanently fed.

To get an optimum performance over the whole flight trajectory, the nozzles are designed for an intermediate pressure ratio pc / pa , at which the exhaust flow is adapted to the ambient pressure. However, this design ambient pressure is chosen high enough to prevent flow separation inside the nozzle during steady-state operation at sea level. Nevertheless, the flow separates from the wall in the divergent part of these nozzles, as long as the chamber pressure pc has not yet reached its nominal value.
Flow separation in rocket nozzles is considered undesirable because an asymmetry in the flow separation can cause dangerous lateral forces, the so-called side loads, which may damage the nozzle. Therefore, flow separation and its theoretical prediction have been and still are the subject of several experimental and theoretical studies. The authors observed ow separationwithin the nozzle, as soon as the wall pressure at the nozzle exit was lower than about 0.35 0.4 times the ambient pressure,dependingon the pressure ratio pc / pw,e . The corresponding formula was soon called the Summer. eld criterion. At the separation point the wall pressure quickly rose to a plateau pressure pp , which is slightly lower than the ambient pressure pa . The source of this pressure rise was shown to be an oblique shock that originatesfromthe separationpoint. In the recirculationzone downstream of the separation point, the wall pressure increases slowly from pp to pw,e . The wall pressure in the exit plane pw,e is, in general, not equal to the ambient pressure pa , but slightly smaller. Nevertheless, the simpli. cation pw,e = pa is frequently used in the literature.

Free-shock separation in overexpanded rocket nozzles, wall pressure pro. le, and phenomenology:, compression waves/shock, and , boundary/shear-layer edge.

Restricted-shock separation in overexpanded rocket nozzles, wall pressure pro. le, and phenomenology:, compression waves/ shock; , expansion waves; and , boundary/shear-layer edge.

Shock Patterns in the Exhaust Plume

In general, the exit wall pressure pw,e of rocket nozzles with operation ranges from sea level to high altitude is chosen high enough to prevent flow separation inside the nozzle. As a consequence, an overexpanded, but attached, nozzle flow is produced at takeoff. At the nozzle exit the lower pressure of the exhaust gases is adapted to the higher ambient pressure by means of an oblique shock. A widespread opinion about the exhaust plume of overexpanded, attached nozzle flows is that there is either a regular reflection of this shock at the centreline or a Mach reflection, resulting in a Mach disk. This is correct for the case of a two-dimensional plane nozzle producing a uniform exit flow as shown in Fig. 3.Which one of these shock patterns will occur mainly depends on the exit pressure ratio pe / pa and exit Mach number Me . Generally, a strong overexpansion and low exit Mach numbers promote the formation of a Mach disk, whereas regular reflection is reached for weaker overexpansion and higher exit Mach numbers. 14 As a rough approximation for air as exhaust gas, regular reflection occurs as soon as the value of the expression [0.81 M ( pe / pa )0.7] exceeds unity This implies that the high-Mach-number ow near the centerline limited by the internal shock is responsible for the cap-shock pattern, as it has been only observed in thrustoptimized rocket nozzles, both experimentally and numerically. In fact, a triple point exists,where the internal shock, the small normal shock, and the cone-shapedoblique shock meet (point c in Fig. 7b), indicatinga connectionbetween the internalshock and the existence of the cap-shock pattern.

Major differences between Mach disk and Cap-shock pattern : 1) Across the small normal shock of the cap-shockpattern (points de in Fig. 7b), the pressurerises to approximatelyambientpressure, whereas downstreamof a Mach disk, the pressure is always higher than the ambient. 2) A cap-shock pattern always includes a cone-shapedshock that

is inclinedaway from the centerline (points c f in Fig. 7b), whereas the oblique shocks occurring upstream of the Mach disk are always inclined toward the axis. 3) A stable vortex is trapped downstream of the small normal shock in the cap-shock pattern, whereas there is no recirculation behind a Mach disk.

a)Mach disk

Cap-shock pattern with trapped vortex

The transition from cap-shock pattern to the Mach disk occurs as soon as the small normal shock at the centerline in the Kernel, which moves downstreamfor increasing thrust-chamberpressures, reaches the intersection of the internal shock with the centerline because from that point on there is no higher Mach-number ow near the centerline any more.