116 Prague, February 17-19, 2021
_______________________________________________________________________
DOI: https://doi.org/10.14311/TPFM.2021.016
A STUDY ON THE FLOW PHYSICS OF ALTITUDE ADAPTIVE NOZZLES
Sandeep Soman1, Prasanth P. Nair2, Abhilash Suryan1
1
Department of Mechanical Engineering, College of Engineering Trivandrum,
Trivandrum, Kerala 695016, India
2
Mechanical Engineering Discipline, IIT Gandhinagar, Gujarat 382355, India
Abstract
Class of rocket nozzles which are capable of changing the effective flow area ratio with
change in ambient pressure as it ascends through atmosphere are called altitude adaptive
nozzles. Aerospike nozzles, expansion deflection nozzles, dual bell and double divergent
nozzles are types of altitude adaptive nozzles. These nozzles are able to adapt to ambient
conditions on account of the complex flow phenomena within. This paper discusses the
topical problem of fluid flow physics in altitude adaptive nozzles such as annular aerospike,
linear plug, expansion deflection and dual bell nozzles as the nozzle operates from low to
high nozzle pressure ratios. The strategy to numerically model the flow within these
nozzles is also discussed.
Keywords: Altitude Adaptive Nozzles, Aerospike Nozzle, Linear Plug Nozzle, Expansion
Deflection Nozzle, Dual Bell Nozzle.
1 Introduction
Space agencies are now concentrating on reducing the total cost of space missions, with more efficient
launch vehicles. The rocket nozzles are the key components in the performance of launch vehicles.
Nozzles perform with maximum efficiency for the design pressure ratio, which corresponds to a particular
altitude in flight. The performance drops at all off design values of nozzle exit pressure. The performance
of a rocket nozzle can be significantly improved if it can adapt to the change of ambient pressure during
flight providing ideal expansion of flow during flight. For this, the nozzle pressure ratio has to be
continuously varied by changing either the throat area or the exit area of the nozzle. This can be achieved
with complex mechanical systems which will increase the overall weight of the launch vehicle. Hence
almost all launch vehicles employ the conventional bell nozzles. The losses are compensated to some
extent by having multiple stages, with each stage nozzle designed to operate at pressure ratios giving
optimum performance within the range of altitudes. However, this causes significant overexpansion at sea
level leading to separation of flow, generating large unsteady asymmetric forces resulting in side loads
that reduce the lifetime and safety margin of the rocket [1].
Altitude adaptive nozzles are suitable for Single Stage to Orbit (SSTO) missions, Reusable Launch
Vehicles (RLVs) or stages, giving improved performance without the need for any mechanical systems.
These nozzles are capable of operating at a higher efficiency in terms of its thrust coefficient, over a wide
range of altitudes compared to a conventional bell nozzle. The improvement in performance is due to the
capability to adapt to ambient conditions and operate at different nozzle area ratios as the rocket ascends
through atmosphere. Aerospike nozzles, plug nozzles, expansion deflection nozzles, dual bell nozzles and
double divergent nozzles are examples for altitude adaptive nozzles [2].
2 Flow Features in Altitude Adaptive Nozzles
Altitude adaptive nozzles operate with better performance by the virtue of flow features that enable
altitude adaption with ambient conditions. Flow physics of different altitude adaptive nozzles such as
aerospike nozzles, plug nozzles, expansion deflection nozzles and a dual bell nozzle at different altitudes
have been extensively studied by many researchers. Grifith proposed the concept of aerospike in 1954 [3].
A number of experimental and numerical studies are available for both annular axisymmetric as well as
linear plug nozzles [4-9, 14-15]. In 1960s, the concept of expansion deflection nozzles was introduced by
G.V.R. Rao [10]. The concept of dual bell nozzles was presented much before aerospikes by Foster and
Cowles in 1949 [11]. Over the years many numerical and experimental studies have been conducted on
these nozzles. The studies consistently reported the performance improvement of altitude adaptive
nozzles over conventional bell nozzle [2, 12-13]. This paper is a review of the studies done by the authors
TOPICAL PROBLEMS OF FLUID MECHANICS 117
_______________________________________________________________________
and attempts to provide an overview on the efforts so far. The prominent flow features captured for the
four major types of altitude adaptive nozzles viz. the aerospike, the plug, the expansion defection and the
dual bell nozzles are discussed in the subsequent sections.
2.1 Aerospike Nozzles
Figure 1: Schematic diagram of an annular aerospike nozzle [3].
Figure 1 describes an axisymmetric conical aerospike nozzle. It consists of a central plug around which
the gases expand. The major difference with conventional nozzles is that the exhaust plume is not
contained by physical boundary of the nozzle. It is exposed to the outer ambient atmosphere and a shear
layer forms between the expanding core jet and the ambient atmosphere. At very low nozzle pressure
ratios (NPR), the exhaust plume remains close to the plug surface. At the design NPR, the plume flows
straight, like a column as in the case of bell nozzle at its design NPR. At very high NPRs, recompressions
and trailing shocks occurs inside the expanding plume that helps to keep it in a column shaped structure.
a) NPR 2.57 b) NPR 3.82
Figure 2: Combined Mach and Numerical Schlieren Contour for full length plug at a) NPR 2.57 and b)
NPR 3.82 [7].
Computational schemes developed by the team are capable of capturing all these flow features
accurately [7-9]. At very low NPRs, the over expansion shocks from the wall of inner nozzle and the plug
surface forms a λ-shock that will result in a restricted shock separation on the plug surface with a
recirculation bubble downstream of the separation location and subsequent reattachment as shown in Fig.
2a. Multiple shock reflections and expansion fans are formed after formation of λ-shock. With increase in
NPR, an expansion fan emerges from the cowl tip that will overexpand the flow through plug surface
resulting in an over expansion shock from the plug surface resulting in flow separation as shown in Fig.
2b. The computational schemes can predict the flow features for the entire range of operation of the
aerospike nozzles. Thus, the methodology can help significantly in the design and optimization of the
nozzles for different launch vehicles.
118 Prague, February 17-19, 2021
_______________________________________________________________________
a) NPR 2.57 b) NPR 7
Figure 3: Combined Mach and Numerical Schlieren Contour for 40% truncated plug at a) NPR 2.57 and
b) NPR 7 [7].
Figure 3 shows the combined Mach and Schlieren contour for a 40% truncated aerospike nozzle at
two different NPRs corresponding to open wake mode (NPR 2.87) and closed wake mode (NPR 7). By
truncating the plug, the flow reattachment can be avoided and prevents further instabilities in the flow.
2.2 Linear Plug Nozzles
Figure 4: Schematic diagram of a linear plug nozzle.
Figure 4 shows the schematic of a linear plug nozzle with dimensions. After the throat, the flow expands
initially in a primary nozzle and then on the plug surface. Unlike annular aerospike, linear aerospike has a
rectangular cross-section throughout and hence the name linear plug. Similar to annular aerospike nozzle,
a continuous altitude adaption happens, theoretically upto the geometrical area ratio, in the case of a
linear plug since the plume is exposed to ambient atmosphere.
The computational methodology developed for linear plug nozzles are similar to the schemes for
aerospike nozzles. Flow structure varies continuously with increase in NPR, with initially the nozzle
operating at over expanded condition with respect to the primary nozzle at low NPR and then the nozzle
operation commences in under expanded condition when the NPR becomes higher than the design NPR
of the internal nozzle. Figure 5 shows the combined Mach and Numerical Schlieren at two different
NPRs, NPR 7.9 (over expanded) and NPR 20 (under expanded) to summarise the change in flow features
and shock patterns with increase in NPR.
TOPICAL PROBLEMS OF FLUID MECHANICS 119
_______________________________________________________________________
a) NPR 7.9 b) NPR 20
Figure 5: Combined Mach and Numerical Schlieren Contours at a) NPR 7.9 and b) NPR 20.
At very low NPR the primary nozzle is in a highly over expanded situation and it favours the
formation of a Mach Disk. A restricted shock separation occurs on the plug side and free shock separation
occurs on the upper primary nozzle wall. Since two different flow separation phenomenon happens on the
two sides of the primary nozzle, the shock pattern formed is asymmetric. With increase in NPR, the over
expansion shocks are weakened and the Mach Disk turns to a regular reflection and the asymmetry in
shock pattern increases. Inside the supersonic core jet, multiple shock reflections happen until the NPR
crosses the design pressure ratio of the primary nozzle. With further increase in NPR, the over expansion
shocks become weak and transforms to curved barrel shock and an expansion fan emerges from the tip of
the primary nozzle as shown in Fig. 5.b. Similar to the annular plug nozzle, truncation prevents the
reattachment of flow over the plug surface and prevents further instabilities as shown in Fig.6
a) NPR 8 b) NPR 20
Figure 6: Combined Mach and Numerical Schlieren Contours for 40% truncated linear plug at a) NPR 8
and b) NPR 20.
120 Prague, February 17-19, 2021
_______________________________________________________________________
2.3 Expansion Deflection Nozzles
Figure 7: Schematic of an E-D Nozzle with flow phenomenon happening at Open Wake Mode [16].
Figure 8: Schematic of an E-D Nozzle with flow phenomenon happening at Closed Wake Mode [16].
Figures 7 and 8 shows the schematic of an Expansion Deflection (E-D) Nozzle with necessary flow
features happening at two different modes of operation in an expansion deflection nozzle, which are open
wake mode and closed wake mode. An E-D nozzle consists of a bell-shaped nozzle with a pintle placed
symmetrically at the centreline and the shape of which affects the flow through the E-D Nozzle. The
exhaust gases expand radially outward through the corner of the plug and the wall contour deflects the
flow axially towards the exit and hence it is called an expansion deflection nozzle. The shape and position
of the pintle is responsible for altitude compensation by altering the effective area of exhaust plume such
that the exit pressure and ambient pressure always matches when working in open wake mode.
At very low altitudes, the pintle deflects the expanding gases radially outwards thereby separating the
flow from the tip of the pintle, creating a viscous and subsonic recirculation zone behind the pintle, as
shown in Fig. 7. With increase in NPR, size of this recirculation zone deceases and at the same time the
effective area ratio increases and finally at a particular altitude, the wake closes as shown in Fig. 8. The
nozzle act as a normal C-D Bell nozzle after this altitude.
As stated above, at low NPR 10, a separation shock emerges from the tip of the pintle and it reflects
from the tip of the nozzle creating a reflected shock as shown in Fig 9a. At high NPR of 30, the flow
attaches towards the centre and the expansion fan from pintle tip reflects from the inner shear layer as a
reattachment shock as shown in Fig. 9b.
TOPICAL PROBLEMS OF FLUID MECHANICS 121
_______________________________________________________________________
a) NPR 10, Open Wake Mode b) NPR 30, Closed Wake Mode
Figure 9. Flow structure in E-D Nozzle [16].
2.4 Dual Bell Nozzles
Figure 10. Flow structure in a Dual Bell Nozzle [17].
As shown in Fig. 10, a dual bell nozzle consists of a base nozzle and an extension nozzle with two
different area ratios. An inflection point is present in between the base nozzle and the extension nozzle.
Unlike other altitude adaptive nozzles discussed above, a continuous area ratio variation is not possible in
the case of dual bell nozzle. Still its performance is compared higher than a conventional bell nozzle with
same effective exit ratio [13]. The flow phenomenon inside a dual bell nozzle is divided into three
different regimes. In the first regime, at low NPR during the engine start up, the ambient pressure is
comparatively high and the flow separates from the base nozzle with formation of a Mach Disc as shown
in Fig. 10.a. With increase in NPR, the separation location moves downstream and finally reaches the
inflection point. Further increase in NPR does not shifts the separation point but remains at the inflection
point till a particular NPR and this phenomenon is called sneak transition. The phenomenon of sneak
transition is observed in all dual bell nozzles of any shape and size. With further increase in NPR, the
separation point jumps downstream towards the end of the extension nozzle and continues to move
downstream with increase in NPR. The separation location can be controlled by a secondary injection
from the inflection point. The details of flow physics with and without secondary injection in dual bell
nozzles at different NPRs are captured numerically as shown in Fig. 11.
122 Prague, February 17-19, 2021
_______________________________________________________________________
a) NPR 15 b) NPR 70
Figure 11. Flow structure in a Dual Bell Nozzle at low NPR (15) and high NPR (70) [17].
3 Strategy for Numerical Computations
Simulation results presented were performed using ANSYS Fluent software by solving finite volume-
based Reynolds Averaged Navier Stokes (RANS) Equations with suitable turbulence model. The
turbulence model for each case was carefully selected based on a turbulence model study involving
models available such as S-A Model, k-ε model, k-ω model, k-ω SST Model and Transition SST models.
Finally, the turbulence model which best suits the experimental data is selected for further analysis.
Figure 12. Representation of Boundary Conditions used in Case of Linear Plug Nozzle
Boundary conditions used in a typical case is shown in Fig. 12 with linear plug nozzle as an example.
Pressure inlet boundary conditions are used at the inlet of nozzle and pressure outlet boundary condition
is used at the edges of extended domain taken for analysis. Axis and symmetry boundary conditions are
used in the case of axisymmetric cases like dual bell nozzle, annular spike nozzle etc. and planar cases
like planar plug nozzles respectively when half of the nozzle is modelled in two dimensions. For steady
case, pressure based coupled solver with double precision can be used for fast convergence [18, 19].
Computational domain and grid independence study can be conducted using methods specified by Roache
[20]. Spatial discretization is done using least squares cell-based method and convective terms can be
solved using second order upwind scheme. FMG initialization gives better results in some cases while
hybrid initialization is enough in case of linear plug nozzles.
4 Conclusion
Flow physics in major types of altitude adaptive nozzles like annular aerospike nozzle, linear plug nozzle,
expansion deflection nozzle and dual bell nozzle are reviewed based on previous work by authors. The
right strategy for numerical modelling of these nozzles using commercial ANSYS Fluent software is
identified and tested successfully. The strategy presented can be generally used for steady two-
dimensional simulation of a scaled nozzle model with slight variations demanded by each case.
TOPICAL PROBLEMS OF FLUID MECHANICS 123
_______________________________________________________________________
Aerospike, Plug and expansion deflection nozzles are capable of continuous area ratio variation up to its
physical design area ratio while only two area ratios are available in the case of a dual bell nozzle. In case
of truncated aerospike, plug and expansion deflection nozzle, the continuous area ratio variation is
capable up to the NPR where the nozzle operation changes to closed wake mode and above which, all
these nozzles act similar to a conventional bell nozzle. Hence these nozzles can be designed at a higher
area ratio so that the nozzle performs better than a bell nozzle at sea level conditions and near optimum
performance can be achieved at higher altitude.
References
[1] Ostlund, J., and Muhammad-Klingmann, B., 2005, Supersonic flow separation with application to
rocket engine nozzles. Applied Mechanics Reviews, 58(3): 143-177.
[2] Hagemann, G., Immich, H., Van Nguyen, T. and Dumnov, G.E., 1998. Advanced rocket nozzles.
Journal of Propulsion and Power, 14(5), pp.620-634.
[3] Arnold, G.A., Rolls Royce PLC, 1954. Jet-propulsion nozzle for use at supersonic jet velocities.
U.S. Patent 2,683,962.
[4] Humphreys, R.P., Thompson, H.D. and Hoffmann, J.D., 1971. Design of maximum thrust plug
nozzles for fixed inlet geometry. AIAA Journal, 9(8), pp.1581-1587.
[5] Migdal, D., 1972. Supersonic annular nozzles. Journal of Spacecraft and Rockets, 9(1), pp.3-6.
[6] Sule, W.P. and Mueller, T.J., 1973. Annular truncated plug nozzle flowfield and base pressure
characteristics. Journal of Spacecraft and Rockets, 10(11), pp.689-695.
[7] Nair, P.P., Suryan, A. and Kim, H.D., 2019. Computational study on flow through truncated
conical plug nozzle with base bleed. Propulsion and Power Research, 8(2), pp.108-120.
[8] Nair, P.P., Suryan, A. and Kim, H.D., 2019. Study of conical aerospike nozzles with base-bleed
and freestream effects. Journal of Spacecraft and Rockets, 56(4), pp.990-1005.
[9] Nair, P.P., Suryan, A. and Kim, H.D., 2017. Computational study of performance characteristics
for truncated conical aerospike nozzles. Journal of Thermal Science, 26(6), pp.483-489.
[10] Rao, G.V.R., 1960. Analysis of a new concept rocket nozzle. Journal of Liquid Rockets and
Propellants, 2, pp.669-682.
[11] Cowles, F.B. and Foster, C.R., 1949. Experimental study of gas-flow separation in overexpanded
exhaust nozzles for rocket motors.
[12] Chutkey, K., Vasudevan, B. and Balakrishnan, N., 2016. Flow and performance analysis of annular
cluster truncated plug nozzle. Journal of Propulsion and Power, 32(6), pp.1442-1453.
[13] Taylor, N., Steelant, J. and Bond, R., 2011, July. Experimental comparison of dual bell and
expansion deflection nozzles. In 47th AIAA/ASME/SAE/ASEE Joint Propulsion Conference &
Exhibit (p. 5688).
[14] Verma, S.B. and Viji, M., 2011. Linear-plug flow field and base pressure development in
freestream flow. Journal of Propulsion and Power, 27(6), pp.1247-1258.
[15] Chutkey, K., Viji, M. and Verma, S.B., 2017. Effect of clustering on linear plug nozzle flow field
for overexpanded internal jet. Shock Waves, 27(4), pp.623-633.
[16] Paul, P. J., Nair, P. P., Suryan, A., Martin M. J. P. and Kim, H. D, 2020. Numerical Simulation on
Optimization of Pintle Base Shape in Planar Expansion-Deflection Nozzles. Journal of Spacecraft
and Rockets, 57(3), pp.539-548.
[17] Choudhury, S. P., Suryan, A., Pisharady, J. C., Jayashree, A. and Rashid, K., 2018. Parametric
study of supersonic film cooling in dual bell nozzle for an experimental air–kerosene engine.
Aerospace Science and Technology, 78, pp.364-376.
[18] Fluent User’s Guide, ANSYS, Inc., Canonsburg, PA, pp. 1–1146
[19] Kurbatskii, K. and Montanari, F., 2007, January. Application of pressure-based coupled solver to
the problem of hypersonic missiles with aerospikes. In 45th AIAA Aerospace Sciences Meeting and
Exhibit (p. 462).
[20] Roache, P. J., Perspective: a method for uniform reporting of grid refinement studies. (1994): pp
405-413.
