The Influence of the Level of the Flow Path Blockage at the Inlet on

the Fan Characteristics

Grigorii Popov, Oleg Baturin, Andrei Volkov, Daria Kolmakova, Vasilii Zubanov,

Anastasia Korneeva and Yulia Novikova

Samara National Research University, Samara, Russia

Keywords: Nonuniformity, Fan, Turbofan Engine, Low Pressure Spool.

Abstract: The paper presents the results of numerical simulation of the effect of flow nonuniformities at the engine inlet

on the working process of the engine fan. Flow nonuniformities is created by pushing the interceptor into the

flow part of inlet device like as it is often done during field tests. The authors have created a numerical model

capable of considering non-stationary processes in the fan using nonlinear harmonic analysis. As a result,

qualitative and quantitative estimates were obtained of the influence of overlapping of the inlet duct by the

interceptor on the main parameters of the fan workflow. It is shown that the more the duct is blocked, the

more its parameters are deteriorated. Moreover, the deterioration is not linear, but according to the dependence

of the 2nd order.

NOMENCLATURE

G mass flow rate of the working fluid, kg/s;

p* total pressure, Pa;

Т* total temperature, K;

n rotor speed, %;

m bypass ratio;

 efficiency;

Y+ non-dimensional wall distance;

 flow angle, degree;

LPC low pressure compressor;

RW rotor wheel;

GV guide vane;

NLH nonlinear harmonic analysis.

Note. The flow angles in this research are measured

from the aerofoil cascade front.

1 INTRODUCTION

Inlet devices of bypass turbofan engines (characterized

by a high degree of bypass ratio) of modern passenger

aircraft have a large flow area and are relatively short

(their length is less than the diameter). It would seem

that losses in such conditions should be minimal in all

typical flight conditions. However, in some cases (for

example, with a strong side wind, flying sideways, etc.)

a separation flow occurs at the inlet edge of the air

intake, which causes the flow at the fan inlet to become

uneven. This, in turn, causes a significant reduction in

the efficiency of the low-pressure compressor and the

engine. In addition, the inlet nonuniformity causes

oscillations of the fan blades, which can lead to their

destruction.

For a long time, the study of the influence of inlet

nonuniformity on the GTE workflow was carried out

during field tests with an aircraft air intake or its

simulator (Figure 1). The latter is a complex of

resistances (grids, plates, struts), located between the

lemniscate attachment and the engine, creating the

same uneven velocity field at the fan inlet, as the

aircraft air intake on the flight mode of interest

(Grigor'ev, 2009).

At present, in connection with the development of

numerical methods for modelling gas-dynamic

processes and strength calculations, it has become

possible to model the influence of the inlet

nonuniformity on the workflow of the fan and the

engine. This will allow an assessment of the influence

of nonuniformity at the stage of calculations, without

manufacturing many prototypes. As a result, at the

initial design stage, the design variants that do not

work satisfactorily under the specified conditions will

be eliminated, which will significantly reduce the

time and cost of engine development.

In this paper, the authors aim to test the possibility

Popov, G., Baturin, O., Volkov, A., Kolmakova, D., Zubanov, V., Korneeva, A. and Novikova, Y.

The Inﬂuence of the Level of the Flow Path Blockage at the Inlet on the Fan Characteristics.

DOI: 10.5220/0007836502470254

In Proceedings of the 9th International Conference on Simulation and Modeling Methodologies, Technologies and Applications (SIMULTECH 2019), pages 247-254

ISBN: 978-989-758-381-0

247

1 - lemniscate tip; 2 - simulator; 3 - instruments controlling

the flow irregularity; 4 - the engine; 5 - thrust measuring

device; 6 - ejector tube

Figure 1: Engine test system with an inlet nonuniformity

simulator (Grigor'ev, 2009).

of conducting a computational study of the influence

of the inlet nonuniformity on the working process

(efficiency) of a fan of turbofan engine, reproducing

tests with an input simulator. There will also be given

a qualitative and quantitative assessment of the

influence of the inlet nonuniformity on the main

parameters of the working process of the engine fan.

2 TEST OBJECT

The object of the study was the fan of the NK-56

turbojet engine, developed at Kuznetsov, PJSC

(Samara, Russia) (JSC "Kuznetsov", 2019) for civil

aviation aircraft in the early 1980s, but not

commercialized. The main parameters of the NK-56

engine are given in Table 1 (Zrelov, 2002).

The appearance of the investigated fan is

presented in Figure 3. The number of blades is 30.

Information about the geometry of the fan and some

of its test results (in the form of internal reports of the

company) was transferred by Kuznetsov, PJSC to

Samara National Research University (Samara

University, 2019) s part of joint research.

Table 1: The main parameters of the NK-56 engine (Zrelov,

2002).

Thrust, kN

177

Pressure ratio

25.5

Gas temperature before the turbine, K

1571

Bypass ratio

4.8

Specific fuel consumption (M = 0, H = 0),

kg/N h

39.1

Specific fuel consumption (M = 0.8, N = 11

km), kg/kN h

63.75

Outer diameter, m

2.05

Weight with reverse, kg

3340

To conduct the research, a computational model

of the LPC of the NK-56 engine was created using the

Numeca FineTurbo software (NUMECA, 2008),

which includes a fan with an add stage.

Figure 2: Fan of the NK-56 engine.

Spalart-Allmaras and k-epsilon turbulence

models were used in the calculations.

The geometry of the computational domain was

created in accordance with the drawings submitted by

PJSC Kuznetsov. The computational area included

the inlet section, the rotor (RW) and the fan guide

vanes (GV), the bypass section and the second stage

of the LPC with the engine annular frame (Figure 3).

In constructing the model, the deformation of the

working blades from the forces acting on them was

considered. For this purpose, a preliminary

calculation of the fan workflow was carried out at a

rotor speed of 100%. Obtained gas loads acting on the

blades, were transferred to the ANSYS Mechanical.

It identified the deformation of the aerofoil, arising

under the action of centrifugal load and gas forces.

Then, information on the geometry of the deformed

aerofoil was transferred to the Numeca FineTurbo to

clarify the gas loads. In total, four such iterations were

performed. The criterion for the convergence of

iterations was the absence of a change in more than

1% of the deformed shape of the blade aerofoil during

the subsequent iteration.

The appearance of the resulting computational

model is shown in Figure 3.

The values of the total pressure p

=101325 Pa and

the total temperature T

=288.15 K were set as the

boundary conditions at the inlet to the computational

domain. The flow rate of the working fluid was set at

the outlet from each circuit. The ratio of mass flow

ratios at the outlet from each circuit advised the

SIMULTECH 2019 - 9th International Conference on Simulation and Modeling Methodologies, Technologies and Applications

248

Figure 3: Appearance of the design model of the fan with

add stages.

required bypass ratio (for the rotor speed n = 95% -

m = 4.9; for n = 100% - m = 4.8; for n = 105% -

m = 4.7).

Two mesh models were created for the

computational model: light (in total, 2.36 million final

volumes - Y+ is more than 7) and heavy (7.06 million

final volumes, Y + is more than 2).

4 VERIFICATION OF THE

COMPUTATIONAL MODEL

At the first stage of the study, the adequacy

assessment and validation of the created

computational model was carried out. For this, the

calculated characteristics obtained using various

stationary models differing in the turbulence model

and density of the finite volume mesh were compared

with the test results on the engine test bench provided

by Kuznetsov, PJSC. Due to the large time that has

passed since the tests, a detailed description of the

experimental setup and the error estimates of the test

data were not provided.

A total of 4 different computational models were

created:

No. 1 — Spalart-Allmaras (SA) turbulence model,

the number of final volumes is 2.36 million;

No. 2 — k-epsilon turbulence model, the number of

final volumes is 2.36 million;

No. 3 — Spalart-Allmaras (SA) turbulence model,

the number of final volumes is 7.06 million;

No. 4 — k-epsilon model of turbulence, the number

of final volumes is 7.06 million.

Comparison of the results obtained as a result of

stationary calculation with experimental data is

shown in Figures 4 and 5 and in Table 2.

Table 2: Comparison of the results of calculations obtained

using the considered numerical models with the data by

PJSC "Kuznetsov".

The deviation

of calculations

by the

numerical

model relatively

data by PJSC

"Kuznetsov"

Model No.

mesh 2.36 mln.

mesh 7.06 mln.

k-ε

Internal circuit

efficiency

0.5...2%

0...1%

1.5%

1...3%

External circuit

efficiency

3...6%

1.5...5

Mass flow rate

of the internal

circuit

6% less

less

2%

less

Mass flow rate

of the external

circuit

2%

𝜋

𝑐

∗

of the

internal circuit

Higher

by 0.05

 by

0.03

Coinci

des

 by

0.02

𝜋

𝑐

∗

of the

external circuit

 by

0.07...

0.13

 by

0.05...

0.12

 by

0.02

 by

0.03...

0.1

Analysing the data presented in Figures 4...5 and

Table 2, we can come to the following conclusions:

 none of the created computational models show

complete agreement with the experimental

characteristics in the whole considered range of

parameters;

 all computational models show significantly

overestimated values of the pressure ratio in the

external circuit, but at the same time, they well

predict the characteristics of the internal circuit;

 all computational models show overestimated

efficiency values, especially in the external circuit

(the difference reaches 6%);

 the value of the working fluid mass flow rate, at

which the maximum efficiency is achieved, for

the external circuit is in good agreement with the

data of the design calculation, while for the

internal circuit, the resulting flow rate is usually

underestimated by 2...4%;

 the smallest discrepancy for the external circuit is

observed at a high rotor speed (n = 105%), and for

the internal circuit - at the small (n = 95%);

 considering the real deformation of the blade with

the help of coupled simulation of strength and gas

dynamics allows to reduce the quantitative

discrepancy between the calculation data and

design data. Qualitatively, the nature of the

calculated characteristics of the LPT does not

change.

The Inﬂuence of the Level of the Flow Path Blockage at the Inlet on the Fan Characteristics

249

a) For external circuit

b) For internal circuit

- experimental data; - model No.1;

- model No.2; - - model No.3;

- model No.4

Figure 4: Comparison of the pressure characteristics

obtained using the created computational models with

experimental data.

Of all the models considered, the best match with

the data of the design calculation is shown by model

No. 3 (fine mesh and Spalart-Allmaras turbulence

model). It is accepted as the final for further research

of the workflow in the fan blade passages. The Mach

number contours obtained using this model are shown

in Figure 6.

5 UNSTEAY (NLH)

COMPUTATIONAL MODEL

At the second stage of the study, based on the created

and verified computational model of the workflow of

the LPC of the NK-56 engine, a computational model

was created to study the influence of the inlet

nonuniformity on the working process of its fan.

The modified model was created in such a way as to

meet the conditions of testing a fan with a simulator

of the inlet nonuniformity on the test benches of PJSC

Kuznetsov. There, the nonuniformity is modelled by

extending the interceptor into the duct between the

lemniscate and the inlet to the fan (Figure 7). During

a) For external circuit

b) For internal circuit

- experimental data; - model No.1;

- model No.2; - - model No.3;

- model No.4

Figure 5: Comparison of the efficiency characteristics

obtained using the created computational models with

experimental data.

the tests, the intensity of the inlet nonuniformity was

regulated by the depth of the interceptor extension to

the flow part of the duct.

Due to the highly variable nature of the flow at the

fan inlet, the task should be solved in a transient

statement. In this case, since the nonuniformity

generated by the interceptor has a different intensity

of direction of the velocity vectors around the

circumference of the RW, the assumption about the

periodicity of the flow cannot be accepted, and the

computational model must contain all the blade

passages.

Due to the highly variable nature of the flow at the

fan inlet, the task should be solved in a transient

statement. In this case, since the nonuniformity

generated by the interceptor has a different intensity

of direction of the velocity vectors around the

circumference of the RW, the assumption about the

periodicity of the flow cannot be accepted, and the

computational model must contain all the blade

passages.

The solution to the problem of studying the

influence of inlet nonuniformity on the workflow of

the LPC using a transient simulation of a full circle

SIMULTECH 2019 - 9th International Conference on Simulation and Modeling Methodologies, Technologies and Applications

250

model (containing all passages of LPC of the turbofan

engine) requires exorbitant computer resources, and

cannot be successfully done in a reasonable time

using available computer equipment. For this reason,

it was assumed to conduct the research using the

method of nonlinear harmonic analysis (NLH)

(Vilmin et al., 2013), which allows to obtain

nonstationary flow patterns several times faster than

using transient simulation. The method allows to

obtain non-stationary flow fields by means of

decomposition of periodic oscillations in Fourier

based on a preselected number of harmonics, usually

associated with the transmission frequencies of the

blades of the turbomachine configuration and their

multiples. At the same time, only one blade passage

is required for analysis. This approach allows to

obtain pictures of dynamic processes by 2 orders of

magnitude faster than with transient calculation

(NUMECA International The Nonlinear Harmonic

module, 2019).

The number of harmonics used in nonlinear

harmonic analysis is 3. A series of calculations was

also carried out with the number of harmonics equal

to 7. The results obtained differed little from the data

obtained with 3 harmonics, but in the case of 7

harmonics the solution process was significantly less

stable.

The NLH method could not be applied to the

required number of blade rows and when the

compressor was operating simultaneously for 2

circuits. For this reason, the geometry of the LPC was

significantly simplified: the separator of the contours

was eliminated (the task became single-circuit). The

computational domain contained only a RW and a

model GV (“lengthened up” GV of the fan of the

internal circuit) (Figure 7) simulating the effect of

downstream elements on the rotor.

An input section imitating the engine inlet channel

with an interceptor was attached to the inlet boundary

of the fan domain (Figure 8). Its geometry was

created in the Numeca IGG software. Several variants

of its geometry were created, differing in the length

by which the interceptor was extended. The total

number of finite volumes of the computational model

shown in Figure 9 is 4 mln. The mesh models of the

RW and GV domains were made with the settings

corresponding to the Model No. 3 of the LPC (see

above).

Hub section of the fan

Shroud section of the fan

The averaged values in the meridional section

Figure 6: Contours of Mach numbers in relative motion in

the fan at the operating point at n = 100%.

Figure 7: Geometry of the simulated variant of the

interceptor in the inlet device when testing the NK-56

engine.

The Inﬂuence of the Level of the Flow Path Blockage at the Inlet on the Fan Characteristics

251

Figure 8: Simplified geometry of the NK-56 engine fan for

conducting research on the effect of inlet nonuniformity on

its workflow.

Figure 9: Computational model for studying the influence

of inlet nonuniformity on the fan workflow.

6 DISCUSSION OF THE RESULTS

Figure 10 shows the contours of Mach numbers in

relative motion when the interceptor extends into the

flow part so that it covers 6.8% of the flow part of the

duct, in two mutually perpendicular planes passing

through the engine axis. It shows that in front of the

interceptor a zone of flow deceleration is formed, and

behind it a developed separation zone reaches the

inlet of the engine, located at a distance more than the

size of its outer diameter (i.e. by more than one

calibre). In this case, part of the flow entering the duct

opposite the interceptor is redirected to the axis of the

engine, causing local flow acceleration, and changing

the flow structure there. That is, the injecting the

interceptor affects not only the structure immediately

near it, but the rest of the duct to a depth of more than

half the diameter of the engine.

These circumstances lead to the emergence of

significant inhomogeneity over the cross section of

the pressure field at the fan inlet (Figure 11). It can be

seen that it is formed behind the interceptor, and, with

an increase of the interceptor extension, an area of

reduced total pressure grows behind it. In this case,

Figure 11 confirms that with an increase in the

overlapping area of the duct, the uniformity of the

pressure field is disturbed over the entire cross

section. And the higher the extension, the greater the

level of unevenness.

The result of a quantitative assessment of the

effect of the interceptor extension level on the

nonuniformity of the pressure field is shown in Figure

12. There, the criterion of non-uniformity applies a

value equal to the ratio of the minimum total pressure

in the cross section to its maximum value. The

smaller this value, the greater the uneven flow. As can

be seen from Figure 12, the flow nonuniformity with

increasing overlap of the flow path by the interceptor

increases linearly. Moreover, when 13% overlap, the

nonuniformity reaches 50%.

a) Section plane No.1 passing through the axis of the

engine perpendicular to the interceptor

b) Section plane No.2 passing through the axis of the

engine perpendicular to plane No. 1

Figure 10: The calculated contours of the change in the

Mach number in relative motion in the “inlet duct+ fan”

system when the interceptor is extended so that it blocks

6.8% of the flow-part of the duct.

The above-described phenomena lead to the fact

that the conditions at the inlet to each blade passage

and, correspondingly, the flow structure there are

unique (Figure 13), disrupting the interaction of

adjacent passages and reducing the pressure ratio,

efficiency, stability margins and air flow through the

fan.

The effect of nonuniformity on the fan workflow

parameters at a rotation frequency of n = 100% can

be estimated from the evolution of the characteristics

as the inlet device is overlapped by the interceptor

(Figure 14). It can be seen from the above data that

SIMULTECH 2019 - 9th International Conference on Simulation and Modeling Methodologies, Technologies and Applications

252

a) No interceptor

b) Interceptor overlaps

4,9% of passage area

c) Interceptor overlaps

6,8% of passage area

d) Interceptor overlaps

8,9% of passage area

e) Interceptor overlaps

11.2% of passage area

f) Interceptor overlaps

13,5% of passage area

Figure 11: Transformation of the total pressure fields at the

fan inlet flange at different levels of extension of the

interceptor (at the top) to the flow part of the supply duct.

Figure 12: Influence of the overlap level of the input duct

by the interceptor on the unevenness of the total pressure

field at the fan inlet.

with the increase in the part of the duct blocked by the

interceptor, all compressor parameters deteriorate:

the working fluid mass flow rate, the pressure ratio,

and the efficiency decrease. In addition, the range of

mass flow rate between the modes of surge and choke

is also reduced, pressure lines become more vertical.

This signals a decrease in the stability of the

compressor.

Figure 13: Contours of Mach numbers in relative motion in

the peripheral part of the fan (height of 98%) opposite the

installation site of the interceptor.

Pressure characteristics

Efficiency characteristics

Figure 14: Comparison of fan characteristics at n = 100% at

different values of the interceptor extension into the flow

part.

A quantitative assessment of the influence of the

overlap of the flow part of the input duct on the

parameters of the compressor workflow, obtained

from the analysis of Figure 14, is shown in Figure 15.

As can be seen, the increase in the level of the

interceptor extension degrades the parameters not

linearly, but by parabolic dependence. The least

change is in the mass flow rate of the working fluid

(with a decrease in the area of the inlet duct by 10%,

it is reduced by 3%). The difference between the

The Inﬂuence of the Level of the Flow Path Blockage at the Inlet on the Fan Characteristics

253

values of mass flow rate at surge and choke changes

most of all (with a decrease in the duct area by 10%,

it is reduced by 40%). If the duct area is reduced by

10%, the efficiency of the compressor decreases by

11% (rel.), and the pressure ratio - by 9%.

Figure 15: Changing the main parameters of the compressor

process at different levels of overlapping the inlet section.

7 CONCLUSIONS

As a result of the performed work, the complex

scientific and technical problem of a reliable

computational study of the influence of the inlet

nonuniformity on the working process of a fan of a

turbojet bypass engine was solved. The solution to

this problem is hampered by the fact that the process

under study is transient and leads to the fact that

different fan blade passages operate in different

conditions. Such a problem should be solved in a

transient setting using a full circle model. However,

such a computational model, in addition to the

enormous calculation time inherent in transient

problems, contains the number of finite elements in

the hundreds of millions, which, due to the limited

possibilities of the computer equipment, is

unacceptable for the authors of the paper.

To solve this problem, the authors developed and

verified a numerical computational model of the

workflow of an inlet device, retractable interceptor

(an inlet nonuniformity generator) and a turbofan fan

stage using NLH approach.

In the future, the results are planned to be

transferred to the module of the strength calculation

and to evaluate the effect of inlet nonuniformity on

the static and dynamic loads on the fan blades.

ACKNOWLEDGEMENTS

This work was supported by the Russian Federation

President's grant (project code МК-3168.2019.8).

REFERENCES

Grigor'ev, V.A., 2009. Ispytaniya aviatsionnykh dvigatelej

(Aircraft engine testing). Mashinostroenie, Moskow.

JSC "Kuznetsov", 2019. Accessed January 10.

http://www.kuznetsov-motors.ru/en.

NUMECA, 2008. User Manual AutoGrid5 Release 8.4,

NUMECA.inc., Belgium.

NUMECA International The Nonlinear Harmonic module,

2019. Accessed January 10.

https://www.numeca.com/product/nlhmethod.

Samara University, 2019. Accessed January 10.

https://ssau.ru/english/.

Venediktov, V.D., Granovsky, A.V., 1990. Atlas

eksperimental'nykh kharakteristik ploskikh reshetok

okhlazhdaemykh gazovykh turbin (The Atlas of

Experimental Performances of Cooled Gas Turbine

Blade Cascades), CIAM.

Vilmin, S., Lorrain, É.,Tartinville, B., Capron, A. and

Hirsch, Ch., 2013. The nonlinear harmonic method:

from single stage to multi-row effects Int J of

Computational Fluid Dynamics 27(2): 88-99.

Zrelov, V.A., 2002. Otechestvennye GTD. Osnovnye

parametry i konstruktivnye skhemy. CHast' 2

(Domestic GTE. Key parameters and design schemes.

Part 2). SGAU, Samara.

SIMULTECH 2019 - 9th International Conference on Simulation and Modeling Methodologies, Technologies and Applications

254