Strategy for Reduction of the Negative Effects of Circumferential

Flow Irregularity in Axial Compressor

Aleksandr Ermakov, Grigorii Popov, Daria Kolmakova and Vasilii Zubanov

Insitute of Engine and Power Plant Engineering, Samara University, Samara, Russia

Keywords: Intermediate Pressure Compressor, Flow Irregularity, Vibrations, Annular Frame, CFD, Optimization.

Abstract: Gas flow nonuniformity is one of the main sources of rotor blade vibrations in the gas turbine engines. Often,

the source of the flow circumferential nonuniformity are the annular frames, located in the flow passage of

the engine. This leads to the increased dynamic stresses in rotor blades and consequently to the blade

destruction. The goal of the research was to find an acceptable method of reducing the level of gas flow

nonuniformity. Thus, this study gives the ideas about methods of improving the flow structure in gas turbine

engine. It allows the selection of the most suitable method for reducing gas flow nonuniformity.

1 INTRODUCTION

Gas flow circumferential non-uniformity is one of the

main sources of oscillation of turbomachinery rotor

blades (Ivanov, 1983; Vorob'ev., 1988; Cohen etc.,

1996). The flow circumferential irregularity is caused

by various factors: the axial asymmetry at the

turbomachinery inlet; the presence of guide vanes

(GV) and nozzle blades (NB) in the turbomachinery

flow path; the influence of the structural elements

(struts and bearings); bypass system operation and air

bleeding; the operation of fuel nozzles of combustion

chambers; the deformation of power casings; the

buckling of the flame tubes of combustion chambers.

Several methods are used in practice to reduce

high dynamic stresses: dynamic methods, gas

dynamic methods and technological (Figure 1)

(Ivanov, 1983; Birger etc., 1993; Logan, 2003).

Each method has its advantages and

disadvantages, as well as the range of applicability.

Methods that suggest design changes are impossible

to apply to already designed engines.

However, it is possible to test the effectiveness of

the proposed methods of dynamic stresses reduction

without resorting to expensive and long experimental

studies. For these purposes, it is possible to use CFD

(Respondek, 2010) of the work processes occurring

in the engine as a preliminary verification step.

The purpose of the research was to evaluate the of

various methods and to indicate the most appropriate

approaches for solving specific problems.

Figure 1: Classification of different methods to reduce high

dynamic stresses in rotor blades.

2 MOTIVATION

The motivation for this research was the real-life

destruction problem of the fifth stage rotor blades of

the intermediate pressure compressor GTE NK-36ST

(RW5 in Figure 2).

The reason of rotor wheel blades’ destruction is

the flow circumferential irregularity, which arises

from the presence of middle annular frame struts in

the GTE flow path. Struts have different profile

thickness and distributed with different angular

displacements relative to each other (Figure 3). Struts

are the cause of the appearance of local zones of high

pressure, which propagate upstream through the

cascade of GV5 fixed blades. Passing through these

zones, rotor wheel blades of the fifth stage experience

dynamic effect, leading to their forced oscillations

and high dynamic stresses (Figure 4).

144

Ermakov, A., Popov, G., Kolmakova, D. and Zubanov, V.

Strategy for Reduction of the Negative Effects of Circumferential Flow Irregularity in Axial Compressor.

DOI: 10.5220/0006887501440150

In Proceedings of 8th International Conference on Simulation and Modeling Methodologies, Technologies and Applications (SIMULTECH 2018), pages 144-150

ISBN: 978-989-758-323-0

Figure 2: The diagram of investigated compressor flow path.

Figure 3: The appearance of middle annular frame.

Because of this, the fifth rotor wheel (RW) blades

in the original design have anti-vibration segments.

However, such a solution has several shortcomings

(Cumpsty, 2004).

a) b)

Figure 4: Mach number (a) and pressure (b) fields.

Thus, the aim of the work was the search of an

alternative approach for reduction of dynamic

stresses level in the fifth rotor wheel blades of

intermediate pressure compressor (IPC). For further

research, the fifth-stage rotor blades were modeled

without anti-vibration segments.

3 METHOD OF DYNAMIC

STRESSES CALCULATION

At the initial stage of the work, the method of

dynamic stresses calculation in the fifth rotor wheel

blades of IPC was developed (Figure 5).

Figure 5: Scheme of method for calculation compressor

rotor blades forced oscillations.

Stage 1. CFD-simulation of IPC sector model on

the main modes of GTE operation. The sector is one

blade channel of blade row, the lateral boundaries of

which are imposed by boundary conditions of

periodicity. The IPC sector model consists of the

intermediate frame domain, the domains of all IPC

blade rows, the middle frame domain and the inlet

guide vane of high pressure compressor (HPC IGV).

When transmitting the flow parameters between

domains, the Mixing Plane interface is used

(NUMECA, 2008), averaging the flow parameters in

the circumferential direction. The purpose of this

calculation is to determine the flow parameters

distribution (total pressure, total temperature, flow

angle) along the flow path height in the cross section

behind the RW of the IPC fourth stage, as boundary

conditions for subsequent calculations.

Stage 2. The CFD modelling of full circle model

will allow to detect the presence of gas flow

circumferential irregularity, which arises from the

nonuniformly located struts of different thickness

along the circumference.

This model consists of GV4, RW5, GV5 of IPC,

middle annular frame and HPC IGV domains. Gas

dynamic load acting on all blades of the fifth stage

rotor wheel is determined in this calculation.

Creation the grid of finite volumes of the full

circle model was carried out in AutoGrid5

(NUMECA, 2008). Parameters of finite volume grids

and computational models tuning were chosen

Strategy for Reduction of the Negative Effects of Circumferential Flow Irregularity in Axial Compressor

145

according to the recommendations given in papers

(Matveev etc., 2014; Popov etc., 2014). The

computational models were verified in studies

(Kolmakova etc., 2014; Ermakov etc., 2014) in

assessing the vibration level of rotor blades. This

indicates the reliability of the calculations performed

in the study described in this article.

Total pressure and temperature, and inlet flow

direction at the inlet, which were obtained from the

sector model calculation at the previous step

(Ermakov etc., 2014); static pressure at the outlet of

the computational model were set as the boundary

conditions for the full circle model.

To transfer parameters between the rotating and

fixed domains, the standard Frozen Rotor interface

was used without their averaging in the

circumferential direction (ANSYS, 2017).

Stage 3. Determination the values of exciting

harmonics amplitudes of the load, as well as dynamic

stresses arising in the RW5 blade. For this purpose, in

the APDL programming language, the program that

imports the load distribution from the finite volume

blade row model in CFX into the finite element blade

row model in Ansys Mechanical APDL, were written.

The gas load, which has a complex character of the

distribution along the circumference, was represented

as a sum of the harmonics components. Each

harmonics component represents a load waves chain

that locate along the flow path circumference. The

load rotates at angular velocity. Thus, the

circumferential stationary inhomogeneity of the gas

flow for a rotating rotor wheel is equivalent to the

action of an infinite set of exciting harmonics, each of

which is a load chain of backward running waves

performing harmonic oscillations in time.

The load is expanded in Fourier series along the

similar blades’ nodes and is represented in the form

of a backward running wave. Campbell diagram is

used to determine the most dangerous operating

modes of the engine, and the most dangerous

harmonics. The dynamic stresses are calculated only

with the most dangerous harmonic.

4 METHODS OF REDUCTION

THE DYNAMIC STRESSES

4.1 Detuning the Rotor Blade from the

Dangerous Harmonics

IPC testing without shroud segment of RW blades

(the blade without shroud segment is a standard blade

airfoil without anti-vibration segment) was carried

out by JSC "Kuznetsov".

The experimental investigations were carried out

using a cold test rig for the IPC with standard pressure

at the inlet and middle annular frame at the outlet.

Two types of the tests were conducted. Strain gauges

were installed at five rotor blades in the first case and

at 12 rotor blades in the second test case (total number

of rotor blades was 84). The maximum rotor speed,

during the tests was the speed at which resonance

occurs with the most dangerous harmonics.

These tests revealed the destruction of the IPC

fifth stage rotor blades at resonance with the 7th and

12th harmonics. Figure 6 shows the histogram of the

exciting load amplitude distribution, calculated using

the method described in clause 3. The average value

on the blade surface that is calculated by adding the

amplitudes of the harmonics in each node and

dividing by the number of nodes is referred to the

mean amplitude. Similar calculations are performed

for all blades in the fifth blade row.

As can be seen from Figure 6, the 12th harmonic

has the highest amplitude (harmonics from 7 to 14 fall

into the region of GTE operating modes).

Figure 6: Level of the amplitudes of the exciting harmonics.

An attempt to turn out from the most dangerous

7th harmonic, and to increase the flexural strength of

the airfoil blade to reduce the dynamic stresses at

resonance with the 12th harmonic, was made at JSC

"Kuznetsov". For this purpose, a special Shvarov's

profile was used. The shape of the suction side

surface coincides with the initial profile; however, the

pressure side surface is convex. This profile shape

allows to significantly reduce the flexural stresses in

the airfoil. Finite element harmonic analysis showed

a double decrease in the dynamic stresses of blade

airfoil at the resonance with the 12th harmonic

(Figure 7). Increase of IPC efficiency was about 0.5%

for the compressor variant with Shvarov’s profile on

SIMULTECH 2018 - 8th International Conference on Simulation and Modeling Methodologies, Technologies and Applications

146

the rotor blades compare to the compressor with

shrouded blades.

Figure 7: Equivalent stresses at the resonance in RW5 blade

(original is on the left, Shvarov's profile is on the right).

Figure 8 shows the Campbell diagram for the 5th

rotor wheel with the original unshrouded rotor blade

and with blade with Shvarov's profile. Thus, the use

of blade with the special Shvarov's profile allows to

significantly reduce the dynamic stresses in the

airfoil, as well as to turn out from dangerous

operating modes. For the blades with Shvarov’s

profile the strongest seventh harmonic is out of region

of the engine operating modes (Figure 6).

Figure 8: RW 5 resonance diagram (--- Blade with

Shvarov's profile, - original blade).

Using the blades with special Shvarov's profile

allow to reduce the level of dynamic stresses by 1.8

times. This result was confirmed experimentally

during strain-gauging at JSC "Kuznetsov". The

stresses were measured using heat-resistant strain

gauge, located on the surface of the blade pressure

side. The single amplitudes were estimated. However,

this reduction was not enough to eliminate the

problems of blade destruction. It is necessary to

influence not only the blades shape, but also the gas

flow circumferential irregularity. In further

researches, we used the blade with Shvarov's profile.

4.2 Design Change

One of the methods of reduction the gas flow

circumferential irregularity in front of the struts is the

use of guide vane with a different pitch along the

circumference and different stagger angle of the GV

blades (Sladojević etc., 2007; Yang etc., 2012). This

method allows to redistribute the flow between

various blade channels and to regulate the position of

high pressure zones (Saren, 1984).

Parametric model of IPC fifth stage GV was

created to implement the method. All GV blades were

divided into 7 groups, according to the location of the

7 struts. Angles and pitch changes occurred within

each group. The key factor was the minimum number

of variable blades.

For the GV5 blades located in the symmetry plane

of the strut, the first and the last blades in the group,

the stagger angle does not change. The change of

blades stagger angles within the groups was carried

out according to a linear law. The blades located on

opposite sides of the strut symmetry plane rotate in

opposite directions, relative to the initial position

(Figure 9). The blades located closer to the strut rotate

a larger angle. If the stagger angle increases, the sign

"+" stands in front of the angle, the sign "-" means

decreasing stagger angle. All angles are measured

from the leading edge plane.

In the case of pitch change, the number of

changeable blades was not limited. Different pitch

was set in the range -0.35 ... + 0.35 of the pitch. The

sign "-" means that in the strut area the pitch between

the blades decreased, and the sign "+" - that the pitch

increased. The number indicates the maximum

increase (decrease) of the pitch between the blades in

the group in relative values from the nominal pitch

with uniformly arranged blades. When different blade

pitches are introduced in groups, the position of the

extreme blades groups did not change. The law of

pitch change is also linear.

In total, 10 variants of the GV5 configuration with

different angles and pitches were tested. Table 1

shows same of them. Variant 1 corresponds to the

original version of the IPC without an anti-vibration

segment on the RW5 blades with Shvarov’s profile.

Table 1: The results of the parametric IPC model calculation.

Variant

number

Parameter of different stagger angles,

maximum stagger angles (number of blades)

for the groups:

Parameter of alternating blade pitch for the

groups:

Number of

variable

blades

Dynamic

stresses

MPa

2, 5, 6 1, 3, 7 4 (3) 1, 7 2, 6 3 4 5

1 0 0 0 0 0 0 0 0 0 86.7

4 3 (6) 3 (6) 3 (8) 0.3 0.3 0.3 0.3 0.3 42 31.877

9 6 (2) 6 (2) 6 (2) 0 0 0 0 0 14 46.737

Strategy for Reduction of the Negative Effects of Circumferential Flow Irregularity in Axial Compressor

147

Figure 9: The diagram of GV blades rotation.

The analysis of the obtained results has shown that

it is possible to achieve a significant reduction in the

level of the exciting harmonics acting on the RW5

rotor blades, and, consequently, the level of dynamic

stresses with the help of different angles and pitches.

For example, variant 9 allows to reduce the

dangerous harmonic amplitude by almost 2 times. In

this variant of the GV5 configuration the number of

changeable blades is 14 (a total of 76 blades). Variant

4 allows to obtain the greatest dynamic stresses

decrease, however the number of changeable blades

will amount to 42, which is more than half of the total

number.

Figure 10 shows the mean of average exciting

harmonics amplitudes for the original compressor

version, GV variant 4 and GV variant 9.

Figure 10: The mean of exciting harmonics amplitudes

acting on RW5.

From a gas dynamic point of view, the positive

effect of different angles and pitches on the dynamic

stresses value can be explained by the following. Due

to redistribution of the flow, its structure becomes

more homogeneous, with smaller high pressures

areas. In addition, the pressure peaks values behind

the 5th rotor wheel is also markedly reduced (Figure

11).

Thus, the introduction of GV of different angles

and pitches can reduce the gas flow circumferential

irregularity and dangerous harmonics amplitude.

However, the use of this method leads to a

complication of production technology. GV blades in

this case will have to be manufactured by

nonuniversal technologies. In addition, due to the

change in the blades pitch, it will be necessary to

develop a new technology for grooving.

Figure 11: Graphs of relative static pressure variation for

basic compressor variant and variant 9.

4.3 Modification of Frame Design

Experimental 13-struts annular frame was made by

JSC ''Kuznetsov'' to reduce the resonant stresses level

in the IPC fifth stage. Variable stresses level at

resonances with the strongest 12th harmonic in the

test with 13-struts annular frame decreased

approximately 2 times. Experimental 13-struts

annular frame calculation using the method described

above confirmed the decrease in the average

amplitude of the 12th harmonic (Figure 12) by 1.8

times. This fact is another confirmation of the

reliability of the developed methodology.

Figure 12: Comparison of excitation amplitudes of

harmonics of the basic 7-struts and 13-struts annular frame

obtained in experiment and CFD calculation.

Comparison of normalized mean harmonic

amplitudes of the basic 7-struts annular frame and

experimental 13-struts annular frame obtained in

experiments and by CFD is shown in Figure 13. The

calculation results were converted into relative values

by considering the mean amplitude of 12th harmonic

of base 7-struts frame as 1 to compare calculation data

SIMULTECH 2018 - 8th International Conference on Simulation and Modeling Methodologies, Technologies and Applications

148

with experimental results.

Experimental 13-struts annular frame proposed by

JSC "Kuznetsov" specialists allowed to decrease the

resonant stresses in the fifth stage rotor blades and to

exclude the anti-vibration segment. However, in this

annular frame in addition to increasing the struts

number, their thickness was reduced. This is critical

towards to the struts through which the various engine

systems of communications pass. Therefore, the use

of this frame on the engine is impossible.

For this reason, it was decided to develop a new

design of 13-struts annular frame with saving the

position and required cross-sections in the struts No.

1, 2, 4, 5 of 7-struts with engine systems.

When designing the new annular frame, all the

previous modifications of the compressor were not

considered (different stagger angles and pitches)

except the implementation of Shvarov’s profile on the

rotor blades of the fifth stage.

It was agreed that the development of new annular

frame design would be carried out using optimization

methods implemented in the IOSO software package

(Egorov etc., 2002; Sigma Technology, 2017). Since

the use of large-scale 3-D model during optimization

requires significant computational resources and

time, it was decided to use surrogate struts models.

With this approach, each strut is represented as a

pressure peak, which the strut causes in the axial gap

between RW5 and GV5. The shape of the peaks is

determined by the geometric parameters of the struts,

and their location in the circumferential direction is

the arrangement of the corresponding struts.

Additional research was carried out (Kolmakova

etc., 2015) using struts parametric models (Figure 13)

to determine the relationship between the struts

geometric dimensions and the peaks’ shape. These

models allowed to change the struts design taking into

account the existing limitations.

As a result, the obtained data array allowed to

derive equations for each strut type, reflecting the

dependence of pressure peaks’ height on the position

of the strut leading edges relative to the outlet edges

of GV blades located upstream.

The found equations after verification were used

to search for a new annular frame design using IOSO.

Figure 13: Strut parametric model.

The optimization problem is the following. Struts

position and the values of the shift of the strut leading

edges are used as variable parameters. Optimization

criteria were to reduce the amplitude values of 12th

and 10th harmonics (the second largest after 12th).

Pareto set was obtained which is the compromise

between the decrease in 12th and 10th harmonics. The

point of the Pareto set, which gives the maximum

decrease in the 12th harmonic amplitude, was chosen

as a point for further research.

The annular frame configuration corresponding to

the selected point of Pareto set is shown in Figure 14.

For this configuration, a CFD simulation was

performed using the method described in Section 3 of

this article. As a result, the amplitude of the 12th

harmonic was reduced (Figure 15).

Figure 14: Optimized 13-struts annular frame.

Thus, the change in the annular frame structure

makes it possible to significantly reduce the exciting

harmonics amplitudes.

Figure 15: Comparison of exciting harmonics amplitudes of

the original 7-struts, 13-struts frame by JSC ''Kuznetsov''

and optimized 13-struts annular frame.

5 CONCLUSIONS

Several approaches to increase the dynamic strength

of GTE compressor rotor wheels were developed.

Each of the proposed approaches, possessing a set of

advantages and disadvantages, allowed to achieve the

set goal which was to reduce the dynamic stresses

level in rotor blades. Also, the following aspects can

be noted as the obtained research results:

Strategy for Reduction of the Negative Effects of Circumferential Flow Irregularity in Axial Compressor

149

1) The calculation method of dynamic stresses

caused by the flow circumferential irregularity

is developed;

2) The use of special Shvarov's profile on the

RW5 blades makes it possible to reduce the

dynamic stresses by almost 2 times;

3) It is revealed that the change in the angular

position of the GV5 blades makes it possible

to reduce the amplitude of the dangerous 12th

harmonic by 2.3 times with the change of 42

blades and by 1.8 times with the change of 14

blades;

4) A technique has been developed to optimize

the struts angular position without using

complex CFD calculations using surrogate

struts models;

5) Different design variants of the engine middle

annular frame different number of struts are

obtained. With an increase in the number of

struts to 13, the amplitude of the most

dangerous 12th harmonic is reduced by 2.4

times;

6) For all annular frame variants, the shape and

angular position of the struts through which

the engine systems pass is saved.

ACKNOWLEDGEMENTS

The work was financially supported by the Ministry

of education and science of Russia in the framework

of basic part of government assignment (Project

number 2496) and in the framework of the

implementation of the Program of increasing the

competitiveness of SSAU among the world’s leading

scientific and educational centers for 2013-2020

years.

REFERENCES

Ivanov, V. P., 1983. Kolebaniya rabochego kolesa

turbomashin (Vibrations of turbomachinery impeller),

Mashinostroenie, Moskow, 224 p.

Vorob'ev, Yu. S., 1988. Kolebaniya lopatochnogo apparata

turbomashin (Vibrations turbomachinery blading),

Kiev, Nauk. dumka, 224 p.

Cohen, H., Rogers, G. F. C., Saravanamuttoo, H. I. H.,

1996. Gas Turbine Theory, 4-th edition, Longman

Group Limited, 455 p.

Logan, E. Jr., 2003. Handbook of Turbomachinery, CRC

Press; 2 edition, 880 p.

Respondek, 2010, "Numerical simulation in the partial

differential equation controllability analysis with

physically meaningful constraints". In Mathematics and

Computers in Simulation 81, pp. 120–132

Birger, I. A., Shorr, B. F., Iosilevich, G. B., 1993, Raschet

na prochnost' detalei mashin (Strength analysis

machinery parts), Mashinostroenie, Moskow, 640 p.

Cumpsty, N. A., 2004. Compressor Aerodynamics, Krieger

Publishing Compan, Edition 2, 552 p.

NUMECA, User Manual AutoGrid5 Release 8.4,

NUMECA.inc., Belgium, January 2008.

Matveev, V. N., Popov, G. M., Goryachkin, E. S.,

Smirnova, Y.D., 2014. “Effect of Accounting of Air

Bleed from the Flow Passage of the Multi-Stage Axial

Low Pressure Compressor on its Design

Performances”. In Research Journal of Applied

Sciences, 9(11), pp. 784-788.

Popov, G., Goryachkin, E., Baturin, O., Kolmakova, D.,

2014. “Development of Recommendations on Building

of the Lightweight Calculation Mathematical Models of

the Axial Turbines of Gas Turbine Engines”. In

International Journal of Engineering and Technology,

6(5), pp. 2236-2243.

Kolmakova, D., Popov, G., Shklovets, A., Ermakov, A.,

2014. “Techniques and Methods to Improve the

Dynamic Strength of Gas Turbine Engines Compressor

Rotor Wheels”. In Proceedings of the ASME 2014 Gas

Turbine India Conference GTINDIA2014, Paper No.

GTINDIA2014-8203.

Ermakov, A. I., Shklovets, A. O., Popov, G. M.,

Kolmakova, D. A., 2014. “Investigation of the Effect of

the Gas Turbine Compressor Supports on Gas Flow

Circumferential Nonuniformity”. In Research Journal

of Applied Sciences, 9(10), pp. 684-690.

ANSYS® Release 12.0, Help System, Ansys CFX-Solver

Modeling Guide, ANSYS, Inc.

Sladojević, I., Sayma, A. I., Imregun, M., 2007. “Influence

of stagger angle variation on aerodynamic damping and

frequency shifts”. In Proceedings of the ASME Turbo

Expo, Vol. 5, pp. 683–700.

Yang, Y., Yang, A., Dong, R., Chen, E., Dai, R., 2012.

“Influence of stagger angle on aerodynamic sound

performance of compressor cascade”. In Hangkong

Xuebao/Acta Aeronautica et Astronautica Sinica,

33(4), pp. 588-596.

Saren, V., 1984. Flow around irregular lattice of plates

placed in front of the cylinder. Technical Report

(Moscow: Central Institute of Aviation Motors) p 36.

Egorov, I. N., Kretinin, G. V., Leshchenko, I. A., Kuptzov,

S. V., 2002, "IOSO Optimisation Toolkit - Novel

Software to Create Better Design". In 9th AIAA/ISSMO

Symposium on Multidisciplinary Analysis and

Optimisation, 04 - 06 Sep. 2002, Atlanta, Georgia.

"Sigma Technology". Novel Optimization Strategy –

IOSO, http://www.iosotech.com/

Kolmakova, D., Popov, G., 2015. “Methods of improving

the axial compressor flow passage to reduce the flow

circumferential nonuniformity”. In Proceedings of the

ASME 2014 Gas Turbine India Conference

GTINDIA2015, Paper No. GTINDIA2015-1276.

SIMULTECH 2018 - 8th International Conference on Simulation and Modeling Methodologies, Technologies and Applications

150