Reducing the Hydraulic Resistance in the Inlet Device

of the Gas Turbine Unit

Oleg Baturin, Daria Kolmakova

, Grigorii Popov

, Vasilii Zubanov

and Yulia Novikova

Samara National Research University, Samara, Russia

Keywords: Hydraulic Losses, Work Shaft, Gas Pumping Unit, Inlet Unit.

Abstract: The paper presents the results of numerical simulation of air flow through a modernized variant of the inlet

filter unit (IFU) of the gas compressor unit GPA-Ts-16. A feature of the IFU design is that to reduce the load

on the filter unit, it is proposed to be as compact as possible, which determines its complex shape. The goal

of the study is to study the hydraulic losses and to develop the measures to reduce them, since it is found that

every 100 Pa of losses in the inlet unit increases the consumption of fuel gas by 2.5 kg/h or reduces the engine

power by 10.5 kW. Calculations of hydraulic losses in IFU are carried out for cases of absence or presence of

wind with a velocity from 0 to 35 m/s, blowing from 5 main directions (0, 45, 90, 135, 180). Studies are

also carried out on the effect of the weather shield shape, presence of baffles under it, and the rack in the shaft

on the hydraulic losses. As a result of the research, recommendations are provided for designing (changing

the shape) of the inlet filter unit that eventually allow to propose a design that will reduce the hydraulic losses

in IFU by 15% relative to the originally suggested variant.

NOMENCLATURE

DIS – de-icing system;

FB – filter box;

GPU- gas pumping unit;

GTU – gas turbine unit;

IFU – inlet filter unit;

PJSC - Public Joint Stock Company;

SAC – Standard Atmospheric Conditions;

p – pressure drop, Pa;

T – temperature, K.

1 INTRODUCTION

The Russian Public Joint Stock Company Gazprom

today is one of the largest energy companies in the

world. Its main activities are geological exploration,

production, transportation, storage and sale of gas,

gas condensate and oil, as well as the production and

sale of electricity and heat. PJSC Gazprom provides

https://orcid.org/0000-0003-2806-3073

https://orcid.org/0000-0003-4491-1845

https://orcid.org/0000-0003-0737-3048

a continuous cycle of gas supply from the field to the

consumer (Gazprom, 2019; Zabelin et al., 2013).

The peculiarity of the gas industry of the Russian

Federation is that its main gas fields are located to the

east of the Urals, and gas consumers are in Europe

and China. Such a geographical location determines

the presence of a large and extensive gas transmission

system, through which natural gas is supplied from

the field to the consumer. For example, the length of

the Druzhba gas pipeline (across the territory of the

former USSR) is approximately 3,900 km, and of the

Soyuz gas pipeline is 2,750 km (Druzhba pipeline,

2019) (Figure 1). In total, PJSC Gazprom owns 161.7

thousand km of trunk pipelines included in the unified

gas supply system of Russia (Zabelin et al., 2013).

An important element of any gas transportation

system is gas pumping stations, which increase the

pressure of natural gas in the gas pipeline and give an

impulse necessary for its movement. Compressor

stations are located evenly along the entire pipeline

every 100...150 km. In 2013, PJSC Gazprom had 215

linear compressor stations, 6 gas processing

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

Reducing the Hydraulic Resistance in the Inlet Device of the Gas Turbine Unit.

DOI: 10.5220/0007836300880099

In Proceedings of the 9th International Conference on Simulation and Modeling Methodologies, Technologies and Applications (SIMULTECH 2019), pages 88-99

ISBN: 978-989-758-381-0

complexes and 25 underground gas storage facilities,

also using compressor stations. 87.2% of compressor

stations have a gas turbine drive (about 3400 gas

turbine engines (GTE) in total) (Zabelin et al., 2013).

Among the GTUs owned by PJSC Gazprom, over

18% are NK-16ST engines (Figure 2) with a capacity

of 16 MW, developed at PJSC Kuznetsov (Zabelin et

al., 2013; JSC "Kuznetsov", 2019) in the late 1970s

based on the NK-8 aviation engine and operating as

part of GPA-Ts-16 (Figure 3).

Figure 1: Gas transportation system of Russian

(INNOVAES, 2019).

1 - Inlet guide vane; 2 - Low pressure compressor; 3 - Middle

annular frame; 4 - High pressure compressor; 5 - Combustion

chamber; 6 - High pressure turbine; 7 - Low pressure turbine;

8 - Free turbine; 9 - Free turbine bearing; 10 - Clutch.

Figure 2: Gas turbine drive of gas compressor unit NK-

16ST (RTEH GTD NK-16ST Vse o transporte gaza, 2019).

GPA-Ts-16 showed good performance and high

reliability. For this reason, a significant part of GPUs

that have been overaged is not replaced with modern

ones, but is being repaired, combining the

replacement of outdated units with modernization

aimed at reducing costs (increasing efficiency) and

eliminating deficiencies identified in operation.

During modernization, baseline unit has low level of

filtration and quickly clogs up in snowy weather.

1 - turning part of the inlet shaft; 2 - inlet sound absorber;

3 - air cleaning device; 4 - cyclic air heating system;

5 - heat recovery; 6 - exhaust sound absorbers;

7 - diffuser; 8 - exhaust support; 9 - engine compartment,

10 - oil system units

Figure 3: Gas pumping unit GPA-Ts-16 (RTEH GTD NK-

16ST Vse o transporte gaza, 2019).

2 RESEARCH OBJECT

The inlet filter unit of an industrial gas turbine is

intended for cleaning cyclic air coming from the

atmosphere to the engine inlet from dust and other

mechanical inclusions at all possible modes of

operation. IFU, despite its apparent simplicity

(compared to the engine), is an important element of

the GPU, in which complex processes take place. IFU

must reliably clean the air entering the gas turbine

unit from impurities to avoid critical damage to the

elements of the flowing part by foreign objects. In this

case, the pressure losses must be minimal to improve

the efficiency of the engine and prevent compressor

surge. The inlet unit must exclude clogging of filter

elements with snow and ice at all operating

conditions. Errors in the creation of IFU can

significantly reduce the operational properties of the

entire gas turbine station, however, this component is

usually referred to the secondary elements of the

engine and less attention is paid to its perfection than,

for example, to the compressor.

One of the companies engaged in the supply of

equipment for the GPA-Ts-16 under modernization is

LLC Volga-Energogaz, which office is in Samara

(Russia) (VolgaEnergoGaz, 2019). The engineers of

this company offered several variants of the IFU of

GPA-Ts-16 and appealed to the Department of

Aircraft Engine Theory (Department of Aircraft

Engine Theory, 2019) of the Samara National

Research University (Samara University, 2019) with a

request to evaluate the level of hydraulic losses in IFU

variants, and to help in choosing the final variant.

Reducing the Hydraulic Resistance in the Inlet Device of the Gas Turbine Unit

The IFU design proposed by LLC Volga-

Energogaz consists of a weather shield with an inlet

from the bottom, a filter box with two-stage Folter

cassette filters with filtration class G4/F8 (Air inlet

Filtration for GAS turbines, 2019), a diffuser, a vertical

shaft with sound absorbing panels, a turning channel,

an engine inlet duct with lemniscate and de-icing

system pipelines (Figures 4 and 9). The proposed IFU

variants differed in the shape of the weather shield

(with straight walls and roundings) and the

presence/absence of baffles in the inlet section below it

and a rounded rack in the turning channel.

The vertical shaft, the turning channel and the

engine inlet duct with lemniscate are adopted from the

existing GPA-Ts-16 mine. The remaining elements are

being developed and will be manufactured again.

The main idea of the modernization is the

replacement of the inertial filters which proved to be

unsatisfactory with the two-stage Folter cassette filter

with the filtration class G4/F8 of the square shape with

dimensions of 592x592mm. A clean filter of this type

has a hydraulic resistance of 75 Pa. The maximum

pressure drop on the contaminated filter reaches,

according to the manufacturer, 450 Pa (Air inlet

Filtration for GAS turbines, 2019). Unfortunately,

more detailed filter permeability characteristics are not

known.

Figure 4: Proposed IFU design.

The required number of filters (124 pieces) is selected

based on the nominal air flow rate through the engine

102 kg/s, considering 15% of the reserve (i.e. based

on the mass flow rate of 117 kg/s). In the baseline

variant, it is assumed that there are 4 filters along the

height of the filter box. This required to increase the

lateral area of the filter box for their placement by the

increasing complexity of its shape (Figures 4 and 5),

considering the required number of filters. In this

case, the designers seek to keep the minimum size of

the filter box to reduce the load on the shaft.

Figure 5: The adopted shape of the filter box (in plane) with

their accepted numbering.

3 THE INFLUENCE OF

HYDRAULIC LOSSES IN THE

INLET UNIT ON THE

EFFICIENCY OF THE ENGINE

GPU

At the first stage, using the thermodynamic model of

the NK-16ST engine, created and verified according

to the data provided by PJSC "Kuznetsov" (JSC

"Kuznetsov", 2019), a study of the influence of the

hydraulic perfection of the IFU of this engine on the

fuel consumption in the case of maintaining a

constant power of 16 MW (Figure 6) is conducted. A

study on the effect of hydraulic perfection of IFU on

engine power at a constant flow rate of fuel gas is also

carried out (Figure 7).

The results show that an increase in hydraulic

losses in IFU leads to a linear increase in fuel gas

consumption and a decrease in power. Every 100 Pa

of losses in the inlet unit increases the fuel gas

consumption by 2.5 kg/h or reduces the engine power

by 10.5 kW.

Figure 6: The increase in fuel gas consumption of the NK-

16ST depending on the hydraulic losses in the inlet unit

under the constant power.

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

Figure 7: Decrease in the power of the NK-16ST depending

on the hydraulic losses in the inlet unit at a constant

consumption of fuel gas.

4 DESCRIPTION OF

COMPUTATIONAL MODEL

The design model for determining the hydraulic

losses in IFU of GPA-Ts-16 is created in the NX and

Ansys CFX software in accordance with the drawings

provided by LLC Volga-Energogaz.

The computational model includes a part of the

atmosphere around the inlet shaft and a flowing part

of the IFU with an engine block simulator (Figure 8).

The modeling of the surrounding atmosphere is

necessary to correctly simulate the fields of

parameters at the inlet to the shaft, considering

changes in the direction and velocity (force) of the

wind. The engine block simulator is used to consider

the effect of gas pumping unit blocks in modeling

winds blowing from the engine side. The model

considers the full geometry of the channels, the

presence of filters, sound absorbing panels, racks in

the transition duct (if available), pressure drop in the

filter elements, ring collector of DIS, etc. The

presence of the operating floor around the perimeter

of the filters is not considered. Filters are modeled

simplified (see below).

In the numerical simulation of the working

process in the inlet shaft, the following boundary

conditions are applied:

 The inlet condition is applied at the boundary of

the considered part of the atmosphere. The air

temperature (in SAC) T

= 288 K, the velocity

(in m/s) and the direction of the wind (as

direction cosines) are set for it. In all cases, it is

assumed that the wind blows parallel to the

surface of the earth.

 The outlet boundary of the model is located at

the engine inlet. The mass flow rate through the

engine is set for it.

 The surface of the earth is an impenetrable wall.

For the calculation, it is assumed that during the

"numerical experiment" the engine operational mode

(mass flow through it) and atmospheric pressure do

not change. It is also assumed that the mass flow rate

of air through the engine is identical for all considered

working conditions. In other words, the response of

the GTE control system to weather conditions is not

taken into account.

Figure 8: Computational model of the inlet shaft of the

GPA-Ts-16.

Because the exact geometry and permeability

characteristics of the filter materials are not known,

they are modeled as follows. The filter geometry is

simplified, and its shape is parallelepiped. A

condition (interface) is imposed on it according to

which the flow falling on the inlet surface 1 is

transferred to the outlet surface 2 in such a way that

the pressure on the surface 2 is less than 1 by the

amount of pressure drop in the filter Δp. Thus, in the

created computational model, total pressure losses on

the filters are considered the same in all filters (the

filters are uniformly polluted), regardless of external

conditions (wind speed and direction).

An ideal gas with air properties and variable heat

capacity and viscosity is used as a working fluid in

the simulation. When calculating, the turbulence

model k-ε with a scalable wall function is used in

calculation.

The calculation model is meshed by a finite

volume grid (Figure 9). It is based on tetrahedral

elements in combination with a prismatic layer on the

surface of streamlined walls. The total number of

final volumes is 11 million. The number of finite

volumes is selected based on the study of mesh

convergence, the results of which are shown in

Table 1. In the course of the study, 4 models were

created based on one computational model

corresponding to the initial geometry (Variant 1,

Table 2), differing in the density of the finite volume

mesh.

Reducing the Hydraulic Resistance in the Inlet Device of the Gas Turbine Unit

Figure 9: The grid of finite volumes of the created

computational model.

Table 1: Results of the study of the mesh convergence of

the computational model.

The number of

model elements,

mln

The drop in total

pressure in the

shaft, Pa

703

728

750

746

It is obvious from Table 1 that the global

numerical values obtained by models with a mesh of

11 and 39 million elements are close, but the

computation time in the latter case is significantly

higher. Therefore, in the future, a mesh with an

approximate number of elements of 10...12 million is

adopted for all studies (the number of elements

changed due to changes in the geometry of the shaft).

Due to the large dimension of the computational

grid, gas-dynamic modeling is carried out on the

“Sergey Korolev” supercomputer (Supercomputer

Center - Samara University, 2019). The calculations

involved 128 cores and 128 GB of RAM.

During the study, 4 variants of the inlet shaft

design (Table 2) are considered:

 Variant 1 - In accordance with the drawing

provided by the Customer (Figure 11) (with

roundings and baffles at the inlet and rack in the

turning channel);

 Variant 2 - Option without rounding and baffles,

but with a rack;

 Variant 3 - Identical to Variant 1, but without a

rack (Figure 4);

 Variant 4 - Identical to Variant 1, but without

baffles.

Table 2: Scheme of differences of the considered variants.

Variant

Rounding at the inlet of the

weather shield

Baffles at the inlet

Rack in the turning channel

All variants are considered in the absence of wind

and in the presence of winds at velocity of 10, 20 and

35 m/s from five main directions (0, 45, 90, 135,

180) (Figure 10).

To simplify the analysis of the obtained results,

the averaged values of the flow parameters and

contours of the parameters (velocity and pressure) in

the flowing part are considered in 7 control sections

along the path (Figure 11):

 0 - the lower edge of the weather shield (is

absent in variant 2);

 1 - lower edge of weather shield after baffles;

 2 - at the inlet to the filters (box-shaped);

 3 - between filter box and diffuser;

 4 - in front of sound absorbing panels;

 5 - after sound absorbing panels;

 6 - at the GTE inlet.

Figure 10: Main wind direction.

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

Figure 11: Considered IFU control sections.

5 STUDY OF THE IFU

WORKFLOW IN CALM

WEATHER

Summarized data on the expected values of total

pressure losses in the considered IFU variants without

the wind effect (wind velocity is 0 m/s) is shown in

Figure 12. Losses are calculated as an algebraic

difference between the area-averaged values of the

total pressure in the control sections. The distribution

of total pressure losses between the section of the IFU

flowing path for the considered variants without wind

is shown in Figure 13.

It is clear from the figures that all the considered

variants show close hydraulic characteristics (total

pressure losses) in calm weather. The difference

between the four variants is no more than 30 Pa. In

calm weather, the total pressure losses vary from 380

(with clean filters) to 750 Pa (with dirty filters).

Figure 12: Total pressure losses in IFU for various

modifications of the baseline design in calm weather.

Figure 13: Total pressure losses in IFU elements for various

modifications of the baseline design in calm weather.

The total hydraulic losses in the flowing part of

IFU in calm weather for all the considered variants do

not differ and are approximately 310 Pa (excluding

losses in the filters).

As can be seen from the obtained results, the

physical picture that occurs with “clean” (Δp = 75 Pa)

and “dirty” (Δp = 450 Pa) filters is identical under the

assumptions made (uniform contamination of filters).

The difference is only a quantitative estimation of the

hydraulic losses (the computational models for

estimating losses with clean and dirty filters are

identical, except for the pressure drop at the interfaces

simulating the operation of the filter (75 Pa for clean

filter and 450 Pa for dirty filter).

For this reason, in the future, all computational

models are calculated only with “dirty” filters (Δp =

450 Pa), since this option is the ultimate in terms of

losses.

The main losses of total pressure in the shaft occur

when the working fluid passes through the filters

(section 2-3) and when the flow turns in the transition

duct and moves in the channel before entering the

engine (section 5-6). In the first two section (0-2) with

zero wind, the loss of total pressure is negligible.

Losses in the diffuser (section 3-4) and in sound

absorbing panels (section 4-5) are relatively small (do

not exceed 40 Pa) and do not depend on the shaft

design.

It should be noted that the losses in the filter box

are largely determined by the losses on the filter

element. Additional losses in this component do not

exceed 70 Pa.

The reason for the increased losses when turning

the flow in the turning channel and in the inlet channel

of the GTE is in the fact that during the flow of the

working fluid, vortex zones appear on the right and

on the left of the engine, as well as a high level of

velocity (up to 70 m/s) in the engine inlet channel.

Reducing the Hydraulic Resistance in the Inlet Device of the Gas Turbine Unit

The vortex in the turning channel is generated as

follows. The internal part of the filter box is divided

by baffles into three parts (two external and one

internal) (Figure 5). The air passing through the filters

of the inner part is directed vertically downward

directly to the engine inlet and is immediately sucked

into it (Figure 14). Only a small part passes by and

interacts with side vortices or falls into the space

between the lemniscate and the front wall of the

engine block.

At the same time, the air trapped in the outer parts,

passes a more complex trajectory. For example, air

passed through filter No. 6 (Figure 5) goes through all

the space of the filter box almost to filter No. 3

(Figure 14) and turns down there. The air that has

passed through the side filters No. 4, 5, 7 (Figure 5)

interacts with the flow through filter No. 6 and almost

immediately turns down. Thus, a substantial mass of

the working fluid from filters No. 4-7 moves in the

direction of the filter No. 3 (Figure 5) inside the box.

As a result, the working fluid passes through the filter

No. 3 only in the lower sections outside. From the

upper sections, the working fluid is pulled out in the

opposite direction (Figure 15). Thus, with the adopted

configuration of the filter box, not all filters work in

the same way. Filters facing the atmosphere pass the

working fluid through only in the “inward” direction.

Part of the filters located in the recesses ("pockets"),

pass the working fluid in smaller quantities, or vice

versa release it back.

In Figure 15 and further on similar graphs (Fig. 21

and 24), a circular diagram of the distribution of air

flow between the vertical rows of filters is shown.

This graph is plotted in the polar coordinate system.

In it, along the radius, the total value of the mass flow

rate through the vertical row of filters in kg/s is

deposited. A positive sign indicates that the working

fluid flows into the filter unit, a negative one indicates

that it flows out. The angular position corresponds to

the number of the filter row according to Figure 5

(indicated on the periphery of the circle).

Figure 15 illustrates this fact, showing the

distribution of working fluid mass flow rates between

the vertical rows of filters for all 4 considered IFU

variants (excluding wind) with maximum and

minimum pressure losses on the filters. Thus, in calm

weather, 4 vertical rows of filters do not work (the

working fluid flows out of them) and another 4 rows

work with reduced flow rates.

Analysis of the flow structure (Figure 16) of the

considered IFU variants shows that in calm weather,

air enters the shaft from all directions, uniformly

filling the entire inlet section.

Central part

Side part

Figure 14: An example of the flow of air through the various

sections of the filter box.

Figure 15: The distribution of the flow rates of the working

fluid between the filters (the numbers correspond to

Figure 5) (the shaded areas correspond to the filters in the

"pockets").

If we compare the flow structure in the variants

without rounding of the weather shield walls (Variant

2) and with it (Variant4), it can be concluded that a

large flow separation is formed on the inlet edge of

the weather shield for the variant without rounding,

which reduces the effective flow area, and increases

the flow rate at the inlet to the filters. Thus, the

rounding at the inlet to the weather shield forms a

more favorable flow structure at the inlet, reducing

the velocity at the inlet and its unevenness. In general,

the rounding of the weather shield in calm weather

reduces hydraulic losses by a relatively small value.

However, in the presence of wind, the gain

significantly increases.

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

Velocity vectors

Variant 1

Variant 2

Figure 16: The flow structure in section 0 (inlet to the

weather shield) for variant 1 and 2.

6 STUDY OF THE IFU

WORKFLOW IN THE

PRESENCE OF WIND

To assess the influence of wind force and direction on

the working process and hydraulic losses of the inlet

shaft, a study is conducted in which the wind velocity

varies from 0 to 35 m/s and the direction from 0° to

180° (Figure 10).

Figure 17 shows how the values of the total

pressure losses change in all the considered variants of

the inlet shaft when the force and direction of the wind

change. As can be seen, with increasing wind velocity,

the hydraulic resistance of the shaft is constantly

increasing. The increase in losses is not linear. With the

wind velocity of up to 10...15 m/s, the losses of total

pressure in the inlet shaft differ little from non-wind

conditions. A further increase in wind velocity leads to

an avalanche-like increase in losses.

In order to understand the causes of losses, the

averaged (among all the calculated variants of the

design and wind direction) total pressure loss values

in characteristic sections of the inlet shaft at various

wind velocities (Figure 18) are calculated.

The most intensive growth of losses is observed at

the inlet to the shaft (section 0-2, more than 500 times

with a change in wind speed from 0 to 35 m/s) and in

filter box (by 70%). In the turning channel (sections

5-6), the growth of losses is moderate (30%). The

greatest losses occur with side winds (45...135°). It is

curious to note that the change in wind velocity and

its force has little effect on the losses of total pressure

in the diffuser and sound absorbing panels, which

confirms the conclusion made earlier that minor

Figure 17: The dependence of the total losses of total

pressure in the inlet shaft with different force and direction

of the wind.

Figure 18: The change in the averaged values of the total

pressure losses in the characteristic sections of the inlet

shaft with increasing wind velocity.

design changes in these parts are not able to

significantly change the shaft hydraulics.

Some changes in the total pressure losses in the

turning channel (section 5-6) is associated with a

change in the flow structure there. In the presence of

wind, there is a redistribution of air flow between the

outer and inner parts of the filter box. The flow rate

of the working fluid through the windward outer part

increases, through the inner and leeward outer part -

decreases. As a result, the intensity of the vortices and

the average level of the flow velocity located on the

right and left of the engine inlet change, the flow loses

a symmetrical character, local flow accelerations

appear (including in the filter box and sound

absorber), the interaction of the vortices changes,

which increases losses.

Losses in the section before the filters are

associated with the formation of vorticity during

flowing past the weather shield inside and outside.

Figure 19 shows the trajectories of air particles before

they fall under the weather shield when the wind

direction is 45°. With other wind directions, the flow

Reducing the Hydraulic Resistance in the Inlet Device of the Gas Turbine Unit

structure does not change in principle. It can be seen

that the working fluid is sucked into the shaft evenly

from all directions in windless weather, but in the

presence of wind the main part of the air enters from

the narrow sector on the windward side. At the same

time, when flowing the weather shield and shaft, a

vortex structure is formed on the leeward side (Figure

19, 20).

In calm weather, all the working fluid goes into

the filters after passing section 0 along the shortest

path. Overflow under the shield is minimal. In the

presence of wind, the working fluid immediately

enters the filters only on the windward side. At the

same time, the air flow velocity increases

substantially compared to calm conditions. The

amount of working fluid that passes under the shield

from the windward side is greater than the filters can

pass through on this side, and the working fluid

begins to flow to the leeward side. This overflow,

coupled with areas of reduced pressure on the leeward

side, caused by vortexes formation in the external

flowing past the shaft, leads to a complex unsteady

vortex structure between the inlet section and filter

box inlet on the leeward side. These circumstances

lead to the fact that the working fluid through the inlet

section under the weather shield (section 0) enters

unevenly, and there is a large number of backflows

(Figure 20), which increases the values of local

velocities there.

Another negative effect associated with such a

flow pattern under the shield is that the DIS tubes

installed in front of the filters is flowed not by air,

which goes in the direction of the filters, but by a

multidirectional and non-stationary flow. This will

blur the field of increased temperature created by

them and reduce the efficiency of the system.

Wind velocity - 20 m/s at an angle of 45



Figure 19: The trajectory of the particles of the working

fluid, before getting under the weather shield.

Wind velocity - 20 m/s at an angle of 135



Figure 20: Flow velocity vectors in the inlet section of the

shaft.

The described structure of the flow under the

shield in the presence of wind leads to the fact that the

distribution of flow rates between filter elements

changes. To confirm this, Figure 21 shows how the

mass flow rates of the working fluid change through

filter groups at a wind direction of 0° at a wind

velocity of 20 m/s. Distribution of mass flow rates for

other wind directions is not fundamentally different.

As can be seen, the presence of wind has a

significant impact on the distribution of mass flow

rates between the filter sections. As a rule, on the

windward side, the flow rate of working fluid through

the filters increases. There is an increase of flow rate

(at 20 m/s up to 3 times relative to windless weather)

at the angles (90...135° and 220...270°) relative to the

wind direction. In the presence of wind, the number

of filter elements (by 4 pcs.) from which the working

fluid is thrown back or the flow rate is significantly

less than the average value (they are mainly

concentrated in the “pockets” and on the leeward

side) on average is 12, which is 1.5 times greater than

in calm weather. The greatest number of filters that

do not work properly takes place with side winds

(45...135°).

The greatest pressure losses during the flowing

past the weather shield occur when the wind blows

into the corner (45° and 135°). This is because an

additional vortex is generated under the shield from

the inner side of the corner (Figure 22).

The presence of baffles streamlines the flow of the

working fluid under the shield. Without baffles, the

working fluid “hits” the lower row of filters, making

the mass flow rate distribution between the filters of

one vertical more uneven. On the windward sides, the

working fluid also moves with a significant horizontal

component, which adversely affects the operation of

the DIS. The variant with rounding the lower part of

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

Figure 21: Distribution of mass flow rates through filters

with a wind of 20 m/s, flowing at an angle of 0° (filter

numbers correspond to Figure 13).

Figure 22: An additional vortex under the shield when the

wind hits "into the corner" (the angle of 135).

the weather shield, as in the case of windless weather,

provides the best inlet under the weather shield,

without additional vortex formation.

7 RECOMMENDATIONS FOR

THE DESIGN CHANGE

As a result of the study, recommendations are made

to change the baseline design of the IFU, which will

reduce hydraulic losses:

 At the inlet to the space under the weather shield,

it is necessary to install baffles, which will

improve the performance of filters and PIC

(especially if there is wind).

 For a more uniform entry of the working fluid and

reduction of inlet losses, the weather shield must

have an extension at the inlet, but its design can

be simplified.

 To reduce clutter on the inlet section, the DIS

collector must be removed from section 0.

 Changes in the design of the diffuser and sound

absorbing unit do not have a noticeable effect on

the hydraulic perfection of the shaft.

 The presence of the rack in the turning channel

does not affect the level of losses, and it can be

removed from the design.

 The layout of the filters must be revised to reduce

the number of abnormally operating filters. To

reduce losses at side wind, it is necessary to

increase the lateral projection of the filter box. In

the ideal case, the filter box must have a square

shape.

 To reduce the backflow from the filter unit inside

it, it is necessary to install impermeable baffles.

8 INVESTIGATION OF LOSSES

IN MODERNIZED IFU

VARIANTS

Based on the recommendations given in the previous

section, 2 variants of IFU are developed.

Variant 5 - In general, it is identical to Variant 3,

but differs in the modified dimensions of the diffuser,

filter box and weather shield, as well as in

simplification of their shape (Figure 23).

Variant 6 - A variant which is characterized by a

rectangular-shaped filter box (Figure 24, in which (8

(in width)*6 (in length)*5 (in height) filters are

located), by the absence of sound absorbing panels and

by a weather shield with additional windows in the side

walls.

For both modernized variants, computational

models are created, similar to those described above.

The calculations show that Variant 5 has the hydraulic

characteristics close to the Variant 3. The magnitude of

the pressure loss increased by 40 Pa (5%).

Variant 5

Variant 6

Figure 23: Modernized IFU variants.

Reducing the Hydraulic Resistance in the Inlet Device of the Gas Turbine Unit

Figure 24: Adopted filter numbering in Variant 6 of IFU.

Variant 6 ensures the achievement of minimal

losses among all the considered variants of IFU. In

the absence of wind, the expected total hydraulic

resistance in the upgraded IFU is: initial - 260 Pa

(taking into account the resistance of clean filters of

75 Pa), final - 635 Pa (taking into account the

resistance of dirty filters of 450 Pa). This is 107 Pa

(15%) less than the baseline variant of the design (No.

1). The reason for the improvement is the absence of

sound absorbing panels, a lower level of losses at the

inlet and a greater uniformity of flow at the inlet to

the filter elements.

The average velocity at the inlet to the filter box

of the Variant 6 in calm weather is 1.78 m/s (for

comparison, in the baseline design (No. 1) is 3.01

m/s), at the inlet to the weather shield is 1.52 m/s (for

Variant 1 is 2.09 m/s). Note. The velocity of flowing

on the filter element recommended by the filter

manufacturer is less than 2.5 m/s.

In calm weather, the distribution of the flow rates

of the working fluid through the vertical rows of

filters of IFU No. 6 is close to uniform

(4.250.25 kg/s). In contrast to the previously

considered variants, in IFU No. 6 under no-wind

conditions there are no sections with a significantly

different level of mass flow rates of the working fluid,

or through which the working fluid flows out (Figure

25).

Variant 6 in windless weather provides a

significantly more uniform "loading" of filters, which

reduces the velocity level of air flowing on them. It

should also be noted a smaller difference in the

distribution of velocities (Figure 26) along the height

of the filter box which suggests that Variant 6

provides large flow rates through the upper filters

and, consequently, reducing flow rates through the

lower ones. These facts indicate more favourable

conditions for the operation of filter elements in

Variant 6.

Figure 25: The distribution of the flow rates of the working

fluid between the filters (numbers for variants1-5 - Figure

5, variant 6 - Figure 23 (pink areas correspond to the filters

in the "pockets" for variants 1-5).

Figure 26: The flow structure through the entrance windows

of the weather shield of the IFU No. 6 in calm weather.

In the presence of wind at 20 m/s, blowing strictly

into the face of the shaft, Variant 6 exceeds the

baseline variant (number 1) in hydraulic losses by

150 Pa. With the direction of the wind "from an

angle" (45° and 135°), the modernized variant shows

a higher level of losses by 150 Pa. The reason for the

increased losses in the characteristics of the air flow

in the space between the weather shield and the filter

box (Figure 27).

To reduce the intensity of vorticity and streamline

the flow structure under the weather shield, its shape

must be optimized, and internal baffles must be

installed similar to the baffles near the lower edge of

the weather shield. To select the shape and position

of these baffles, considering the effect on the

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

complexity of the maintenance of IFU, additional

research is necessary.

Figure 27: Dependence of the total loss of total pressure in

IFU of the main variants of the inlet shaft at a wind of

20 m/s with its different direction.

9 CONCLUSIONS

In the current research, a numerical simulation is

carried out of the air flowing in various variants of the

modernized inlet unit of the GPA-Ts-16 gas-pumping

unit with a two-stage cassette filter.

The values of total pressure losses in all elements

of IFU are calculated and the structure of the flow

inside is studied in detail for the absence or presence

of wind at a velocity from 0 to 35 m/s blowing from

5 main directions (0, 45, 90, 135, 180). Studies

are also carried out on the effect of the weather shield

shape, the presence of baffles under it, and the rack in

the shaft on the hydraulic losses.

As a result, the main sources of energy losses are

identified and recommendations for the design of

inlet shafts of this type are provided.

Based on the experience gained, an IFU

configuration is found that provides a reduction in

hydraulic losses relative to the baseline variant by

15%. Changing the IFU design will save 2.67 kg or

3.42 m3 per hour for each engine.

At the same time, the obtained design has reserves

for improvement due to streamlining the flow under

the weather shield using baffles.

The work carried out should be continued in the

direction of analyzing the work of the UFI de-icing

system and ensuring its effective operation in the

whole range of operating parameters.

Moreover, it is planned to validate simulation

results by experimental data after the inlet device will

be manufactured and tested.

ACKNOWLEDGEMENTS

This work was supported by the Russian Federation

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

REFERENCES

Air inlet Filtration for GAS turbines, 2019. Access mode:

http://www.folter.ru/en/catalog/vozdushnye-filtry-

folter-dlya-gazovyh-turbin/.

Departmenr of Aircraft Engine Theory, 2019. Access mode:

http://tdla.ssau.ru.

Druzhba pipeline – Wikipedia, 2019. Access mode:

https://en.wikipedia.org/wiki/Druzhba_pipeline.

Gazprom, 2019. Access mode: http://www.gazprom.ru/.

INNOVAES | Innovación en Asuntos Estratégicos, 2019.

Access mode: innovaes.com.

JSC "Kuznetsov", 2019. Access mode:

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

RTEH GTD NK-16ST Vse o transporte gaza, 2019. Access

mode: http://www.turbinist.ru/6830-rte-gtd-nk-

16st.html.

Samara University, 2019. Access mode:

https://ssau.ru/english.

Supercomputer Center - Samara University, 2019. Access

mode: https://ssau.ru/it/supercomputer.

VolgaEnergoGaz, 2019. Access mode: http://www.v-eg.ru/.

Zabelin, N.A., Lykov, A.V., Rassokhin, V.A., 2013.

Calculation of available heat power of smoke fumes of

gasocompressor units of russia’s united gas

transmission system. Nauchno-tekhnicheskie

vedomosti Cankt-Peterburgskogo gosudarstvennogo

politekhnicheskogo universiteta. 4-1(183): 136-144.

Reducing the Hydraulic Resistance in the Inlet Device of the Gas Turbine Unit