TECHNICAL REVIEW

No.17 / March 2022

echnical eview
MAPNA TURBINE ENGINEERING & MANUFACTURING CO. (TUGA)

Willpower to Empower Generations

TECHNICAL REVIEW

Editorial & Production

Editorial Director:
Owliya, Mohammad

Associate Editors:
Rashidi, Saeid
Sarmast, Mohammad
Hajizadeh, Hamed

Production Director:
Jabery, Roohollah

Coordinator:
Azizi, Fakhrodin

Graphic Designer:
Radfar, Kianoosh

Cover Page:
A MAPNA Turbine technician standing in
front of an MGT-70 assembly platform
Editorial
Dear Colleagues, Partners and Professionals,
At MAPNA Turbine we are constantly learning, adapting and evolving to help our clients by
delivering world-class solutions with beneficial impacts on their bottom lines. A brief account
of a few recent technological achievements is presented to you, our valued readers, in this
edition of MAPNA Turbine Technical Review.

The first article takes an in-depth look into the design of MST-40C as the first three-level-pressure
steam turbine designed by MAPNA Turbine engineers to be mated with our new MGT-75 gas
turbine product in the near future to deliver 59.5% efficient combined cycle. It spells out in
detail all meticulous steps taken to design different parts and components of the machine
including but not limited to the last stage blades, the steam path, the rotor and casings as well
as all the related auxiliary systems.

The second article takes a sneak peek into the latest upgrade scheme already laid out and
pursued at MAPNA Turbine for our flagship MGT-70 gas turbine. One of the key attributes of
the new upgrade scheme is, among others, the extension of the current periodic inspection
intervals of 33,000 EOH to an incredibly high 50,000 EOH, equivalent to almost 6 years of normal
operation. Therefore, significant operating and maintenance cost reductions and fewer power
plant outages would be expected, to the full benefit of our clients.

Steps taken to improve MGT-40 gas turbine output at some ambient conditions via compressor
blades’ stagger angles are elaborated on in the third article. This would be beneficial for some
customers of the machine. A number of analysis and optimization tools, including an internally
developed cycle analysis code, have been taken advantage of in the new design to reduce
the associated computational time and costs to a significant extent.

The fourth article features a thorough investigation into the effects of different scanning
strategies and support geometries and dispositions on subsequent residual stress and distortion
of Inconel 625 SLM additively manufactured parts. The results of the combined experimental–
computational analyses shed more light on the complex nature of the process and allow for
more precise and reliable outcomes.

The last article reflects on two of the latest improvements made to make final assembly process
of MGT-70 gas turbines more efficient and less time consuming. This was achieved by ruling out
and optimization of the prolonged processes of using dummy shafts for the casing/bearing
alignments and the two-step rotor blades grinding.

Please join us in relishing a detailed account of these subjects in this issue of the Technical
Review.

Respectfully,

Mohammad Owliya, PhD

Managing Director

MAPNA Turbine Company (TUGA)

March 2022

MAPNA Turbine • March 2022 1

 Contents

3-11 22-31

1
An In-depth Look at MST-40C Steam Turbine
Design

4
12-15 Evaluation of Residual Distortions through
Simulation of Scan Patterns & Supports for
Inconel 625 Additive Manufacturing Process

2
MGT-70(4); A Sneak Peek into the Latest
32-40
MGT-70 Gas Turbine Upgrade Scheme

16-21

3 5
MGT-40 Gas Turbine Compressor Modification More Streamlined & Efficient MGT-70 Gas
Turbine Manufacturing Process

2 echnical eview No.17

 1
Introduction

M APNA Turbine has started manufacturing

a brand new gas turbine branded as
MGT-75 with nominal power output of 220 MW
and 217 MW in simple and combined cycle
applications respectively. This machine falls
into F-Class category with TIT (Turbine Inlet
Temperature) of 1230 °C and TET (Turbine
Exit Temperature) of 578 °C. MGT-75 was
domestically designed by MAPNA Turbine
engineers, ideally suiting relatively low budget
projects, taking into account its relatively lower
power in comparison with other major OEMs’ A n I n -d ep t h
Look at
products of the same class, while maintaining
the efficiency at a world-class level.
In order to better exploit the heat of the
exhaust gas of the newly designed gas turbine, M S T-4 0 C
it was decided to design a 3-level-pressure
steam turbine designated as MST-40C. The S t e a m Tu r b i n e
Des i g n
condensing steam turbine produces 102 MW
with reheat steam flow, meant to increase
the combined cycle efficiency to 59.5% at
standard ISO conditions.

Modarres, Ali

MAPNA Turbine Engineering & Manufacturing Co.

(TUGA)

MAPNA Turbine • March 2022 3

 Fig. 1 – MST-40C steam turbine 3-D model

The last stage blades of this machine have been designed and sized to have the best possible
performance in typical domestic sites with relatively high condenser pressures. Moreover,
MAPNA Turbine’s in-house technology has been utilized for designing upstream stages, as well
as rotor features and casings.

Design Features
■■ Steam Stream
Steam first enters the HP section at a temperature of 565 °C and exits the section as CRH (Cold
Re-Heat) stream, flowing back to the HRSG to become heated up again to the main steam
temperature. The HRSG’s second pressure steam (IP) mixes with the CRH steam right before
reheating. The HRH (Hot Re-Heat) steam enters the IP section at the middle of the IP outer
casing and at the same temperature of 565 °C where it passes through the section blading. The
HRSG’s third pressure steam (LP) then blends with the IP outlet steam within the IP outer casing
area, flowing to the LP section. The exhaust steam eventually flows on to the condenser. There
are 31, 16 and 6 stages in the HP, IP, and LP sections respectively.
To find a balance between the equipment price and cycle efficiency, it is economical to
use three-level-pressure steam turbines for F-Class gas turbines meaning that the HRSG would
produce steam at three different pressure levels (115 bar, 26 bar and 5 bar for instance). This
ensures that water absorbs as much energy as possible from the gas turbine exhaust, while
keeping the equipment price at a reasonable level. Although having more pressure levels
would slightly increase the overall efficiency, it is not an economically viable option.

■■ Key Points Considered

There are a number of key points that had to be considered while designing the machine.
First of all, the turbine had to be compatible with ACC (Air-Cooled Condenser) system. This
is because Heller type condensers have been crossed off MAPNA Group’s standard cooling
system options and are only used on client’s request and unless the plant is near the sea or a
big river allowing for application of once-through cooling systems, ACC cooling systems are
the only viable option for newly built plants or those that are about to be launched.

4 echnical eview No.17

Secondly, attempt was made to keep the resultant thrust force at the minimum permissible
value. This has been achieved by selecting the steam path configuration, so that the HP and
IP sections face the generator, whereas the LP section is in the reverse direction towards the
condenser. As a result, there is no need for huge thrust bearings to counter large thrust forces.
Thanks to this configuration, impermissible temperature difference between the upper and
lower halves of the IP outer casing is ruled out as well, as it does not let the steam get trapped
between the IP inner and outer casings. When steam does get trapped in there, hotter steam
might be collected within the upper half and less hot steam within the lower half resulted from a
weak ventilation caused by low LP steam inlet flow. In the current configuration, there is always
enough steam flow between the IP inner and outer casings, giving rise to strong ventilation and
blending in that area.
Thirdly, the IP-LP rotor had to comprise three parts welded together. This is mainly because the
material has to be able to withstand high thermal stresses or high centrifugal forces wherever
required. Since the IP steam enters at the middle of the outer casing, the part in the middle
has to be resistant to high steam temperatures. The other two parts however, experience high
centrifugal forces and stresses. Moreover, the welding feature gives the designer more freedom
to choose higher steam temperatures compared to the case when the entire rotor is made of
a single part. Welded rotor also gives the designer another advantage: Manufacturing on a
modular design basis. Unlike gas turbines, steam turbines are normally subject to customization
and change of the LP end, based on site conditions. The more volumetric flow at the exhaust
because of low condenser pressure, the larger the exhaust annulus area. This should be
noted that each last stage blade set has its own setup of tip and hub diameters, axial chord,
clearances, etc., so it is not simply replaceable. Therefore if a larger LP end is to be used, the
rotor will have to be thicker. Thanks to the welding feature of the rotor, the LP end is able to be
selected and customized individually while keeping the upstream stages unchanged.

■■ Steam Parameters
Steam parameters, i.e., pressure, temperature and flow rate were finalized in a mutual
information exchange within MAPNA Group, so that all parties involved in designing the steam
cycle came to a consensus. The rather high condenser pressure was particularly chosen in a
way that the turbine would be compatible with typical Iranian site conditions. As a result, the
last stage blade set was designed. The key issue was calculating the exhaust annulus area so
that the exiting flow velocity would be about 240 m/s which is quite normal for a steam turbine
LP end.
The agreed upon steam parameters are listed in Table 1.
Table 1 – MST-40C steam parameters
Item Parameter Value Unit
1 Main steam flow rate 60.8 kg/s
2 Main steam temperature 565 °C
3 Main steam pressure 112.6 bar
4 Reheat steam flow rate 68.3 kg/s
5 Reheat steam temperature 565 °C
6 Reheat steam pressure 26.6 bar
7 LP steam flow rate 6.7 kg/s
8 LP steam temperature 233 °C
9 LP steam pressure 3.7 bar
10 Condenser pressure 0.082 bar
11 Power output 102 MW

MAPNA Turbine • March 2022 5

 ■■ Last Stage Blades Design
As mentioned earlier, the last stage sizing was carried out so as to have a good performance
at high condenser pressures, keeping an eye on normal operating conditions with low back
pressures. The exhaust velocity of 240 m/s was selected on account of the ACC cooling system
type. Among different cooling systems, ACC has rather large pressure fluctuations, particularly
when the turbine is in the startup phase. Therefore an LP end sized to cope with high back
pressures, and still providing an acceptable performance in low back pressure conditions
would be of great benefit. The hub diameter of the upstream stages depended on the sizing
of the LP end.
The second issue following setting the exhaust annulus area was the layout of the LP end
stages. There are various types of LP last stage rotary blades, e.g., free-standing, snubber type,
integrally shrouded, and combinations of the above. For MST-40C, bearing in mind that the last
blades were the first set of the kind to be designed, free-standing type was selected which is
easier to design and also simpler to assemble/disassemble. However, significant tip leakage
associated with this type of layout, could be a drawback.
The material of the blades was selected on the basis of water droplet erosion index. This index
is calculated taking into account the linear blade velocity, the steam quality at the exhaust,
the turbine speed and other parameters. This very index determines which of the two blade
hardening methods (flame or laser hardening) shall be used. For the designed 3.7 m2 set of LP
end blades (TF-37) particularly, flame hardening method was chosen to be used.

Fig. 2 – TF-37; Last stage blade set designed for MST-40C steam turbine

6 echnical eview No.17

Considering the fact that typical Iranian sites have rather high condenser pressures, it was
deemed necessary to design the LP blades capable of withstanding such operational conditions
that happen mostly in the summertime. The higher the allowable condenser pressure, the better
and more economical the cooling system performance. For this reason, 0.5 bar was targeted
as the relevant trip value. This should be noted that the designed set of LP blades is applicable
to a range of turbine power outputs. Roughly speaking, TF-37 may be used in turbines with
outputs from 70 MW up to 120 MW. Thanks to rotor welding feature, it is possible to manufacture
the IP-LP rotor on a modular design basis. In other words, this rotor is able to be customized for
different condenser pressures.

■■ Steam Path Design

The first step following the design/selection of the LP end, was calculating the number of stages
for each section. The curve of efficiency against velocity ratio, i.e. the ratio of linear blade
velocity over equivalent velocity of isentropic enthalpy drop across the stage, has a dome
shape for each section. The dome indicates that both too low and too high velocity ratios
would result in stage efficiency drop. Having too few stages means high enthalpy drop and
thus low velocity ratio, whereas too many stages would mean the opposite. Therefore the
number of stages shall be meticulously calculated to bring about the best possible efficiency
for each section of blades. As mentioned earlier, 31, 16 and 6 stages were calculated for HP,
IP and LP sections, respectively.
Aside from the above, there are various configurations of the steam path each with their
own pros and cons. For instance, HP and IP sections may be on a single rotor in a separate
cylinder and LP in another one, or HP section in one cylinder and IP and LP sections in another.
In addition, each section can face either the generator or the condenser. For each option,
the direction of the thrust force applied on the blades by the steam flow must be taken into
account. This is because the resultant force on the entire rotor train shall not be more than a
pre-assumed limit regardless of the number of rotors. This becomes important when it comes to
thrust bearing design. If the bearing was to withstand too much of thrust forces, it would need
to be bulky, resulting in huge oil consumption in the lube oil system design. Moreover, the rotor
collar meant to face thrust pads would also need to be too large.
The profile used for drum blades was our in-house technology which not only provided
acceptable efficiency but also was more familiar to our manufacturing department staff
translating into easier production processes and fewer non-conformities involved. This profile
was made use of throughout the turbine steam path, with different stagger angles, heights,
and scales though.
The hub slope angle of steam path is usually a multiplication of 1.5 degrees and may be selected
based on the designer’s preferences. This parameter affects the hub-to-tip ratio of each blade
row. Lower values mean longer blades and thinner rotors and vice versa. As the linear velocities
of the blade hub and tip get close to each other, they do not require twisted airfoil. This is
because the added efficiency of the twisting feature is not worth the manufacturing difficulty
of the airfoil. However, thicker rotor resulted from high hub-to-tip ratio of the blades brings
about heavier components, bigger bearings, more oil consumption and higher price. On the
other hand, lower values of hub-to-tip ratio (say 0.75) lead to longer blades which may need
twisting. In addition to difficult manufacturing, special provisions are needed for roots of the
blades on which higher centrifugal force is applied. A compromise is inevitable considering all
of the above mentioned mechanical, aerodynamic, economical and manufacturing aspects.
Attempt was made to design cylindrical type HP and IP stages, allocate twisted blades to the
LP section only, and keep the rotor diameter in the common practice range for a typical 100
MW turbine.

MAPNA Turbine • March 2022 7

 The main steam enters the HP section through one set of HP stop and control valves. In a similar
way, the reheat steam has one set of IP stop and control valves integrated into a common
casing. The LP steam has its own stop and control butterfly valves. The position of the HP and IP
valves relative to the casing, whether left or right, was meticulously selected in order to avoid
steam turbulence when reaching the leading edges of the first row. Bearing in mind that the
valves have been positioned sideways, the steam flows in a circular motion downstream the
valves. This means that the incidence angle at the first stage may be biased either to positive
or negative relative to 0°. The latter has a worse impact like braking against the blade. So, the
valves’ position has been chosen to avoid such detrimental consequences by establishing
positive incidence angles.
Two types of sealing including labyrinth and tip-to-tip were considered in this machine. The
labyrinth seals have better function and less leakage; however, since this type of sealing has
limited relative axial expansion, they can only be used for portions close to the thrust bearing.
For the regions such as the LP section that are away from the axial fixed points, tip-to-tip
seals were applied. Proper function of the seals in both hot and cold working conditions was
thoroughly investigated.

■■ Rotor Design
Following the steam path design, the rotors were designed so as to allow for higher steam
temperatures at the inlet while having higher strength against centrifugal stresses wherever
required, fewer bearings needed and also offer a modular manufacturing scheme.
It was deemed necessary at the beginning of the project to consider exhausts with different
annulus areas for different site conditions. However the change in the LP end should not give
rise to a new IP-LP rotor design. This can expedite the procurement and manufacturing of the
rotor for mass production. In other words, the IP-LP rotor was designed with three parts welded
together with two welding seams. The middle part and the IP section can be procured on
a “make-to-stock” basis, whereas the LP portion is supplied as per each specific order in a
“make-to-order” fashion.
Fewer bearings is considered a substantial benefit which was taken advantage of in the design
of this machine. The weight of the HP rotor, shown in Fig. 3, is borne by two bearings, a journal
as well as a combined journal-thrust bearing. The IP-LP rotor however has only got one bearing
located at the LP exhaust section. The other end of this rotor is coupled to the HP rotor and
its weight is borne by a common combined journal-thrust bearing. The details of stiffness and
damping coefficients have been finalized with a bearing supplier and fine-tuned to not allow
the natural lateral and torsional frequencies to fall within the impermissible ranges.

Fig. 3 – 3-D model of the MST-40C HP rotor

8 echnical eview No.17

Choosing the thrust bearing to be located between the HP and IP-LP sections brings about
the advantage of moderate axial expansions to the possible extent for different areas of the
rotors/casings including the most extreme ends. The advantage lies in the selection of the
sealing type, so that labyrinth seals could be used for more areas compared to the case with
large axial expansions. This would provide far less steam leakages, leading to a higher overall
efficiency.

Fig. 4 – 3-D model of the MST-40C IP-LP rotor

Care was also taken to spare the operators the difficulties typically associated with balancing
of the rotor on site by designing an opening on the outer casing’s upper half precisely aligned
with the rotor’s balancing bolts.
A hydro-motor was located at the very end of the rotor train behind the generator, as the
turning gear to rotate the rotor at 60 rpm during startup and shutdown operations. The motor is
fed by the oil coming from the main lube oil system. A clutch is also considered to disconnect
the motor from the rotor train when the train rotating speed reaches 600 rpm.

■■ Casing Design
Steam inlets at both HP and IP-LP sections are located in lower halves, so that the operator
does not need to cut/decouple the inlet pipes to open the upper halves. This would be most
beneficial during maintenance operations and the turbine can get back to operation more
quickly.
The openings allocated for fastening balancing bolts can also be used for both borescope
inspections as well as air dryer connections that come in handy in safe long-term preservation
of the machine.
Both HP and IP-LP outer casings are constrained in three directions within the combined journal-
thrust bearing pedestal, so that they cannot move in any direction at this point. The LP exhaust
casing is fixed laterally and vertically, and can only expand axially. An expansion joint provided
in the cooling system prevents this axial expansion from being transmitted to the condenser.
Similar to the outer casings, the inner casings, i.e. HP and IP inner casings are fixed at one end
and can move at the other within the outer casings.

MAPNA Turbine • March 2022 9

 Fig. 5 – 3-D model of the MST-40C IP-LP outer casing

Casings’ thicknesses are calculated based on steam pressure at each specific compartment.
The HP inner casing is the thickest and LP stationary blade rings are the thinnest of all parts.
There are thermocouple insertion points considered on the HP and IP inner casings to allow
for inner surface temperature readings. These values are assumed equal to the rotor surface
temperature at those points.
Knowing the rotor surface temperature in addition to the steam mass flow rate and temperature,
the temperature gradient across the rotor section could be calculated. This input is required to
provide startup and shutdown curves for the machine.
Water drainage nozzles have been placed within the outer casings wherever needed. The
nozzles are welded onto an inter-connected piping which in turn is routed down to the flash
pipe through drain valves. The valves are open when the turbine is in cold condition and
is starting up. In this phase a mixture of steam and condensate is drained off to ensure no
condensate is accumulated in the lower half of the outer casings. If this does not happen for
any reason, the machine would experience an oval-shape distortion giving rise to blade tips
clearance reduction which might lead to seal rubbing in extreme circumstances.

■■ Auxiliary Systems
MST-40C steam turbine has a set of auxiliary systems arranged round the turbo-generator set,
as follows:
• Sealing system (seal steam, leak off, gland steam condenser)
• Lube oil system (lube oil module, clean and dirty oil tanks, oil purifier)
• Control fluid system
• Drainage system
• Bypass system (IP and LP bypass stations)
• Condensation system (hood spray, flash pipe spray, bypass station spray, etc.)
• Control air

10 echnical eview No.17

The sealing system ensures that there is no air ingress and no steam leakages into the atmosphere
during turbine operation.
The lube oil system feeds the turbine and generator bearings with lubricating and lifting oils at
two different pressures. The lifting oil system lifts the rotor train by 0.1 mm to make sure that no
metal-to-metal contact is happening. This system is kept in service until the train speed exceeds
600 rpm.
The control fluid system provides hydraulic oil for electro-hydraulic actuators. The consumers
include HP, IP and LP steam stop, control and bypass valves. Although the HP bypass valve is
out of MAPNA Turbine’s scope of supply, its required hydraulic oil is provided by the turbine’s
control fluid system.
The drainage system is meant to ensure that there is no condensate accumulation within the
valves and casings. The drain valves are of fail-to-open type to make sure that in case of any
malfunction in the control air system, the drainage system operates properly. The drain valve
actuators are of pneumatic type and provided with control air supplied by an air compressor.
Bypass valves allow for the passage of the steam coming from the HRSG to the condenser
whenever not needed by the machine. They are mainly used during turbine startups and load
rejection modes. At startups, part of steam is admitted to the turbine and the rest flows to the
condenser via the IP and LP bypass valves. In a similar way, the unwanted portion of the HP
steam bypasses the turbine and flows to the CRH pipe. At the design load, the bypass valves
are fully closed.
The condensation system includes water sprays wherever required, e.g. downstream of
bypass valves, downstream of the last moving blade row of the LP section and the flash pipe.
The purpose is to cool down the hot steam to avoid heating and harming the bearings or
condenser. In case of very low steam flow (in load rejection, for instance), steam is circulated
back and forth within the LP end stages causing a temperature rise in that area. This gives rise
to the heating of the LP bearing, leading to non-uniform bearing distortion. As a consequence,
blade tip clearances might be altered which could end up in rubbing in extreme cases.

Concluding Remarks
MST-40C is a three-level-pressure machine designed by MAPNA Turbine. The machine was
designed in close coordination and data exchange with other MAPNA Group companies.
Various LP ends, improved steam parameters and improved auxiliary systems are as among
potential product upgrades in the future.

MAPNA Turbine • March 2022 11

 2
Introduction

M APNA Turbine have unveiled their

ambitions to introduce a new upgrade
for MGT-70 gas turbine platform entitled ‘MGT-
70(4)’ soon. The main goal of this new upgrade
scheme is to achieve enhanced lifecycle for
gas turbine hot components along with higher
power generation and thermal efficiency.
The Turbine Inlet Temperature (TIT) will be
increased to achieve outstanding 190 MW
power output mark, which makes it necessary

MGT - 7 0 (4 ); to reanalyze cooling schemes of gas turbine

vanes and blades and to introduce some

A S ne ak
design modifications to cope with the new TIT.

P eek in to th e
Lates t M G T - 7 0
Gas T u rbine
Upg r ade
S che m e

Faghfoor, Saeed
Poursamad, Amir
Hashemi, Omidreza

MAPNA Turbine Engineering & Manufacturing Co.

(TUGA)

12 echnical eview No.17

Operating MGT-70 Gas Turbine Fleet
There are a number of criteria by which a successful gas turbine upgrade is judged, including
performance, availability, starting reliability, equipment lifecycle and operational flexibility.
Considering current competitive environment, improvement in any of the above-mentioned
areas could lead to product adoption in the market. Following this approach and providing
the operators with beneficial upgrades and improvements, MGT-70 gas turbine has achieved
a remarkable success in the market in recent years.
To date, three upgraded versions of the machine have been introduced to the market, besides
several other improvement packages such as a series of MGTboost-70 platform, IGV+, Wet
Compression and MAPTune-70 to name a few.
Evolution of MGT-70 gas turbine product since its initial version in 2002 by MAPNA Turbine is
shown in Fig. 1. The last upgraded version of MGT-70 gas turbine product, i.e., MGT-70(3), was
introduced in 2017. Thus far, more than 15 new units have been delivered to the customers and
more than 14 older units have been upgraded to this version in Iran and abroad. In about 4
years of operation, the fleet has amassed more than 260,000 operating hours demonstrating
remarkable availability and run/start reliability.

Fig. 1 – MGT-70 gas turbine evolution

MGT-70(4) Upgrade Scheme

Following successful previous product upgrades for MGT-70 gas turbine, the new MGT-70(4)
upgrade scheme has been planned taking into account the following considerations:
The new upgrade scheme is aimed at providing the customers with the minimum operational
costs possible whilst increasing the performance parameters and extending the unit lifecycle
as much as physically practicable.

MAPNA Turbine • March 2022 13

 The main focus of the new upgrade scheme would be to extend the present inspection
intervals, i.e. 33000 EOH (about 4 years of normal operation) to a new outstanding value of
50,000 EOH (about 6 years of normal operation) for a power generation gas turbine unit. To
achieve this ambitious goal, a comprehensive study on mechanical conditions of hot gas path
components in addition to a defect survey on components during operation was conducted.
New gas turbine inspection interval was suggested to reduce the outage period of the unit
required for inspection and refurbishment of parts, leading to improved availability of the gas
turbine. Meanwhile, total costs associated with overhaul inspections during gas turbine lifecycle
will also be reduced.
The new upgrade scheme will bring about improved performance parameters to the highest
values attainable by E-class gas turbines as a result of modifications implemented on gas turbine
vanes and blades cooling scheme allowing for an increased TIT to an unprecedented value
of 1105 °C. Consequently, several hot gas path parts and components will also be modified to
better withstand the higher operating temperatures at these sections of the machine.
While former upgrades sometimes warranted fundamental changes in the plant, this one can
be carried out for in-service units through minor alterations and hence reduction in time and
money spent on the process. Although there were several upgrade scenarios to improve the
performance of different components, some of them did not contribute to the intended goal
of reaching 50,000 EOH limit and hence were ignored.
As a result of design enhancements and modifications made, the upgraded MGT-70(4) gas
turbine will boast increases of up to 31 MW in power, and 2.2% in efficiency with respect to the
original V94.2(3) gas turbine product in simple cycle applications.
Performance parameters of the last two upgraded versions of MGT-70 gas turbine product are
listed in Table 1.

Table 1 – Performance parameters for MGT-70(3) & MGT-70(4) gas turbines

Product Power Output [MW] Efficiency [%]

MGT-70 (3) 185 36.4
MGT-70 (4) 190 36.6

Considering design modifications and preparation of related manufacturing prerequisites and

procedures, the new upgraded product is expected to be launched by mid-2023.

■■ Mechanical Design Improvements

Due to mechanical power increase at the end of driving shaft, there were concerns about the
rotor integrity and attachments. Increasing TIT would also affect hot gas path components within
combustion chamber, as well as vanes and blades of the turbine section. These components
have been studied for strength and lifetime considerations.
In MGT-70 gas turbine platform and to integrate rotor parts, all shafts and disks are stacked
together via Hirth-serration couplings and a central tie-rod and nut fastens all components
to ensure a reliable rotor assembly. With the increase in the mechanical power output of the
gas turbine, the strength and capacity of those couplings and connections of the turbine-
generator coupling shaft needed to be checked. The power increase would also introduce
some minor modifications in the rotor components. The effects of these changes were proved
to be negligible and hence no concern about the shaft train’s dynamic behavior.
The aim of the study conducted on combustion chamber parts was to evaluate the strength
as well as creep and fatigue life assessment. Related mechanical analyses were carried out
based on temperature distributions obtained from the CFD analyses performed.

14 echnical eview No.17

Although creep analysis is deemed to be the most important analysis in investigation of hot gas
path parts and components, the stress analysis is a prerequisite for subsequent analyses. The
results of stress analysis would be used to demonstrate critical points which may have high risks
for initiation and development of any kind of damages such as fatigue crack or creep rupture
initiation.
For life assessment procedure, creep behavior of components exposed to high temperature
must be analyzed. Creep strain limits the life of combustion chamber components such as
mixing chamber and hot gas casing and hence must be kept at a minimum. An innovative
approach to take into account positive effects of compression stresses within a pressure vessel
for creep life assessment of related parts and components was also developed. This scheme
allows for more accurate creep life assessments to be performed in such huge shells.

■■ Gas Turbine Section Modifications

In redesigning the first stage vane of the MGT-70(4) gas turbine and despite the increased TIT,
an attempt was made to reduce the maximum and bulk temperatures of the vane by making
some improvements in the gas turbine blade cooling system. The sum of changes made to
the cooling system resulted in 40 and 9 degrees of Celsius decrease in the maximum and bulk
temperatures of the airfoil, respectively, while keeping the cooling air flow almost unchanged.
For the first stage blade of the MGT-70(4) gas turbine, the airfoil shape and serpentine cooling
design of the MGT-70(3) were left unchanged. However, some modifications were made in
the cooling system to reduce the bulk and maximum temperatures, to change the location
of maximum temperature, and to keep both the blade life and coolant mass flow rate same
as the baseline design despite the increased TIT. Thanks to these modifications, the maximum
temperature decreased by about 10 K, and some major hot spots in the baseline design were
also removed.
Modifications performed on the cooling system of the 2nd stage vanes of the MGT-70(4) gas
turbine, have also led to reductions in critical lifetime parameters such as maximum and bulk
temperatures as well as temperature gradients compared to the base design.

Concluding Remarks
After successful introduction of the last upgraded version of MGT-70 gas turbine product, i.e.,
MGT-70(3) in 2017, a new upgrade scheme is now planned and pursued at MAPNA Turbine to
provide the customers with several outstanding benefits.
The new upgrade scheme is primarily aimed at extending the unit life cycle while improving
the performance parameters of the machine. In this upgraded version, the TIT would be
increased by 15 °C with respect to the previous version. To reach these goals, all major
parts and components of the machine are assessed and analyzed to make sure that they
can withstand the more severe working thermo-mechanical conditions they are going to be
exposed to, considering the new extended maintenance intervals of the machine. As a result
of the analyses performed, only the first stage vanes and blades in addition to the second
stage vanes of the gas turbine would be redesigned and replaced in MGT-70(4) with respect
to MGT-70(3) gas turbine product.

MAPNA Turbine • March 2022 15

 3
Introduction

G as turbines are generally designed for

ISO standard conditions, and changes
in ambient conditions such as temperature,
pressure, and humidity substantially affect
their performance; therefore, it is essential to
provide solutions to prevent or diminish such
unfavorable degradations. Temperature is
the most critical ambient condition of all
and an increase in this parameter leads to
decreased air mass flow rate and hence a

MGT - 4 0 decrease in the power output and efficiency.

It is estimated that, every 1 K increase in the

Gas T u rbine
ambient temperature from ISO conditions,
leads to 0.5 to 0.9% drop in the power output

Com pre s s o r
[1] and 1% decrease in thermal efficiency of
the machine [2].

Modif ic atio n In the present article, a method is presented

to improve the performance of MGT-40 single
shaft gas turbine at high ambient temperatures
by changing the angle of compressor
stator vanes. First, a three-dimensional
simulation was carried out for the compressor
compartment using a commercial CFD solver;
afterwards, the gas turbine’s performance
was calculated using an in-house cycle
analysis code. Next, design variables with the
most significant impact on the power output
were selected using the design of experiment
(DoE) approach and finally, optimization
was performed to find the optimum stagger
angles. The optimization process is based on
a genetic algorithm coupled with an artificial
neural network.

Azizi, Rohollah
Ziamolki, Alireza
Nazari, Mohammad

MAPNA Turbine Engineering & Manufacturing Co.

(TUGA)

16 echnical eview No.17

Numerical Simulations
Performance of the MGT-40 gas turbine compressor comprising Inlet Guide Vanes (IGVs), 17
stages of rotary and stationery blades, as well as two rows of Exit Guide Vanes (EGVs), as shown
in Fig. 1, was investigated in the present study using a 3-D finite volume CFD solver. Steady-state
solution was used to solve the Reynolds-averaged Navier-Stokes (RANS) equations within the
domain under investigation.

Fig. 1 – Meridional view of MGT-40 gas turbine axial compressor

Cycle Performance Analysis

The objective of this investigation was to enhance the performance of the MGT-40 gas turbine’s
cycle during hot seasons. Therefore, the effects of the compressor blades’ re-staggering on the
cycle performance had to be studied. Related thermodynamic simulations were performed
using in-house software called MapCycle, previously developed for performance analysis of
single and multi-shaft gas turbines.
MapCycle inputs include environmental conditions, axial speed, pressure drop at the
compressor inlet, and inlet mass flow rate. It provides parameters such as temperature, pressure
and mass flow rate at different gas turbine components, in addition to the characteristics
of the compressor performance such as pressure ratio, efficiency, and corrected mass as
outputs. A graphical representation of MapCycle’s architecture indicating various inputs and
outputs at different parts of the code corresponding to different parts and components of a
turbogenerator set is shown in Fig. 2.

Fig. 2 – MapCycle architecture for performance analysis of a turbogenerator set

MAPNA Turbine • March 2022 17

 This software is used to evaluate performance of different gas turbines of various configurations,
including MGT-40 gas turbine, as schematically presented in Fig. 3.

Fig. 3 – Schematic representation of a typical single shaft gas turbine simulated in MapCycle

The results provided by the cycle analysis code are in good agreement with the experimental
data, as shown in Fig. 4.

Actual correction curve

Cycle analysis code results
Power Factor (%)

Ambient Tempreture (°C)

Fig. 4 – Power factor versus ambient temperature for MGT-40 gas turbine

An optimization procedure was utilized to enhance performance of the MGT-40 gas turbine in
hot ambient conditions. First, a commercial CFD solver was used to find the impacts of stagger
angles on the compressor performance, and subsequently, MapCycle code was utilized to
calculate the engine performance.
According to the design of experiment (DoE) analysis results (Taguchi method [3]), engine
power at hot ambient conditions is mostly affected by the first five compressor stator vanes.
Hence, these five design parameters were considered for optimization purposes. An automatic
optimization process based on an internally developed code was used in this study. Genetic

18 echnical eview No.17

algorithms (GAs) and artificial neural networks (ANNs) were coupled with a commercial CFD
solver to speed up the process. A flowchart of the applied optimization procedure is depicted
in Fig. 5.

Start

Generating the initial database

Adding new results
to the database
Training the ANNs

Optimization by GA and
calculation of the objective
function by ANNs

Solving the optimum configuration

by CFD solver

Conducting thermodynamic
cycle analysis

Meeting the No
convergence criteria

Yes

End

Fig. 5 – Optimization procedure used for MGT-40 gas turbine performance improvement

Direct use of CFD solver leads to a significant increase in calculation time; to reduce this
time, ANN was used to predict the engine performance. ANN is made up of several primary
processing units called neurons or nodes. The primary purpose of a neural network is to use the
information stored in a database to relate design variables of a problem to the corresponding
objective functions.

Results & Discussion

Optimal compressor stators’ stagger angles contributing to MGT-40 gas turbine performance
improvement as the objective function of this study were calculated following implementation
of the developed optimization procedure up until the relevant convergence criteria were met.
Compressor maps are presented in Figs. 6 & 7 for original and optimal geometries at various
rotational speeds. From these figures it is clear that compressor mass flow rates are noticeably
increased in cases of optimum compressor stators’ stagger angles at all operating conditions.
MGT-40 gas turbine performance parameters at standard ISO and hot ambient conditions
are listed in Table 1 for both the original and optimized compressor blades’ geometries. As it is
indicated in Table 1, the power output increased by 1.69% and 3.64% at standard ISO and hot
ambient conditions respectively.

MAPNA Turbine • March 2022 19

 Table 1 – MGT-40 gas turbine performance parameters

Operating Original Optimum

Parameter Variation
Conditions Geometry Geometry
683
Power (kW) 40410 41093
Standard ISO (+1.69 %)
Efficiency (%) 30.67 30.59 -0.08 %
1,222
Power (kW) 33614 34836
Hot Ambient (+3.64%)
Efficiency (%) 28.81 29.02 +0.21%

Comparison between the results obtained for original and optimal gas turbine compressor
blades’ geometries clearly indicates that the aim of this investigation, i.e. increasing the MGT-
40 gas turbine power output at hot ambient conditions whilst maintaining its power output and
cycle efficiency levels at standard ISO conditions, has been achieved.

Fig. 6 – MGT-40 gas turbine compressor map for baseline and optimal compressor blades at
90%, 95% and 100% rotational speed

Fig. 7 – MGT-40 gas turbine compressor map for baseline and optimal compressor blades at
100%, 105% and 110% rotational speed

20 echnical eview No.17

Concluding Remarks
The present work was aimed at improving MGT-40 gas turbine performance at higher
ambient temperatures, while maintaining its performance at standard ISO conditions through
optimization of stagger angles of gas turbine compressor stator blades. To do so, a variety of
tools and software including a commercially available CFD solver, an internally developed
cycle analysis code, an evolutionary algorithm coupled with artificial neural networks were
utilized. Because of the rather large number of compressor stages, the main challenge for
the optimization algorithm to proceed was the high computational time and costs associated
with it. So, DoE techniques were utilized to find out which variables significantly impacted the
engine performance and it was eventually found that the first five stator blades’ stagger angles
had the dominant effects on engine power output. Therefore, these five angles were used
as design variables in the optimization process. Additionally and due to high computational
time associated with the utilized CFD solver, a Genetic Algorithm was coupled with a Neural
Network to significantly reduce the time required to calculate the objective function. Eventually,
the results indicated that the gas turbine’s power output increased by 1.69% and 3.64% at
ISO standard and hot ambient temperature conditions, respectively. This was achieved via
changing only the first four rows of compressor stators’ stagger angles as well as that of the
IGV’s.

References
[1] M. Ameri, H. Nabati and A. Keshtgar, “Gas turbine power augmentation using fog inlet
air-cooling system”, Engineering Systems Design and Analysis: ESDA2004-58101, PP. 73-78, 2004
[2] A. D. Sa and S. Al Zubaidy, “Gas turbine performance at varying ambient temperature”,
Applied Thermal Engineering, Vol. 31(14-15), PP. 2735-39, 2011
[3] S. D. Bolboacă and L. Jäntschi, “Design of experiments: Useful orthogonal arrays for number
of experiments from 4 to 16”, Entropy, Vol. 9, PP. 198-232, 2007

MAPNA Turbine • March 2022 21

 4
Introduction

S elective Laser Melting (SLM) is an additive

manufacturing (AM) technique during
which high-power-density laser is used to form
consecutive metal layers on top of each other
by melting and fusing a metallic powder bed
according to a predefined model. Since each
layer is gradually cooled before the formation
of the next, the component, not cooled
homogenously, is subject to local heat input
and severe thermal gradients causing residual

E v a l u at ion deformations/distortions in addition to stresses

which can render the final AM part unfit for

o f Residu al
use.
Local melt pool geometry, scanning strategy,

D i s t or t ion s and part geometry are some of the contributing

factors to the extent of residual macroscopic

t h ro ugh distortions in additively manufactured parts.

While the melt pool geometry and scanning

S i m u lat ion of
strategy (including laser speed, path/pattern,
hatch spacing and strip width) mainly affect
the local microstructures and anisotropy
S c a n P at t er n s of the resulting distortions as well, the part
and support geometries as well as building
& S uppor t s orientation play vital roles in the build-up of
macroscopic residual distortions/stresses in
f o r I ncon el additively manufactured parts [1].
The complex nature of the SLM method and a
62 5 Addit iv e plethora of contributing parameters affecting
design and manufacturing processes of the
M a nu f act u r in g parts have prompted the designers to take
advantage of the available CAE and FEA tools
Pro c ess to identify potential failure modes, eradicating
the need for numerous trial and error attempts
and allowing for prior adjustment of the parts’
geometries and machine parameters to
minimize the risk of costly printing errors.
There are many parameters to be considered
to achieve an acceptable work piece during
the SLM process. In the second article of
MAPNA Turbine’s Technical Review No.16, the
authors determined the effects of heat density
on uniformity of the work piece surfaces,
amount of residual stresses, dimensional
Fathi, Alborz tolerances, microstructures as well as size and
Ghodsi, Mahsa shape of defects. It was also mentioned that
Davarzani, Hossein the heat density level is directly impacted by
scan strategy. Scan strategy describes all the
MAPNA Turbine Engineering & Manufacturing Co.
parameters to determine how to move laser
(TUGA)
point on the powder bed to scan the whole

22 echnical eview No.17

selected area. The scan strategy mainly affects residual stresses as well as roughness of the
part’s final surfaces [2].
The CAE software packages are needed to be calibrated and validated to be reliable in
predicting thermal residual stresses within final additively manufactured parts. The goal of the
calibration process is to obtain related correction factors for the deployed CAE software. These
correction factors help designers to simulate and evaluate the parts prior to the start of AM
processes with more reliability, making it possible to modify the supports and redesign the parts
if needed.
The present study focuses on the effects of scan patterns and support geometries on residual
stresses and distortions during SLM additive manufacturing process of Inconel 625 metal powder
on a NOURA M100P machine with a layer thickness of 30 microns. It also investigates how FE
software may be used to predict and obtain more accurate residual distortion and help select
suitable support geometries prior to the fabrication of any part utilizing the SLM process.
The SLM machine was used to manufacture several parts with specific dimensions. Two different
supports and various scan strategies were adopted, and then the actual and numerical results
for strain distortions were compared. Correction factors with different scan patterns or scan
strategies were obtained. Finally, to confirm and validate the FE coefficients and the suitability
of the scan pattern, the numerical and actual results were re-evaluated for a selected group
of parts with optimized support.

Simulation Process & Parts Manufacturing

Simulating the build process may be performed at various points along the overall design/
manufacturing process depending on the goals pursued. FE software generates solutions to
residual stress, distortion, and building failure issues, hence enabling the user to improve the
product design, and to inform and validate the build preparation. The overall workflow of the
steps taken to attain final distortion values in the present study is shown in Fig. 1.

STEP 1 STEP 2 STEP 3

Calculating the new SSF and
Simulating the manufacturing
Manufacturing the part with (ASCs) and obtaining the scan
process and obtaining the
SLM machine and measuring pattern and thermal strain
distortions via simulation
the distortions analysis results for support
results
optimization
• Models are designed with • Assumed strain and scan
primary support pattern simulations are run for • Corrected SSF is calculated

• The Scan Pattern is primary support • Scan pattern simulation is run

determined • Machine parameters are set with corrected SSF and new
ASC is found
• Models are built on the • Linear Elastic stress mode is set
platform • Support’s design is optimized
• Strain Scaling Factor (SSF) is based on simulation results
• Distortions are measured set
before supports’ cut-off • Simulation distortions are
• Distortions are obtained compared with recorded
• Distortions are measured after for the required points and measurements
supports’ cut-off directional component

STEP 4

• Corrected SSF and ASC factors are set and validation analysis is run for the design with new support
• Simulation results are compared with control sample measurements and the simulation error is calculated

Fig. 1 – Overall process workflow of the steps taken to attain final distortion values in this study

MAPNA Turbine • March 2022 23

 1st Step: Manufacturing the Parts with SLM Machine & Measuring the
Actual Distortions
■■ Models Designed with the Primary Support
For convenience and to maintain the same manufacturing conditions, all parts were built on
the same building platform with the same process parameters optimized using the method
described in the second article of MAPNA Turbine Technical Review No.16 [2]. All simulation
and manufacturing steps were performed on the same cantilever beam model with only
scanning strategies and supports (primary and new design) varied. 3-D representations of the
cantilever beam model and the holding supports are shown in Fig. 2.

Fig. 2 – Primary model and the holding supports used for simulation and manufacturing
purposes

All parts were printed using a NOURA M100P SLM machine in an inert argon atmosphere using
Inconel 625 powder with a layer thickness of 30 µm and with two types of support structures
directly mounted on the base plate (Stainless steel 316L, 25 cm in diameter and 4 cm high).

■■ Design & Determination of Scan Patterns

The CAD model and the STL file were prepared for manufacturing. In the slicing software,
primary supports were designed and the scan strategies as well as the hatch spaces were
specified.
There are several standard machine parameters, also known as process parameters, used in
additive manufacturing simulations. All necessary parameters needed to be determined for
scan pattern analysis are listed in Table 1. Each simulation (except the assumed strain simulation)
uses some of or all these parameters.

Table 1 – Specifications of different scan patterns used with the manufactures parts
Layer
Scan Pattern Starting Hatch Model with Primary
Scan Direction Rotation
Strategy Angle feature support
Angle
Parallel
Scan Pattern 1 0° 0° Meander
to X direction

Perpendicular to
Scan Pattern 2 90° 0° Meander
X direction

Rotating stripe
Scan Pattern 3 (optional scan 0° 67° Meander
pattern)

Scan Pattern 4 TUGA typical 17° 37° Chess-board

24 echnical eview No.17

 ■■ Building the Model with Primary Support on the Platform
After modeling and determining the manufacturing strategies, supports were designed for
each model. The .CLI files were also created with the slicing software for SLM Machine, and
finally the samples were manufactured accordingly, as shown in Fig. 3.

Fig. 3 – Designed part with the primary supports (left);

built models using the SLM machine (right)

■■ Measurements of distortions prior and after cutting off the supports

For quantitative characterization of residual distortions after printing according to the proposed
instructions, the parts were meticulously measured using digital gauges in two instances of
before and after cutting off the supports from the building platform.
The measurement paths for each of the mentioned cases are shown schematically in Fig. 4.

X component of displacement
is of interest

Measurment location A Measurment location B

Fig. 4 – Schematic representation of measurement paths A & B

The line charts of actual distortion values measured along path B of the additively manufactured
parts are plotted in Fig. 5. According to the results, scan patterns 3 & 4 yielded nearly the
same results and so the maximum distortion values belonging to the scan pattern No. 4, i.e.
chessboard, was considered and presented in Fig. 5.

MAPNA Turbine • March 2022 25

 Distortion ]mm[

Measurement points along path B

Fig. 5 – Actual distortion values measured along path B for different AM parts manufactured
with different scan patterns following cutting off the supports

2nd Step: Simulation of the Manufacturing Process

The objective of the second step was to determine the Strain Scaling Factor (SSF) and Anisotropic
Strain Coefficients (ASCs) for thermal strain analysis calculations. The calibrated SSF and ASCs
will significantly improve the prediction accuracy of the simulation software, and hence the
chance of successful builds along with reducing the cost of trial-and-error experiments. In this
study, a purely elastic model was used to achieve a more accurate prediction in the residual
stress level as well as the associated residual deformation. The values for SSF and ASCs depend
on the following parameters:
• Material
• SLM machine used
• Machine parameters including laser power, scan speed, layer thickness, base plate
temperature, hatch spacing, slicing strip width, scan pattern, etc.
• Simulation type performed (assumed strain, scan pattern, or thermal strain)
• Selected stress mode (linear elastic or elastic-plastic)

■■ Assumed Strain (AS) Method

Assumed strain (AS) analysis is the simplest and fastest simulation type. It assumes a constant,
isotropic strain at every location within a part as it is being built. Anisotropic effects or process
dependencies are not directly considered and AS relies solely on material and machine-
specific calibrations of the isotropic SSF. Neither process nor scan strategy parameters are
required for this simulation type.
SSF is an important factor quantifying the variables specific to each build scenario. It must be
experimentally determined for each combination of machine/material/strain/stress mode.

■■ Scan Pattern (SP) Method

The scan pattern (SP) mode uses the same strain as the assumed uniform strain, but it takes
the scanning direction into account, as well. Larger distortions occur along the laser scanning
path than the perpendicular direction. Unlike isotropic distortions in which only SSF needs to be
calibrated, in anisotropic distortions, calibration of ASCs is also necessary.

26 echnical eview No.17

For each scan pattern, the following equations were used to achieve new SSF and ASCs factors
which are dependent on the scanning direction.

Where:
ASC stands for anisotropic strain coefficient
δePar means parallel to the scan direction and
δever means perpendicular to the scan direction

■■ Thermal Strain (TS) Method

The thermal strain (TS) simulation type is the most accurate approach as thermal ratcheting
effects and inhomogeneous cooling are taken into account. Since this method requires a
thermal prediction for every scan vector, a much longer computation time is required. Like the
SP approach, both isotropic and anisotropic strain scaling factors need to be calibrated based
on the measured data.
In addition to the material and scan strategy parameters, TS approach also requires the actual
process and scan strategy parameters like the laser power and speed, base plate temperature
and layer thickness as well as hatch spacing and stripe width to be set.

3rd Step: Calculating New SSF & ASCs and Designing a New Support
■■ Assumed Strain (AS) Analysis
Maximum distortion values were recorded once the AS simulations were conducted. Since this
approach does not account for anisotropic effects, the predicted width deviations were equal
in both X and Y directions. As mentioned before, the AS simulation approach is not dependent
on the scan pattern and hence chessboard pattern was used. The simulations were repeated
until the errors became less than the allowable limit (<1%).

■■ Scan Pattern (SP) Analysis

Consideration of the anisotropic effects through using this approach improves the prediction for
both bidirectional and rotating scan strategies. The maximum distortion values of the cantilever
beam and the supports calculated using this analysis method is shown in Fig. 6.

Fig. 6 – Maximum distortion values in the main part (left) and in the supports (right) calculated
using the SP simulation method

MAPNA Turbine • March 2022 27

 For both the parallel and orthogonal modes, the preliminary results were extracted. For error
rates greater than the allowable limit, i.e. above 1%, the analyses have been repeated till the
errors were brought within the acceptable range.

■■ Thermal Strain (TS) Analysis

After the scan pattern analysis, the thermal strain analysis using the primary data was conducted.
The results for both parallel and orthogonal modes are displayed in Fig. 7. The analyses were
again performed using the new data acquired at each step till the errors reached the allowable
limit, i.e. below 1%.

Fig. 7 – SSF & ASCs calibration for TS simulations

■■ Simulation of Residual Deformations & Extraction of Correction Factors

Preliminary results from the first simulations carried out for each type of analyses were recorded
and recalculated based on the relationship between the new coefficients and the error rate.
For error rates greater than 1%, the simulation process had to be repeated with the latest
acquired data. The final results of the analyses carried out to obtain the SSF parameter, as well
as ASCs at different modes are summarized in Table 2.

Table 2 – Final results obtained for SSF & ASCs following conclusion of all analysis schemes

Summary of Results
Stress
Material Assumed Strain Scan Pattern Thermal Strain
Mode
SSF 1.5 SSF 0.711
Linear ASC ║ 1.29 ASC ║ 1.331
Inc625 SSF 0.93
Elastic ASC ┴ 0.71 ASC ┴ 0.669
ASC_z 1 ASC_z 1

Results of the TS simulations performed for calculation of the maximum distortion utilizing the
data specified in Table 2 indicate 0.0004 and 1.671 mm distortion values for the beam and
supports respectively prior to and after removal of the supports, as shown in Fig. 8.

28 echnical eview No.17

 Fig. 8 – Maximum distortions obtained by TS simulations prior (left)
and after (right) removal of the supports

■■ Designing a New Support

To reduce the maximum distortion encountered, a new support was designed based on the
primary simulation results from the TS analysis. More supports with less distance between them
were put in place. Once all the simulations and AM process of the new parts concluded, the
results were obtained and compared to each other. An integrated model of the part and
supports as well as associated AM built models are shown in Fig. 9.

Fig. 9 – Designed part with newly developed supports (left);

built models using the SLM machine (right)

4th Step: Validation of Calibration Factors with the New Support

■■ Residual Deformations
In order to confirm the results, other tests were also performed on the same AM part using
the newly developed support. The dimensions of the part and initial settings - except for the
manufacturing strategy which was chosen to be of the chessboard pattern with changing
angles in each layer as well as the rotation angle - were all kept unchanged.
The line chart of actual distortions measured at the specified points previously indicated along
the AM parts are plotted in Fig. 10. The maximum actual distortion measured to be 1.48 mm
in the case of scan pattern No. 4 which yielded acceptable results in terms of mechanical
properties.

MAPNA Turbine • March 2022 29

 Scan pattern_4
Distortion ]mm[

Measurement points along path B

Fig. 10 – Actual distortion values measured along path B of the AM part manufactured with
scan pattern No.4 following cutting off the new supports

■■ Validation of Calibration Factors With Newly Developed Supports

Thermal strain analysis was performed for the part with the new support taking advantage of
the correction factors listed in Table 2. According to the results of the simulations performed, the
maximum distortion values of 0.003 and 1.45 mm were predicted for the beam and supports
respectively prior to and after removal of the supports, as shown in Fig. 11.

Fig. 11 – Maximum distortions obtained by TS simulations prior (left) and after (right) removal of
the new supports

Concluding Remarks
Simulation of the additive manufacturing process plays an important role in the design of AM
parts, and is a key to the transition from a mere experience-based to a more knowledge-driven
design process that exploits the full potential of this manufacturing technique.
This study was aimed at presenting a fresh insight into the current capabilities of FE software
and may serve as a benchmark study for future investigations. The following points should be
considered to get a realistic view of the applicability and reliability of the simulation approaches:
• All presented simulation approaches provide reliable predictions of shape deviations in
the investigated parts that are within measurement uncertainties of less than 1%.
• According to the results obtained from the calibration, applying the determined
correction factors to any other Inconel 625 AM part printed by NOURA M100P can
guarantee probability error of less than 1% for residual deformations in FE analysis.

30 echnical eview No.17

 • The final simulation results and measured distortion values subsequent to using newly
developed supports show that residual distortion could be reduced by a support that is
redesigned based on the simulation data.
• The residual distortion is sensitive to the scan pattern; and chessboard pattern is more
suitable than merely longitudinal or transverse scanning schemes.
It is highly recommended to recalculate the correction factors for FE software upon any change
in the machine, material or layer thickness. It is also suggested that the same calibration process
be performed for different materials (specifically 316L Stainless Steel, Titanium and other Inconel
alloys), geometries, support structures, process parameter sets and other machines.
Having the discussed software values also contributes to prediction of the adequacy of the
supports, the suitability of the part concerning the platform, and the level of residual stresses
on the part before manufacturing. This can significantly reduce the cost and time of the AM
process.
Verification and validation test determined that regardless of the support, part model, and
scan strategy, if the machine and material are not changed, the simulation results will remain
valid with a high probability percentage.

References
[1] T. Mayer, “Simulation and validation of residual deformations in additive manufacturing of
metal parts,” Heliyon, Vol.6(5), 2020
[2] “First MGT-30 Gas Turbine Additively Manufactured Parts set to render conventional
manufacturing process of the part obsolete,” MAPNA Turbine Technical Review, No. 16, Sep.
2021

MAPNA Turbine • March 2022 31

 5
Introduction

C ontinuous development and looking for

new solutions in bringing more efficiency
into our production lines have always been
given a top priority at MAPNA Turbine.
MGT-70; our flagship gas turbine, has been
in our product portfolio since 2002 and
the present article outlines two recent
improvements made in its assembly and
machining processes.

More
S treamline d
& E ff ic ie n t
MGT - 7 0
Gas T u rbine
Manufacturing
P roc e ss

Shiri, Alireza
Mehrali, Hossein

MAPNA Turbine Engineering & Manufacturing Co.

(TUGA)

32 echnical eview No.17

Eliminating Turbine Dummy Shafts
Traditionally dummy shafts were used for the adjustment of lower casing and hot gas parts as well
as to align the turbine and compressor bearings with blade carrier I. Taking novel approaches
in the assembly of the MGT-70 gas turbine components such as doing the assembly of the
center casing and exhaust casing vertically and not horizontally has rendered turbine dummy
shafts obsolete.
This way, casings’ diameters are extracted from quality control reports at the conclusion of the
machining processes and hence the thickness of the needed shims to fill the gap between the
casings can be calculated. By using appropriate shim sizes evenly placed on both sides, the
casings will be put in their best aligning positions.

Fig. 1 – MGT-70 gas turbine final assembly processes in progress at MAPNA Turbine workshops

This method is also utilized for all lower casings, so as the corresponding piping runs for the
casing I and the pedestal are implemented in the vertical position with much better ergonomic
outcomes as well as reduced assembly time.
So, in addition to reaching the mentioned goal of eliminating turbine dummy shafts, some
machining operations and dimensions were also changed. For instance, the diameter of the
hot gas drain hole was decreased by more than 15 mm.

Elimination of the Second Step of Rotor Blades Grinding Process

Grinding of the MGT-70 gas turbine blade tips was traditionally carried out via a two-step
procedure. In the first step, blade tips were roughly ground and the rotor was then installed
within the casings. Radial gaps were then measured following tightening of the casings’ bolts
and final grinding totals required were calculated. The gas turbine was then disassembled in

MAPNA Turbine • March 2022 33

 the second step to bring out the rotor to carry out the final blade tip dressing. The finished rotor
was once again installed in place and the gaps were checked to verify the blade tip gaps in
a trial and error procedure.
So, a careful analysis was performed in an effort to eliminate the second step grinding of the
rotor blades and to do it all at once via a single-step procedure.
Initial assessments of the procedure represented more stable and uniform grinding values for
compressor blades in comparison with gas turbine blades and hence brought about different
treatment of each.

■■ Compressor Blades
Initial grinding values were analyzed for 13 recent gas turbines manufactured to find out
an optimum value for each compressor stage. Initial grinding values measured at the 2nd
compressor stage are represented in Fig. 2. Excluding the initial grinding value corresponding
to the 164th gas turbine unit as an outlier in the provided dataset, would yield an optimum
grinding value of 0.2 to 0.3 mm for this stage of the compressor as shown in Fig. 3.

Fig. 2 – Initial grinding values at the 2nd compressor stage of the MGT-70 gas turbine;
units 164-177

Fig. 3 – Initial & optimum grinding values at the 2nd compressor stage of the MGT-70 gas
turbine; units 165-177

34 echnical eview No.17

In the next step, 0.25 mm final grinding value was theoretically considered for several gas
turbines deemed critical (units ground with values far from the mentioned optimum value) to
calculate final resulting gaps. Final gaps calculated for the 3 critical units are listed in Table 1.
According to the results, applying final grinding value of 0.25 mm for the 2nd stage compressor
blades would have resulted in acceptable final gap values in 12 of the 13 cases investigated.
Measured gap values at just 3 points of the rotor belonging to the 164th gas turbine unit
exceeded the acceptable range by less than 0.1 mm (marked in red color in Table 1).
Table 1 – Critical cases gap measurement data

Critical Units (Grinding 0.25mm)

Nominal Gap: 1.8-2.2 mm
Gap Real
U L UL LR LL UR
2.13 2.06 2.09 1.93 2.06 1.95
U164
Gap IF grinding 0.25
U L UL LR LL UR
2.32 2.25 2.28 2.12 2.25 2.14
Gap Real
U L UL LR LL UR
2.12 2.19 2.09 2.1 2.1 2.06
U174
Gap IF grinding 0.25
U L UL LR LL UR
2 2.07 1.97 1.98 1.98 1.94
Gap Real
U L UL LR LL UR
2.09 2.15 2.02 2 1.98 2
U176
Gap IF grinding 0.25
U L UL LR LL UR
1.98 2.04 1.91 1.89 1.87 1.89

The same approach was applied on all 16 compressor stages to find out optimum grinding
values for each stage. For the 182nd gas turbine unit, 50% of the predicted grinding values
were applied on compressor blades and these values were corrected following final gap
measurements (other parameters such as the over speed test precedence could have
impacted these values).
For the next gas turbine, 70% of the corrected values were applied on compressor blades and
finally for the 185th gas turbine unit, the predicted values were applied completely and final
gap measurements revealed that all gap values were acceptable, as shown typically in Fig. 4,
for the first stage compressor blades.

MAPNA Turbine • March 2022 35

 Fig. 4 – First stage compressor blades gap measurement data

■■ Turbine Blades
Initial investigation of the grinding values applied at the turbine side revealed that the variations
at several blade rows were more than radial gaps’ tolerances, as shown in Fig. 5.

Fig. 5 – Measured grinding values at the 4th stage of the MGT-70 gas turbine; units 170-180

36 echnical eview No.17

So, it seemed highly risky to determine an appropriate fixed grinding value for each gas turbine
stage and hence the two following approaches were taken:
►► Direct Measurement of Rotor & Stator Diameters
Theoretically, gap values can be calculated when the diameter of male (rotor) and female
(stator) pieces are measured. To do so, it is necessary to find out the exact axial position of the
rotor inside the turbine carrier, as represented in Figs. 6&7, so the axial locations for measuring
gaps on the stator part could be determined.
This was carried out for the 182nd gas turbine unit and the theoretical results obtained showed
a quite acceptable conformity with the actual gap measurements carried out.

Fig. 6 – Pinpointing the axial position of the stator diameter on the MGT-70 gas turbine blade
carrier

MAPNA Turbine • March 2022 37

 Fig. 7 – Calculations performed for determination of the axial position of the stator diameter
on the MGT-70 gas turbine blade carrier

Although application of such a method yielded quite acceptable results, there are still some
difficulties as mentioned below:
• Setting the turbine carrier on the measuring platforms for diameter measurement is
difficult and highly risky for the carrier part
• Any mistake in finding the exact axial position of measuring points on the turbine carrier
would have a significant impact on the obtained results. A negligible mistake of 1 mm
in the estimation of the exact axial position would lead to a 0.6 mm change in the
measured diameter.
• While doing measurements, the stator is measured and once calculations are finalized,
the rotor should be ground simultaneously. Planning and synchronizing these activities
are quite difficult as a routine manufacturing program.
Therefore, a second approach was taken to eliminate gas turbine second step grinding
procedure in the turbine side, as mentioned below:
►► Assembly Optimization Combined With Statistical Methods
Utilizing this method, all parameters with potential significant impacts on variation of gas turbine
rotor blades’ grinding values were addressed to decrease the amount of observed variations.
One of the most important parameters greatly impacting the grinding value of the turbine side
rotor blades of different stages is the axial position of the turbine blade carrier.
To address this issue, a comprehensive study was carried out to install the turbine blade
carrier in a fixed position using the shims sized accurately following some careful analyses and
calculations.

38 echnical eview No.17

Other major parameters were also adjusted and dealt with delicately to rule out the second
step of rotor blade tip grinding. These parameters are as follows:
• Retrospective analysis of axial clearances in previous gas turbines showed that in 85% of
cases, it is possible to move the rotor for about 0.6 mm in the positive direction, resulting
to almost 0.2 mm change in the radial gaps at the turbine side
• Review and revision of manufacturing processes of stator blades to have quite fixed
final stator diameters
• Analyzing out-of-range gap values which could be acceptable in non-conformed cases
and the direction of such non-conformities (plus or minus) for increased maneuverability
Taking into account all of the above, and providing appropriate solutions, grinding values were
measured in several gas turbines, as shown in Fig. 8. The predicted values were then calculated
and applied in part for the next two subsequent gas turbine rotors being manufactured. Finally,
following some fine tuning, the predicted values were applied completely on a gas turbine
rotor with acceptable gap values in all directions, as shown typically in Fig. 9, for the 4th stage
gas turbine blades.

Fig. 8 – Measured & optimum grinding values at the 1st stage of the MGT-70 gas turbine; units
176-180

MAPNA Turbine • March 2022 39

 Fig. 9 – 4th stage gas turbine blades gap measurement data; unit 185

The Bottom Line

This article outlined the steps taken to make manufacturing processes of MGT-70 gas turbines
more efficient by ruling out and/or optimizing some burdensome and time consuming processes
of using dummy shafts for the alignment of different gas turbine components and two-step rotor
blade tip grinding procedure. In the course of this project, statistical tools and methods were
utilized and several assembly procedure change orders were issued to successfully accomplish
the jobs. Consequently, the machining hours and man-hours required for the grinding procedure
of the MGT-70 gas turbine rotor were decreased by 50 hrs and 500 hrs, respectively. The critical
path of gas turbine final assembly was also reduced by almost 6 working shifts.

40 echnical eview No.17

 Head Office:
231 Mirdamad Ave. Tehran, I.R.Iran.
P.O.Box: 15875-5643
Tel: +98 (21) 22908581
Fax: +98 (21) 22908654

Factory:
Mapna blvd., Fardis, Karaj, I.R.Iran.
Post code: 31676-43594
Tel: +98 (26) 36630010
Fax: +98 (26) 36612734

www.mapnaturbine.com
tr@mapnaturbine.com

© MAPNA Group 2022

The technical and other data

contained in this Technical Review
is provided for information only
and may not apply in all cases.
GR530-0