Available online at www.sciencedirect.

com

ScienceDirect
Procedia Engineering 138 (2016) 337 – 346

“SYMPHOS 2015”, 3rd International Symposium on Innovation and Technology in the Phosphate
Industry

Gas Turbine: Optimization of Energy Production and High

Efficiency by Using Power Electronics
Aziza Benabouda*, Alfred Ruferb
a
Royal Navy School, Sour Jdid, Casablanca, 22001, Morocco
b
Swiss Federal Institute of technology, LEI / EPFL, Lausanne, Switzerland

Abstract

The use of gas turbines in the power generation industry has been increasing. Their demand represents more than 50% of the
world market of thermal power plants. Electric power generators using gas turbines as power sources are connected to the
turbines through a mechanical gearbox, in order to adapt their synchronous speed to the optimal rotation speed of the turbine,
which is very often much higher than the synchronous speed. However, due to direct network connection, the generator speed
cannot be variable: it is imposed by the network and constant. To overcome this problem, we propose to replace the mechanical
gearbox by a flexible electronic solution which offers a high efficiency. Using this approach, the turbine is directly connected to
the synchronous generator, which is connected to the grid through an indirect static frequency converter. However, this type of
converter is not common in this application because of very high switching losses due to the high frequency of the PWM
technique used for its control. In this paper, a new control strategy is proposed for the multi-level converter, characterized by its
high efficiency due to the use of square-wave modulation. The main advantage of this mode is the quasi absence of switching
losses. In this mode, only the frequency can be varied between the input and the output voltage, but their magnitudes are not
freely controllable. A voltage magnitude adaptation can be done by the generator's excitation. The produced active and reactive
power can be controlled by the generator excitation as well as both the angle shift between the generator and rectifier voltages
and between the inverter and network voltages. Simulation and experimental results for different operating points highlight the
capabilities of the proposed control strategy. These include the ability to operate with unity power factor and better current
quality.

©©2016
2015TheTheAuthors. Published
Authors. by by
Published Elsevier Ltd.Ltd.
Elsevier This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of the Scientific Committee of SYMPHOS 2015.
Peer-review under responsibility of the Scientific Committee of SYMPHOS 2015

* Corresponding author.
E-mail address: [email protected]

1877-7058 © 2016 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of the Scientific Committee of SYMPHOS 2015
doi:10.1016/j.proeng.2016.02.094
338 Aziza Benaboud and Alfred Rufer / Procedia Engineering 138 (2016) 337 – 346

Keywords: gaz turbine, multilevel converter, square wave modulation, switching angles, high efficiency, angle shift control;

1. Introduction

The gas turbine has seen a recent rise in popularity within the power generation industry. Until the end of the
1960s gas turbines were almost exclusively the preserve of the aviation industry. During the 1970s and 1980s they
started to find favor as standby and peak power units because of their facility for rapid start-up. It was during the
1990s, however, that they became established, so that by the end of the twentieth century the gas turbine had
become one of the most widely used prime movers for new power generation applications [1]. It is known that to
optimize its efficiency, the gas turbine must operate at high rotational speed (5000 to 6000 rounds / min in a power
range from 10 to 40 MW) [2][3]. For this power range, a mechanical gearbox is inserted between the turbine and the
generator in order to inject electrical power into the grid (50 or 60 Hz). However, in spite of the convincing
economic reasons, the mechanical gear boxes present major disadvantages; Lubrication oil increases the fire risk, it
requires more protection, heavy weight and unacceptable noise level [4][5]. And due to direct network connection,
the generator speed cannot be variable: it is imposed by the network and is constant. See Fig.1. Its elimination is
then wished.

Start-up gas turbine gear box Synchronous transformer Network

motor (High speed) generator (50 or 60Hz)
(50 or 60Hz)

Fig. 1. Schematic of the electrical power generation system using gear box.

To overcome this problem, many authors propose to replace the mechanical gearbox by a flexible electronic
solution which offers the ability to operate with very high power, increases turbine efficiency by using variable
speed [6][7]. Figure 2 proposes one schematic of the electrical power generation system using frequency converter.

gas turbine Syncronous precharge Network

(High speed) generator Frequency converter (50 or 60Hz)
(High speed)

Start-up direction

Energy production direction

Fig. 2. Schematic of the electrical power generation system using frequency converter.

As shown in this figure the static converter can be used also to start the synchronous motor and the gas turbine
[2]. Electronic solution presents then reduction in equipment and noise [8].
This paper is organized as follows: Section 2 proposes some related works; Section 3 presents multi-level NPC
inverter in Square Wave Modulation with selective harmonic cancellation; Multi-Level NPC Inverter Required for
High Power Application is proposed in Section 4. Section 5 shows some verification results. Finally, Section 6
concludes the paper.
 Aziza Benaboud and Alfred Rufer / Procedia Engineering 138 (2016) 337 – 346 339

Nomenclature

SWM Square Wave Modulation

NPC Neutral Point Clamped
3L3PNPC inverter Three Level Three Phase Neutral Point Clamped inverter
THD Total Harmonic Distortion
PWM Pulse Wave Modulation
δn ,δn1,δn2 The switching angles
ʋ
U The output-voltage of a three-level converter

2. Related works

The idea to replace the mechanical reducers by static converters existed for a long time because of the
availability of the power electronic components and great flexibility of speed. There are two main groups of
frequency converters: The direct converters, that transform a given input electric power at frequency f g directly into
another power at frequency fn. Typical examples are the cyclo-converter as well as the matrix converter [9][10]
[11][12]. Problems affect one or both types of direct converter: extra forced commutation at synchronous speed and
poor quality converter currents [13][14]. Moreover, cyclo-converters suffer from a limitation of the maximum
output frequency to values smaller than half the mains frequency. The indirect converters are characterized by a
double frequency conversion AC to DC transformation followed by the DC to AC one. The intermediary circuit
brings the advantage of decoupling of harmonics currents as well as of decoupling of reactive power between the
generator and the network. This type of converter is not common in high power application, because of very high
switching losses due to the high frequency of the PWM technique normally used for its control. A new control
strategy is proposed for the multi-level converter, characterized by its high efficiency due to the use of square-wave
modulation. The main advantage of this mode is the quasi absence of switching losses. In this mode, only the
frequency can be varied between the input and the output voltage, but their magnitudes are not freely controllable. A
voltage magnitude adaptation can be done by the generator's excitation. The produced active and reactive power can
be controlled by the generator excitation as well as both the angle shift between the generator and rectifier voltages
and between the inverter and network voltages. In this paper the author proposes to replace the mechanical gearbox
by a flexible electronic solution which offers a high efficiency. Due to its significant advantages in high-power,
medium and high voltage applications, the multi-level Neutral Point Clamped (NPC) converter has received special
attention [15][16]. With the NPC topology, multi-level converters allow the output voltage to spread on several
levels and thus increase it above the voltage limits of classical semiconductors and also reduce its harmonic
distortion.

3. Multi-Level NPC Inverter in SWM, Selective Harmonic Cancellation

3.1. Three Level NPC Inverter in SWM

A three level Neutral Point Clamped inverter is widely used for medium voltage applications. It allows the output
voltage to spread on three levels and thus to increase it above the voltage limits of classical semiconductors.
Figure (3.a) shows the main circuit of the NPC inverter, where for clarity only one phase is shown. In this
configuration, the DC side will be connected to the capacitor midpoint N (neutral point) via diodes to provide the
third level in the output waveform. Figure (3.b) shows an output-voltage’s waveform of a three-level converter
operated in square-wave mode. In this figure, δn represents a selected switching angle, which depends on the
harmonics component according to (1) [17].

υ Û 2
U cos(υ G n ) (1)
υπ d
340 Aziza Benaboud and Alfred Rufer / Procedia Engineering 138 (2016) 337 – 346

a b

Fig.3. (a) Phase circuit diagram of a 3 level NPC inverter; (b) output voltage’s waveform of 3Level converter.

Two important criteria must be taken into consideration when selecting the switching angle δn: fundamental
voltage evolution and harmonic distortion (THD) of the output voltage. To illustrate this point, Figure (4.a) presents
in function of the switching angle δn, the ratio, given by (2), between the fundamental wave 1Û and the DC voltage
Ud. In Figure (4.b) the THD of the output voltage calculated according to (3) is also presented in function of δn,
1
Û 2
cos G n (2)
Ud S
f § cos( XG · 2
100
THD ¦¨ n
¸ % (3)
cos G n X 5© X ¹

a b

Fig.4 (a) Ratio between fundamental wave and DC voltage; (b) The output voltage’s THD in function of switching angle δn.

The curve in Figure (4.a) shows that the ratio in (2) hardly changes, as a function of δn, in the range between 0°
and 20°. In this range the output voltage’s THD presents a minimum value, exactly when δn =15.58°. The switching
angle δn can also be optimized to eliminate some harmonics. For example, the 5th harmonic disappears when δn =18°.
See fig.5. By choosing the switching angle, the minimum of harmonics can be obtained (δn =15.58°֜ TDH = 16%).
The traditional two level converter has more harmonic contents, which corresponds to (δn = 0°эTDH = 31%).
The advantage of this method with SWM is that the lower harmonics are reduced or eliminated without increasing
the higher-order harmonics.
 Aziza Benaboud and Alfred Rufer / Procedia Engineering 138 (2016) 337 – 346 341

Time (seconds) Harmonics order

Fig.5 (a) The output 3LNPC inverter voltage; (b) Output voltage.harmonics analysis.

3.2. Five Level NPC Inverter in SWM

Five level inverter is composed by 8 switches and connected to neutral point via diodes as illustrated in
Fig.(5.a). The output voltage’s wave form is drown in Fig.(5.b).

a b

Fig.6. (a) Phase circuit diagram of a 5 level NPC inverter; (b) The output voltage’s waveform of 5 Level converter.

In this wave form, δn1 and δn2 present switching angles which can be chosen according to (4) in order to
eliminate for example the fifth and the seventh harmonics.

­5 2U d § G · G
° Û cos 5¨ G n1  n 2 ¸ cos 5 n 2 0
° 5S © 2 ¹ 2
® (4)
°7 Û 2U d § G · G
cos 7¨ G n1  n 2 ¸ cos 7 n 2 0
° 7S © ¹
¯ 2 2

The fifth and the seventh harmonics are eliminated when δn1=5.14° and δn2=25.71°, see figure (7.b), the output
voltage’s THD is 11%. It is also possible to optimize the THD by calculation according to (5). For example for
δn1=6.4° and δn2=23.68° the optimum of the output voltage THD is 9.35%.
342 Aziza Benaboud and Alfred Rufer / Procedia Engineering 138 (2016) 337 – 346

2
100% §1 § G · G ·
THD ¦ ¨¨ cos X¨ G n1  n 2 ¸ cos X n 2 ¸¸
§ G · G ©X © 2 ¹ 2 ¹ (5)
cos¨ G n1  n 2 ¸ cos n 2
© 2 ¹ 2

a b

Time (seconds) Harmonics order

Fig.7 (a) The output 5LNPC inverter voltage; (b) Output voltage harmonics analysis.

4. Multi-Level NPC Inverter Required for High Power Application

For high power applications, for example: (73MVA: (9.4kV, 5.23kA)), high voltage IGBTs are needed. To
achieve the required output voltages of (9.4kV), which means DC-link voltages above 15 kV, the use of three or five
level technology with the available 2.5 kV and 2 kA devices, is not sufficient, so an internal series connection is
necessary (see Figures (8.a) and (9.a)).

4.1. Three Level NPC Inverter with internal series connection

a b

Fig.8. (a) 3L-NPC inverter, with internal series connections. (b) Three blocs of 3L-3P-NPC inverter operating in parallel.
 Aziza Benaboud and Alfred Rufer / Procedia Engineering 138 (2016) 337 – 346 343

4.2. Five Level NPC Inverter with internal series connection

a b

Fig.7. (a)

Fig.9. (a) 5L-NPC inverter, with internal series connections. (b) Three blocs of 5L-3P-NPC inverter operating in parallel.

In addition, to fulfill the demands of high power applications, it is necessary to use several leg modules in parallel
(see Figures (8.b) and (9.b)). This solution allows the output voltage and current to increase above the
semiconductors limits, so that it can be used for high power applications [18]. However, the series connection of
power semiconductors can have negative side effects, such as power semiconductors voltage equilibrium. To ensure
the voltage stabilization between the power semiconductors, an appropriate converter control method is needed
[19][20].

5. Simulation and Experimental Results

5.1. Simulation results

In order to have an estimation of the converter efficiency, the model in Figure 2 has been simulated using the
following characteristics: Vn = 9.4kV is the network voltage. In = 5.23kA is the line current. Lni = 1.71mH is the line
inductance. Sn = 73.8MVA is the apparent power. Ud = 15kV is the DC voltage at the input of the inverter. The
semiconductors used in this model are IGBT (5SNR20H2500), and they are driven by using a square wave
modulation. Figures 10 and 11 plot our results. For three and five level NPC inverters, the active and reactive
powers are represented respectively in fig (10.a) and (11.a), together with total losses calculated using the model in
Figures 8 and 9 and shown in fig (10.b) and (11.b).
In comparison with a maximal power, the total losses in those converters represent only 0.21%, which means
approximatively 99.97% efficiency. It is important to notice that for a complete back to back frequency converter
(see fig.2), the produced power is affected twice by this efficiency value, i.e. 99.5%. This high efficiency is achieved
by the use of Square Wave Modulation which yields a quasi-absence of switching losses.
344 Aziza Benaboud and Alfred Rufer / Procedia Engineering 138 (2016) 337 – 346

a b
x107 x105

Time (seconds) Time (seconds)

Fig.10.(a) Active and reactive power, (b) converter losses, in 3L3PNPC inverter

a b

Fig.11.(a) Active and reactive power, (b) converter losses, in 5L3PNPC inverter

5.2. Experimental results

The same system as shown in Fig. 2 has been built using three level three phase neutral point clamped inverter
and tested by a low voltage laboratory prototype. To replace the gas turbine a Direct Current Machine is used as
power source, it is directly connected to the synchronous generator, which is connected to the grid through an
indirect static frequency converter with an intermediary DC circuit.

Fig.12. Laboratory prototype using 3L3PNPC inverter.

 Aziza Benaboud and Alfred Rufer / Procedia Engineering 138 (2016) 337 – 346 345

Active and reactive powers are measured and presented in figures (13) and (14) together with the alternative line
current, inverter and network voltages. When the reactive power is near zero, the alternative line current is in phase
with the network voltage. The reactive power can be compensated rapidly and without DC component in the
alternative line current.

Fig.13. Measured results for cosφ = 1

Fig.14. Measured results for cos φ ≠ 1

346 Aziza Benaboud and Alfred Rufer / Procedia Engineering 138 (2016) 337 – 346

6. Conclusion

In this paper, power electronic solution is proposed to improve the energetic efficiency of a thermal plant and
to optimize energy production. Indeed, Electric power generators using gas turbines as power sources are connected
to the turbines through a mechanical gearbox, in order to adapt their synchronous speed to the optimal rotation speed
of the turbine, which is very often much higher than the synchronous speed. However, due to direct network
connection, the generator speed cannot be variable: it is imposed by the network and constant. The proposed
frequency converter replaces the mechanical gearbox and offers a high efficiency due to the use of SWM for its
control; SWM presents a quasi-absence of switching losses. Using this approach, the turbine is directly connected to
the synchronous generator, which is connected to the grid through an indirect static frequency converter. The
capacitive intermediary circuit brings the advantage of decoupling of harmonics between the generator and the
network currents. A control method is also proposed to resolve some problems incurred by using SWM, in particular
to reduce the harmonics distortion of the output inverter voltage and current, simply by choosing well-defined
switching angles. As a second important contribution, this paper proposes two types of Multilevel NPC inverters,
three and five level with internal series connection. These multilevel converters are required for high power
applications. Simulation and experimental results for different operating points highlight the capabilities of the
proposed power electronic solution. These include the high energetic efficiency of a thermal plant, the ability to
operate with unity power factor and better current quality with fewer harmonic.

References

[1] P. Breeze, Power Generation Technologies, book, Elsevier, 2005

[2] F. Peneder, R. Lubasch, and A. Voumard, Dispositif statique de démarrage pour centrales de pompage, compensateurs et turbogroupes à gaz,
Rev.Brown Boveri 440-447, 9/10-1974.
[3] P. Chambadal, La turbine à gaz. 1976.
[4] H. Polinder, Comparison of direct-drive and geared generator concepts for wind turbines, IEEE Transaction on Energy Conversion, vol 21N°3
Sept. 2006 pp. 725-733.
[5] D. Mc Millan, Techno-Economic Comparison of Operational Aspects for Direct Drive and Gearbox-Driven Wind Turbines, IEEE
Transaction on Energy Conversion, vol 25N°1 March 2010 pp. 191-198.
[6] H. Stemmler, High-power industrial drives, proceeding of the IEEE, vol.82, Issue 8, 1994, pp 1266-1286.
[7] H. Stemmler, Variable speed, 3 phase drives fed by static frequency changers, state of the art and applications, Brown Boveri, Revue vol 67,
issue 1 pp. 55-63, 1980.
[8] A.Benaboud, A. Rufer, Power electronic solution for high energetic efficiency of a thermo plant, world Academy of science, engineering and
technology (waset), vol. 73,n. 1,Junuary 2013, pp. 511-518.
[9] S. Turri, Etude d’un alternateur desynchronize, Thèse de Doctorat, U.F.R. des Sciences Techniques et Gestion de l’industrie de l’Université
des Franche-Comte, Belfort, 03.11.2000.
[10] M. Visintin, The active generator principle for high power electric supply: properties and limitations, these EPFL, no. 3324, 2005.
[11] C. Schauder, “Theory and design of 30-hp matrix converter”, IEEE Transactions on industry Applications, vol.28 no 3, pp 546-551, 1992.
[12] A. Schuster, “A matrix converter without reactive clamp elements for an induction motor drive system” PESC 98 Rec, 29th Annual IEEE
17-22 May 1998, vol.1 pages: 714 – 720.
[13] O. Carlson and J. Hylander, “Variable speed AC-generators in wind energy converters”, MELECON’89 Miditerranean, pp. (91-94), 11- 13
April 1989.
[14] M. Dahlgren, “Production à grande échelle d’électricité d’origine éolienne », Revue ABB, pp. 31-37, Mars 2000.
[15] R. Baker, Bridge converter circuit, United States Patent N. 4 270 163, May 1981.
[16] A. Nabae, I. Takahashi, and H. Akagi, A new neutral-point-clamped PWM inverter, IEEE Transactions on Industry Applications, vol. IA -
17, pp. 518 – 523, September/October 1981.
[17] H. Bühler, Convertisseurs statiques. Lausanne: presses Polytechniques et Universitaires Romandes, 1991.
[18] A. Benaboud, convertisseur de fréquence indirect à rapport de tension fixe: interface entre turbo-alternateurs à haute vitesse et réseau
Electrique, Thèse n3733 EPFL, Lausanne 2007.
[19] S. Mariethos, A. Rufer, Multisource DC-DC converter for the supply of hybrid multilevel converter, IEEE Industry Applications
Conference- 41th IAS Annual Meeting, Tampa, florida, USA, 8-12 October 2006.
[20] A. Elserougi, HVDC shunt tap based on series-input parallel-output DC-AC multi-module VSCs with DC voltages balancing, Power
Electronics, Machines and Drives (PEMD 2014), 7th IET International Conference, Manchester, 8-10 April 2014