Benefits and weak points of various control strategies in enhancing variable speed wind turbine transient performance

Renewable Energy Focus  Volume 19–20



2017

www.renewableenergyfocus.com

ORIGINAL RESEARCH ARTICLE

Benefits and weak points of various control strategies in enhancing variable speed wind turbine transient performance Kenneth E. Okedu Department of Electrical and Electronic Engineering, Kitami Institute of Technology, Hokkaido, Japan

It is important to keep grid connected wind farms in operation during disturbances in the power network. This paper investigates various control techniques of wind farm composed of Doubly Fed Induction Generator (DFIG) and fixed speed induction generator. Several enhancement schemes were considered with the aim of stabilizing the mixed wind farm. Some schemes were used to enhance the performance of other schemes. The schemes considered were subjected to the same transient condition at the rated wind speed of the wind turbines. The simulation results in Power System Computer Aided Design and Electromagnetic Transient Including DC (PSCAD/EMTDC) platform show the improved performance of the wind turbines and the wind farm for the various cases. In addition, a comparison of the benefits and weak points in terms of economic costs and performance of the various approaches to enhance the stability of the variable speed wind turbines were analyzed. 1. Introduction The demand for wind power is on the rise daily, thus, it is imperative to take into account the stability of wind farms in order to ensure safe and good power delivering. According to the recent grid requirements, wind farms must attend to transient disturbances as quickly as possible. Various Flexible AC Transmission Systems (FACTS) have been proposed in the literature to enhance wind turbine stability, ranging from the use of Static Var Compensator (SVC), Static and Unified Power Flow Controller (UPFC), Dynamic Voltage Restorer (DVC), Solid State Transfer Switch (SSTS), Energy Capacitor System (ECS), and SMES [1–6]. A comparative study of the transient stability improvement of fixed speed based grid connected wind farm with the help of SVC and a STATCOM was presented in [7], while the improvement of power quality considering voltage stability in grid connected system by FACTS devices was discussed in [8]. The transient stability improvement with electrical braking and reactive compensation of large scale wind power generation was carried out in [9], where it was concluded that the combination of braking resistors and reactive power compensation could significantly improve the stability of wind turbines. E-mail address: [email protected].

A comparative study using FACTS devices and variable speed drive to stabilize wind farms was presented in [10], where it was established that the variable speed drive is more favourable because no external reactive power compensation is required. In [11], a comparative stability analysis of DFIG-based wind farms and conventional synchronous generators was analyzed, with the conclusion that the oscillatory behaviour of the synchronous generators was improved by the DFIG system when connected. The tunning method for parameter optimization to achieve effective stability and tune damping controller considering eigenvalues were carried out for the DFIG system in [12,13]. The effects of the increase stability margins and penetration effects for including a DFIG based system at point of coupling connection at transmission lines were presented in [14–17] respectively. According to these papers, the voltage control capabilities of the wind farm based on DFIG, improves the voltage stability margin at distribution and transmission stages. In [18,19], the study of using trajectories eigenvalues of the DFIG were calculated and a four generator system was used to know the effect of replacing synchronous generators by wind turbines, with the aim of improving the stability of the wind farms when they are grid connected. 1755-0084/ß 2017 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.ref.2017.06.001

104



2017

Since DFIGs are becoming more popular, the idea is shifting from using external reactive compensation devices for wind farm stability [20,21] to the use of the DFIG systems. Although, an alternative use of variable speed wind turbine for wind farm stability is the implementation of the Permanent Magnet Synchronous Generator (PMSG) [22]. The major drawback of the PMSG system is the full power converter rating as compared to DFIG that has only 20–30% power converter rating. Based on this fact, many Fault Ride Through (FRT) methods on DFIG system have been proposed. In [23,24], the use of a crowbar switch was proposed to improve the FRT of DFIG system during transient, while in [25,26], a DC-chopper voltage was implemented. The combination of the crowbar system and the DC-chopper system was investigated in [27], while a study comparing the operation of the DC-link chopper with the crowbar protection mechanisms was presented in [28]. The effective use of a Static Series Compensator (SSC) and a Dynamic Voltage Restorer (DVR) was reported in references [29–31] and the use of STATCOM device to boost and support DFIG was presented in [32]. The use of a series dynamic braking resistor (SDBR) to boost the FRT of large wind farms made up of only induction generators was reported in [33]. The connection of SDBR switch to the rotor side converter of the DFIG and the stator side were presented in [34] and [35] respectively. Superconducting Fault Current Limiter (SFCL) [36], Passive Resistance Network (PRN) [37], and connections involving series antiparellel thyristors [38] to the stator side of a variable speed DFIG system, already exist in the literature. A study of low voltage ride through for DFIG considering rotor current dynamics was analyzed in [39]. In this paper, the rotor transient analysis and differential equations were derived in a bid to improve the ability of the DFIG during grid fault. The use of hybrid current controllers to enhance DFIG reactive power and capability to withstand low and high voltage ride through was reported in [40]. In [41–44], the authors discussed the problems and possible solutions to improve the performance of DFIG during transient by proposing active voltage control and fast coordinated control strategy based on the characteristics of the DFIG. Enhanced field oriented control technique (EFOC) [45,46] was proposed for the rotor side converter of the DFIG system to facilitate the power flow transfer and also enhance the dynamic and transient stability of the wind generator. The analysis of asymmetric faults considering stability [47,48] of the DFIG system already exists in the literature. Also, analyses of DFIG connected wind farms and a review of the impact of the DFIG wind farm on transient stability of power systems were reported in [49] and [50] respectively. During grid faults the DFIG experiences overcurrents which lead also to increasing DC-voltage on the converter side. The behaviour of the DFIG during grid fault is described as follows. If there is no protection system, the DFIG can suffer from large transient currents in the stator during a grid fault since its stator circuit is directly connected to the grid. Because of the magnetic coupling between the stator and the rotor, large currents and high voltages appear also in the rotor circuit. Furthermore, the surge following the fault includes a rush of power from the rotor terminals towards the rotor side converter. Therefore, there can be a possibility that the desired rotor voltage cannot be maintained and thus the rotor currents cannot be controlled. This means that the rotor side converter can reach to its operating limit and as a consequence

ORIGINAL RESEARCH ARTICLE

it may lose the independent control of real and reactive powers during the grid fault. On the other hand, as the grid voltage drops in the fault moment, the grid side converter is not able to transfer the power supplied from the Rotor Side Converter (RSC) to the grid, and therefore, the excess energy is stored in the DClink capacitor, resulting in rapid increase of the DC-bus voltage [51–53]. It is therefore necessary to protect the power converters against over-currents and the DC-link against overvoltage. This paper focuses on the Fault Ride Through (FRT) methods of a wind farm composed of DFIG and Induction Generator (IG). Eight scenarios were considered in this study. In the first case, the DFIG based wind turbine is operated using a Voltage Controlled Voltage Source Converter (VC-VSC) system including a DC-link chopper protection scheme. The second case considers a crowbar active switch having different crowbar resistances with the DFIG VC-VSC system. A STATCOM was considered to further enhance the DFIG based system during grid fault in case 3, since the crowbar switch in case 2 would disconnect the RSC of the DFIG thereby loosing controllability of active and reactive power control. In case 4, an SDBR is connected at the rotor side of the VC-VSC DFIG based wind turbine. The performance of the SDBR system was also investigated in case 5, when it was connected to the stator of the DFIG and the grid side converter instead of the rotor side of the wind generator. A proper sizing, best time of insertion and duration of operation of the SDBR was considered in this case. Case 6 considers when the DFIG-based wind turbine converter system is operated using a Current Controlled Voltage Source Converter (CC-VSC) system with no protection mechanism circuitry. In case 7, a hybrid approach of controlling the variable speed wind generator power converters using VC-VSC and CC-VSC were considered respectively. Case 8 uses a parallel interleaved configuration of the DFIG power converters. Simulations were run in PSCAD/EMTDC [54] environment. Furthermore, a comparison was carried out with regards to some of the benefits and weak points of the schemes used in this study.

2. Wind turbine modelling The parameters of the wind turbines used in this study are given in Table 1. Basically the primary components required in the modelling of a wind turbine system are the turbine rotor or prime mover, a shaft and a gearbox unit. The wind generators ratings are 2 MW for the DFIG and IG respectively. The aerodynamic torque and the mechanical power of a wind turbine are given by [26,35]: T M ¼ 0:5rCt ðlÞpR3 Vw2 ½NM

(1)

Pwt ¼ 0:5rCp ðl; bÞpR2 Vw3 ½W

(2)

TABLE 1

Generator parameters. Generator type

IG

DFIG

Rated voltage Stator resistance Stator leakage reactance Magnetizing reactance Rotor resistance Rotor leakage reactance Inertia constant

690 V 0.01 pu 0.07 pu 4.1 pu 0.007 pu 0.07 pu 1.5 s

690 V 0.01 pu 0.15 pu 3.5 pu 0.01 pu 0.15 pu 1.5 s 105

ORIGINAL RESEARCH ARTICLE

Renewable Energy Focus  Volume 19–20

ORIGINAL RESEARCH ARTICLE

Renewable Energy Focus  Volume 19–20

where r is the air density, R is the radius of the turbine, Vw is the wind speed, Cp(l, b) is the power coefficient. The wind turbine characteristics with the rotor speed limits are shown in Appendix (Figure A1).

ORIGINAL RESEARCH ARTICLE

The model of this study with the detailed structure of the various fault ride through methods for the DFIG variable speed wind turbine connected to the fixed speed IG system is given in Figure 1. In Figure 1, the various FRT schemes considered to improve the wind turbine response and the entire wind farm are summarized as follows: Case 1: The use of VC-VSC and the DC chopper protection scheme. Case 2: The use of VC-VSC with an active Crowbar Switch at the RSC of the DFIG VSWT Case 3: The use of VC-VSC with STATCOM FACTS reactive power compensation Case 4: The use of VC-VSC with SDBR at the rotor side of the DFIG system Case 5: The use of VC-VSC with SDBR at the stator and the grid side of the DFIG system Case 6: The use of a CC-VSC with no DC-link or chopper circuitry Case 7: The use of a combined VC-VSC for RSC and CC-VSC for GSC approach for the DFIG. Case 8: The use of parallel interleaved voltage source power converters for the DFIG. Apart from assessing the performances of the various schemes, another aim is to know how much impact each scheme has on the connected fixed speed wind turbine, as not all the schemes would

V dcstat STATCOM

Gearbox Blade

DFIG Ps,Qs

Wind Energy

Case 5

Pg,Qg

SDBR Case 4

SDBR Crowbar Switch Case 2

Pitch Angle Control

AC

MPPT Control

IG

j0.1 Qc= j0.39

0.0005+j0.005

0.0005+j0.005

Vdc

GSC

VC-VSC Case 1

Ps, Qs Control

Case 8 Interleaved Parallel Converters

j0.1

Case 6

CC-VSC

RSC

P

LQ L

AC

Vdc, PL, QL Control

Vdc-Ref

Case 7

VC-VSC & CC-VSC

have impact on the fixed speed wind turbine during transient. Although, all the schemes would have impact on the DFIG variable speed drive since they are directly connected to its circuitry.

The control strategies for the various FRT schemes considered in this study are given as follows.

4.1. Case 1 The schematic representation of power flow breakdown in a classical DFIG wind turbine configuration that uses bidirectional power converters is shown in Figure 2a [55,56]. One of the salient features of the DFIG is the economical utilization of the power electronic voltage source converters only in the rotor circuitry as compared to the synchronous based wind generators that operates using fully rated converters. From, Figure 2a, optimum active power is obtained during normal operation based on the wind power conditions by the setting of the d-axis reference current in relation to the maximum power point tracking characteristics. The pitch angle control of the DFIG could also help in the active power regulation by providing power margins which allows frequency control and grid support. The rotor d-axis current could be given preference in the control system to generate active power in the Grid Side Converter (GSC) which indirectly controls the DC-link voltage by maintaining it at 1.0 pu. The rotor and grid side converters can contribute to the reactive power dissipation by regulating the grid voltage as shown in Figure 2a. The DC-link chopper circuitry is shown in Figure 2b.

4.2. Case 2 The coordinated control of the crowbar system and the RSC is shown in Figure 3a. The crowbar system shown in the figure is used to assess the overall impact of protection of the DFIG during transient conditions. The crowbar is made up of one resistance fed through a three phase thyristor system (T1 to T6) which is connected to the rotor side through a controllable three phase switch. The switch is normally open and is closed in order to short circuit the rotor, if the DC-link voltage exceeds the set maximum voltage of 150% pu during transient. A comparator is used to compare the signals as shown in the figure. If Vdcmax is higher than Vdc, the comparator gives an ON state sign of 1 and thyristors (T1 and T6) are triggered, hence disconnecting the RSC and current flows through the resistance Rcr connected in the circuit. The value of the crowbar resistance has a great effect on the DFIG system during transient, as will be shown in the simulation results. Hence, in this work, proper crowbar resistance sizing was taken into consideration during the grid fault analysis.

4.3. Case 3 0.1+j0.6

Infinite bus

j0.2 0.1+j0.6

Fault

Self-capacity base (6.6kV) System base (50Hz, 66KV, 100MVA)

FIGURE 1

Model system of study with the various enhancement schemes considered. 106

2017

4. The control strategies for the various enhancement schemes considered

3. The model system of study and various enhancement schemes for DFIG wind generator

Case 3



The output current of the STATCOM in Figure 3b(i) is adjusted to control the nodal voltage at the point of common coupling by injecting reactive power at the bus. The output of the voltage source converter can be expressed as 8 V V > < P ¼ 1 2 sind X (3) 2 > : Q ¼ V 1 V 2 cosd V1 X X



2017

ORIGINAL RESEARCH ARTICLE

ORIGINAL RESEARCH ARTICLE

Renewable Energy Focus  Volume 19–20

FIGURE 2

(a) DFIG control strategy. (b) DC-link chopper scheme (case 1).

where P and Q are the active and reactive power of the STATCOM voltage source converter, and V1 and V2 are the bus voltages respectively. X is the reactance of the coupling transformer and d is the phase difference between voltage V1 and V2. The d-axis current is used to control the q-axis reference voltage, while the qaxis current is used to regulate the d-axis reference voltage. The control system is effectively regulated via Proportional Integral tunings. The control strategy of the STATCOM is shown in Figure 3b(ii).

4.4. Cases 4 and 5 In Figure 3c(i), SDBR is connected to the rotor side (A), the grid side converter (B) and the stator side (C) of the DFIG system respectively. The best SDBR position was determined based on a severe 3 LG fault applied to the system as shown so that the effects of the duration and switching strategy of the SDBR could be ascertained. The control strategy and switching signals of the SDBR system are shown also in Figure 3c(ii) in the wind generator stator side. Figure 3c(iii) shows the transfer of power across the DFIG wind 107

ORIGINAL RESEARCH ARTICLE

Renewable Energy Focus  Volume 19–20

ORIGINAL RESEARCH ARTICLE FIGURE 3

(a) Control strategy of crowbar scheme (case 2). (b) STATCOM scheme (case 3). (c) Series Dynamic Braking Resistor Scheme (cases 4 and 5).

108



2017

Renewable Energy Focus  Volume 19–20



2017

ORIGINAL RESEARCH ARTICLE

TABLE 2

Switch-in time (ms)

Switch-out time (ms)

Duration of operation (ms)

120 150 180

200 250 300

80 100 120 ORIGINAL RESEARCH ARTICLE

SDBR switching times.

generator system, with extra dynamic power possessed in its drive train system. The three switching signals for the SDBR that were investigated based on set values are the DC-link voltage, Edc, Rotor current Ir, and the grid voltage Vg. Table 2 gives the switching times of the SDBR considering a 100 ms fault. The best insertion time and optimum duration of operation of the SDBR during transient was investigated also.

4.5. Case 6 A CC VSC structure shown in Figure 4a is used to enhance the DFIG wind generator during transient instead of the conventional voltage source converter topology. A comparator based on the grid voltage signal is used to determine the reference signal of the active and reactive power control during normal and fault conditions of the DFIG wind turbine system. During normal operation, the grid voltage is above the set value of 0.9 pu, and PMPPT of the wind turbine real power. The reactive power of the wind turbine generator is maintained at 0 pu with the help of the controllers. During periods of grid disturbances, that is, in the scenarios were the grid voltage is below 0.9 pu, the comparator regulates the signal by switching the reference signals of the DFIG real and reactive power to zero. The Phase Lock Loop (PLL) system helps in the dq to abc conversion, consequently, *Irabc switching signals are compared to the measured rotor currents Ir(abc) signals. It could be observed that only one Proportional Integral (PI) system is obtainable in this strategy, hence dq to abc transformation is done once instead of twice for the conventional voltage source converter system, consequently, there is reduced controller design complexities.

4.6. Case 7 Figure 4b shows the combined approach of the VC-VSC for the RSC and CC-VSC for the GSC of the DFIG variable speed wind turbine system respectively. This hybrid converter topology is compared to the other converter topologies earlier discussed.

4.7. Case 8 Figure 4c shows the parallel interleaved IGBT converter for the DFIG wind turbine system. The most suitable scheme for the parallel interleaved VSCs is the active zero state PWM (AZSPWM) scheme that uses two adjacent active voltage vector and two near opposing active vectors. In the two level converter system presented in earlier cases, only eight voltage vectors that includes two null is been able to be generated. Although, the complex switching circuitry of the 2 level interleaved VSCs system allows much possibilities for appropriate voltage vector selection to satisfy the commutation condition.

5. Analysis of simulation results Some of the presented simulation results are given in Figs. 5–8 for the most affected wind turbine variables in the variable and fixed

FIGURE 4

(a) Control system for DFIG using CC-VSC topology (case 6). (b) Combine approach of VC-VSC and CC-VSC topology (case 7). (c) 2-level parallel interleave IGBT inverter scheme.

speed wind generators for the various cases considering a severe three line to ground fault (Figure 1). During the transient analysis, the wind speed is constant because the grid fault is assumed to occur for a very short time. Figure 5a shows the response of the DFIG DC-link voltage for cases 1 and 2 respectively. The fault in 109

ORIGINAL RESEARCH ARTICLE

Renewable Energy Focus  Volume 19–20

(b)

DC link voltage (pu)

(a) 2.0

150

Rotor current[A]

1.0

ORIGINAL RESEARCH ARTICLE

0.5

100 50 0 -50 -100

0.0

-150

0.4

0.8 1.2 Time[sec]

-200

2.0

5

1.2

(d)

1.0 0.8 0.6 0.4 0.2

6

7

8

1.0 0.8 0.6 0.4 0.2 0.0 10

11

With STATCOM Without STATCOM

1.0

0.8

12

13

14

15

Time[sec]

Terminal Voltage of DFIG[pu]

(f) 1.2

12

13

14

15

Time[sec]

(e)

11

8

With STATCOM Without STATCOM

Time[sec]

0.6 10

7

1.2

0.0 5

6

Time[sec]

Crowbar resistance 0. 2 Crowbar resistance 0. 8 Crowbar resistance 1. 0

(c)

Terminal Voltage[pu]

1.6

Terminal Voltage of DFIG[pu]

0.0

Rotor Speed of IGs[pu]

2017

Crowbar resistance 0. 2 Crowbar resistance 0. 8 Crowbar resistance 1. 0

200

No Protection With Crowbar With DC Chopper

1.5



1.4 With crowbar only With crowbar and FACTS

1.2 1.0 0.8 0.6 0.4 0.2 0.0 0.0

0.2

0.4

0.6

0.8

1.0

Time[sec]

FIGURE 5

(a) DC-link voltage response (cases 1 and 2). (b) Rotor current of DFIG (case 2). (c) Terminal voltage of DFIG (case 2). (d) Terminal voltage of DFIG (cases 1 and 3). (e) Rotor speed of IG (cases 1 and 3). (f ) Terminal voltage of DFIG (cases 2 and 3).

this case occurred at 0.1secs and a case where no protection was implemented is additionally shown in the figure. For no protection scenario, the DC-link voltage nominal value is almost twice and this is not healthy for the wind generator converters. However, in case 1, the DC-link voltage was mitigated within acceptable range by the DC-Chopper scheme. Also, the DC-link was protected by the crowbar switch considering case 2, although apart from the expensive nature of this scheme, over shoots also occur when disconnecting the rotor side converter of the DFIG system. Thus, the cheaper and cost effective case 1 scheme is preferred in this 110

situation. This is because with the use of the DC chopper control in Figure 2b, the unbalance of electrical power injection into the power system by the GSC and the electrical power injected into the DC-link by the RSC is dissipated. In Figure 5b and c, the fault is said to occur at 5.1secs to illustrate the response of the crowbar system for case 2 as shown for the rotor current and the DFIG terminal voltage respectively. The value of the crowbar resistance shown in Figure 3a is very important because it determines how much reactive power the DFIG will draw while the crowbar is inserted during grid fault. Hence, as

Renewable Energy Focus  Volume 19–20



2017

ORIGINAL RESEARCH ARTICLE

Active Power of DFIG[pu]

1.2

1.0 0.8 Rotor current switching DC-link voltage switching Grid voltage switching

0.6 0.4 0.2 0.0

0.2

0.4 0.6 Time [sec]

0.8

1.0 0.8

Rotor current switching DC-link voltage switching Grid voltage switching

0.6 0.4 0.2 0.0 0.0

1.0

0.2

0.8 0.6

SDBR position A (rotor) SDBR position B (grid side converter) SDBR position C (stator)

0.4 0.2 0.0 0.2

0.4

0.6

0.8

1.0

1.3 1.2 1.1

1.0

0

1

2

3

4

5

Time [sec ]

Time[sec]

(e)

(f) 1.2 1.0 0.8 0.6 0.4 0.2 0.0 -0.2 -0.4 0.0

1.33 Rotor Speed of DFIG[pu]

Active Power of DFIG[pu]

0.8

1.4

1.0

0.0

0.6

SDBR position A (rotor) SDBR position B (grid side converter) SDBR position C (stator)

1.5

1.0

0.4

Tim e[sec] (d)

1.2

ORIGINAL RESEARCH ARTICLE

Terminal Voltage of DFIG[pu]

(c)

(b) 1.2

Rotor speed of DFIG[pu]

DC-link voltage of DFIG[pu]

(a)

SDB R 0.05pu SDB R 0.10pu SDB R 0.15pu No SDB R,

0.5

1.0 Time[sec]

1.5

2.0

20ms Insertion time after fault, 80ms duration 50ms Insertion time after fault, 100ms duration 80ms Insertion time after fault, 120ms duration

1.32 1.31 1.30 1.29 1.28

0

1

2 3 Time[sec]

4

5

FIGURE 6

(a) Terminal voltage of DFIG (cases 4 and 5). (b) Rotor speed of DFIG (cases 4 and 5). (c) Rotor speed of DFIG (cases 1 and 5). (d) Reactive power of DFIG (cases 1 and 5). (e) Active power of DFIG (cases 1 and 5). (f ) Rotor speed of DFIG based on timing (case 5).

shown in Figure 5b and c, different values of crowbar resistance were investigated for this case to determine the most efficient in protecting the variable speed wind turbine during transient conditions. When the crowbar resistance is too low (0.2 pu), there is a delay in the settling time of the wind turbine variables, and when the crowbar resistance is too high (1.0 pu), high oscillations were observed in the rotor current of the DFIG, leading to poor performance. However, a proper crowbar resistance value of (0.8 pu) gives a better performance as shown in the figures because, it could effectively damp the oscillations in the rotor circuitry during transient compared to other considered crowbar resistances. If the fault is not eliminated before the critical rotor speed is achieved the rotor may accelerate and loose its stability. The period of time that the wind turbines must remain connected to the network depends on the injection of reactive power during

the fault by the RSC which further contributes to increase the terminal voltage as shown in Figure 5c. Figure 5d and e shows the responses for cases 1 and 3 respectively for the DFIG terminal voltage and rotor speed of the induction generator. In these figures, the grid fault occurred at 10.1 s. Case 1 shows the response of the DFIG terminal voltage and rotor speed of the induction generator fixed speed without the STATCOM external reactive compensation device. In case 3, where the STATCOM control was implemented, the variables of the wind turbine recovered faster during the grid fault because more reactive power was injected by the STATCOM into the grid to support the DFIG power converters. This scenario has a great impact on the connected fixed speed wind turbine as shown in Figure 5e because the rotor speed was not allowed to accelerate during the grid fault, thus, do not lead to loss of synchronism of the fixed speed wind turbine. The 111

ORIGINAL RESEARCH ARTICLE

Renewable Energy Focus  Volume 19–20

(a)



2017

(b) W it h VC-VS C W it h CC-VS C

1 .0 0 .8 0 .6 0 .4 0 .2 0 .0

0

1

2

3

4

With VC-VSC With CC-VSC

1.2

DC-link Voltage of DFIG[pu]

ORIGINAL RESEARCH ARTICLE

Terminal Voltage of DFIG[pu]

1 .2

1.0 0.8 0.6 0.4 0.2 0.0

5

0

1

2

Time [sec]

(c)

(d)

DFIG-1 (with VC-VSC) DFIG-1 (with CC-VSC)

2

4

5

No DFIG Control Using VC-VSC Using CC-V SC

1.2 1.0

Terminal Voltage of Induction Generator[pu]

Rotor Current of DFIG-1[pu]

3

3

Time[sec]

1 0 -1 -2 -3 0.0

0.2

0.4

0.6

0.8

0.8 0.6 0.4 0.2 0.0

1.0

0

1

2

Time[sec]

3

4

5

Time[sec]

FIGURE 7

(a) Terminal voltage of DFIG (cases 1 and 6). (b) DC-link voltage of DFIG (cases 1 and 6). (c) Rotor current of DFIG (cases 1 and 6). (d) Terminal voltage of IG (cases 1 and 6).

(b)

Terminal voltage of variable speed wind turbine [pu]

1.2 Current-control based for RSC and GSC systems

1.0

1.2

0.8 0.6

Voltage-control based for RSC and GSC systems Hybrid of voltage-control based for RSC + current-control based for GSC systems

0.4 0.2 0.0 0.0

0.5

1.0

Fixed speed w ind turbine term inal voltage [pu]

(a)

Current-control based for RSC and GSC systems

1.0 0.8 Hybrid of voltage-control based for RSC + current-control based for GSC systems

0.6 0.4

Voltage-control based for RSC and GSC systems

0.2 0.0

1.5

0.5

Time[sec]

1.5

2.0

(d)

1.0 0.8 0.6

No Converter Interleave only RSC only GSC Both RSC and GSC

0.4 0.2 0.10

0.15

0.20 0.25 Time[sec]

0.30

Dc-link voltage of DFIG[pu]

Dc-link voltage [pu]

(c) 1.2

0.0

1.0 Time[sec]

1.2 1.0 0.8 0.6 No SDBR, No converters Interleave Only SDBR, No converters Interleave SDBR + converter Interleave Only converters Interleave

0.4 0.2 0.0

0.1

0.2

0.3 0.4 Time[sec]

0.5

0.6

FIGURE 8

(a) Terminal voltage of DFIG (cases 1, 6 and 7). (b) Terminal voltage of IG (cases 1, 6 and 7). (c) DC-link voltage of DFIG (case 8). (d) DC-link voltage of DFIG (cases 5 and 8). 112



2017

rotor acceleration and disturbance was reduced by almost 50% (Figure 5e), with a faster settling time as compared to when the STATCOM was not implemented. To further enhance the wind turbine system with the conventional crowbar system, cases 2 and 3 were implemented together as shown in Figure 5f, whereby the STATCOM injects reactive power to the system, when the crowbar disconnects the RSC of the variable speed wind turbine during grid fault. Since the DFIG would behave as a conventional IG during the period of crowbar disconnection, the STATCOM could help to achieve fast recovery and improve stability of the wind turbine system by boosting the reactive power via its converters described in Figure 3b. With the use of the STATCOM controller, the high voltage specks during transient were reduced from almost 140% to 110% with faster recovery of the terminal voltage of the wind turbine (Figure 5f) because of the extra reactive power compensation from the STATCOM power converters. Figure 6a and b shows the situation where the SDBR was switched using grid voltage, DC-link voltage and rotor current signals based on the control strategy of Figure 3c(ii). The grid voltage signal gave better performance of the system. When the SDBR is considered in the rotor, stator and grid side of the DFIG system for cases 4 and 5 respectively, better performance of the wind generator variables were obtained for the stator connected SDBR used in case 5 as shown in Figure 6c and d. This is because the current limitation reduces the charging current to the DC-link capacitor, due to the series SDBR connection topology in Figure 3c at the stator of the DFIG. Hence, DC-link overvoltage which could damage the DFIG power converter is mitigated. Consequently, more analysis for the connection of the SDBR at the stator side of the DFIG system are shown in Figure 6c–f to show the effects of the proper sizing and timing of the SDBR insertion during grid fault for case 5. A situation where SDBR was not used in case 1 is compared with situations where various SDBR values were considered. The use of SDBR as seen from the figures improved the performance of the wind turbine during grid fault. Furthermore, the sizing of the SDBR is very important as shown in Figure 6e for the active power of the DFIG wind turbine. A small value of 0.05 pu SDBR gives a better performance compared to higher values based on the reasons presented in Section 4 for the SDBR control strategy. Considering the best performance of SDBR size of 0.05 pu, the time of insertion of the SDBR was also investigated as shown in Figure 6f, where a shorter timing of 20 ms insertion time and 80 ms duration of operation gives a better performance of the variable speed wind turbine response. Thus, the distinctive merit of connecting SDBR is justified based on the fact that its effect is related to current magnitude rather than voltage magnitude. Also, the generated power is transferred across the wind generator system, while the excess dynamic power is stored in its drive train and heat is dissipated by the SDBR. The analyses for the use of CC-VSC are shown in Figure 7a–d for the wind turbine variables. The terminal voltage of the DFIG and its DC-link voltage gave a better performance as shown in Figure 7a and b respectively. It should be noted that while a DC-link protection scheme was used in case 1, there is no protection scheme for the DC-link in case 6 (Figure 7b) based on the control topology employed in Figure 4a(i) with reduced abc/dq/abc transformation technique. Furthermore, there are many oscillations in the rotor current for case 1 compared to case 6 (Figure 7c). The

ORIGINAL RESEARCH ARTICLE

response of the connected fixed speed wind turbine terminal voltage for this case is shown in Figure 7d. A case when no control was implemented for reactive power compensation leads to voltage collapse, which could lead to immediate shut down of the wind farm based on the grid requirement, while the response using the CC-VSC (case 6) gives a better performance compared to case 1. Figure 8a and b shows the FRT scheme for cases 1, 6 and 7 respectively. Reponses for case 7, where the VC-VSC and CC-VSC DFIG based converter topology was used together gave a better response than case 1, where only VC-VSC was employed. This is because the current based DFIG control gives a faster response as earlier discussed due to its reduced circuitry. However, the responses of the wind turbine variables for case 6 were slightly better compared to case 7. The DFIG variable speed wind turbine response for the single step back to back converter system in case 1, is compared with when only the Grid Side Converter (GSC) is interleaved, only the Rotor Side Converter (RSC) is interleaved, and when both converters were interleaved (case 8). The DC-link voltage in Figure 8c indicates that when both sides of the DFIG wind turbine back to back power converters are interleaved in parallel configuration, more reactive power is achieved in addition to more circulating current. Consequently the DC-link voltage response is enhanced, thus, protecting the power converters more during transient conditions. A combination of cases 5 and 8 in Figure 8d shows that the SDBR can further enhance the DC-link voltage performance because the induced overvoltage caused during transient periods is limited. This will not cause the power converter damage and loss of control, hence controlling the rotor overvoltage and at the same time limiting the high rotor current of the wind turbine. An improved performance of the wind turbine variables were achieved using the proposed control scheme of the parallel interleave converter and SDBR because the Space Vector Modulation (SVM) of the parallel interleave converters result in maximum value of the change in common mode voltage, leading to improved switched output voltage of the voltage source converter leg as shown in the schematic diagram in Figure 4c.

6. Benefits and weak points of the various FRT schemes A summary of some of the benefits and weak points in terms of economic cost and performance of the various protection schemes for the DFIG system are highlighted in Table 3. From the table, some schemes have advantages over the others, though all the schemes are faced with their own shortcomings in protecting the wind turbine. Also, some schemes require the mechanism of other schemes to function properly. Consequently, no scheme has all it takes to effectively enhance the performance of the variable speed wind turbine in terms of economic cost and performance during transient. A combine approach of two or more protection schemes is usually recommended. However, the current controlled voltage source converter topology for both power converters of the DFIG system seems to have slight edge over the other presented schemes because it does not require external circuitry, have reduced switching complexities and good performance during transient conditions. Table A1 in Appendix shows the numerical key performance index of the DFIG wind generator variables for the various solutions considered in this study. 113

ORIGINAL RESEARCH ARTICLE

Renewable Energy Focus  Volume 19–20

ORIGINAL RESEARCH ARTICLE

Renewable Energy Focus  Volume 19–20



2017

TABLE 3

Benefits and weak points of the various schemes.

ORIGINAL RESEARCH ARTICLE

Cases

Benefits

Weak points

Performance during transient

1.

One of the cheapest schemes. Cheaper than Schemes 2 and 3 for DC-link over voltage protection. Does not require rotor circuit of wind turbine disconnection during transient. Simple to implement with less switching circuitry because triggering signal is basically the DC-link voltage.

Faster in response with better performance than Scheme 2. Effects not noticed in the connected IG during transient

2.

Protects the power converters during grid faults. Can use either the rotor current or DC-link voltage or grid voltage as trigger signal for the IGBTs during operation. Has ability to absorb the excess currents in the rotor side converter during fault.

3.

Very effective for stability enhancement of the entire wind farm system because it provides additional reactive power during transient.

Cannot take care of DC-link under voltage protection, only suitable for over voltage protection. Dissipation of heat in DC-link resistive element. Protects only the power converter, does not have effects on the active power, rotor speed, active power etc. of the turbine. Disconnects the rotor circuit during grid fault, thus making the turbine loose active and reactive power control. More switching circuitry is required, hence more expensive than Scheme 1. Less effect in the responses of other variables of the wind turbine. Most expensive scheme because of external circuitry involved with additional capacitors and its control ancillary for reactive power provision.

4.

Cheaper than other schemes except Scheme 1. Little heat dissipation in rotor circuitry since its resistive value is low.

5.

Cheaper than other schemes except Scheme 1. Can mitigate DC-link under voltage. Has a great effect on the rotor speed, active power and other variables in fast recovery during post fault condition. Do not require external circuitry because it is embedded in the control system. Less circuitry in switching because of reduced dq-abc and abc to dq transformation. No DC-link protection for over voltage is required, hence reduced cost than VC-VSC Schemes. No external circuitry required

6.

7. 8.

Can take care of DC-link under voltage protection without external circuitry.

Require Schemes 1 or 2 for improved performance of DC-link under voltage. Cannot take care of DC-link over voltage. Limited ability to affect other variables of the wind turbine system during transient. Causes little amount of heat in stator circuitry. Require Schemes 1 and 2 for improved performance of DC-link under voltage. Cannot take care of DC-link over voltage.

Gives fewer oscillations of the wind turbine variables. Used to enhance Scheme 2 for more efficient performance and has effects on the connected fixed speed IG stability. Has longer settling time and causes distortions in the wind turbine variables. No stability effect on the connected fixed speed wind turbine. Faster settling time with less distortion of the wind turbine variables. No stability effect on the connected fixed speed wind turbine.

Cannot control DC-link under voltage during transient, unless implemented with Scheme 5.

Faster performance than VC-VSC schemes with good settling time. Has greater effect on the connected fixed speed wind turbine than other schemes considered.

Unable to take care of under voltage DC-link protection Complex switching circuitry, hence more expensive than conventional VC-VSC schemes

Slower in response than Scheme 6 but faster than VC-VSC schemes. Performance is enhanced using Scheme 5. No effect on the connected fixed speed wind turbine variables.

7. Conclusion Various Fault Ride Through (FRT) protection schemes were investigated in this paper to improve the transient stability of a wind farm composed of fixed and variable speed wind turbines. The FRT schemes considered in this study are; the use of a DC-link chopper in a Voltage Controlled Voltage Source Converter (VC-VSC) based DFIG wind turbine, crowbar, Static Synchronous Compensator (STATCOM), Series Dynamic Braking Resistor (SDBR), Current Controlled Voltage Source Converter (CC-VSC) topology, a combination of VC-VSC and CC-VSC and a parallel interleaved power converters for DFIG based wind turbine system. All the schemes considered were able to improve the transient stability of the variable speed wind turbine. The benefits and weak points of all the schemes were highlighted. Some schemes have more benefits than others. It was discovered that the CC-VSC based DFIG control offers better performance during transient periods. Therefore, this method could be effective in designing power converter system of 114

Slower in response than Scheme 1 and also causes oscillations in the wind turbine variables. Effects not noticed in the connected IG during transient conditions.

variable speed wind turbines. Also, new and existing wind farms could be a mixture of variable speed and fixed speed wind turbines because the stability of the wind farm would be enhanced.

Conflict of interest None declared.

Appendix The dynamics of the DFIG converter controller is a follow [57]: The electromagnetic torque and stator active power can be derived as Te ¼

3P lds iqr 22

(A1)

Ps ¼

3P ve lds iqr 22

(A2)



2017

ORIGINAL RESEARCH ARTICLE

Also, with the stator flux remaining unchanged, the reactive power can be derived as 3P ve lds ids 22

(A3)

Therefore, the d-axis component of the rotor current, idr, can be controlled to regulate the stator reactive power. Neglecting power losses in the converter, capacitor current can be described as follows:

lds ¼ Ls ids þ Lr idr

(A4)

idc ¼ C

Qs ¼

dV dc 3 ¼ migcd idcr 4 dt

(A5)

where igcd stands for the d-axis current flowing between the grid and grid side converter, idcr is the rotor side DC current, C is the DClink capacitance and m is the pulse width modulation index of the grid side converter. From (A5), it can be seen that DC-link voltage can be maintained constant, though small ripples might be present due to the instantaneous inequality between power entering into the grid side converter and the power at the rotor side converter.

where lds is the d-axis stator winding flux, ids, is d-stator current, iqr and idr are d- and q-axis rotor currents respectively, Ls and Lr are the stator leakage and rotor self-inductances, and ve is the electrical angular velocity (Figure A1). From (A1), the torque control can be achieved by controlling the rotor current component orthogonal to the stator winding flux in the q-axis. Then from (A2), stator active power is subsequently controlled. In (A4), ids is controllable by idr, with lds unchanged.

TABLE A1

Numerical key performance index of the DFIG variables. Control strategy

Parameters/variables

Dip/suppression (%)

Overshoot (%)

Recovery duration (sec)

Case 1 Case 2 No 1 or 2 Case 2 Case 2 Case 2 Case 2 Case 2 Case 2 Case 1 Case 1 + 3 Case 1 Case 1 + 3 Case 2 Case 2 + 3 Case 4 + 5 Case 4 + 5 Case 4 + 5 Case 4 + 5 Case 4 + 5 Case 4 + 5 Case 1 + 5 Case 1 + 5 Case 1 + 5 Case 1 + 5 Case 1 + 5 Case 1 + 5 Case 1 Case 6 Case 1 Case 6 Case 1 Case 6 Case 1 Case 6 Case 7 Case 8 Case 8 Case 8 Case 8 Case 5 + 8 Case 5 + 8 Case 5 + 8 Case 5 + 8

DC-link voltage DC-link voltage DC-link voltage Rotor current (c = 0.2) Rotor current (c = 0.8) Rotor current (c = 1.0) Terminal voltage (0.2) Terminal voltage (0.6) Terminal voltage (0.8) Terminal voltage Terminal voltage Rotor speed Rotor speed Terminal voltage Terminal voltage DC-link (rotor signal) DC-link (DC-V signal) DC-link (GridVsignal) Active power Active power Active power T. voltage (SDBR@A) T. voltage (SDBR@B) T. voltage (SDBR@C) Rot speed (SDBR@A) Rot speed (SDBR@B) Rot speed (SDBR@C) Terminal voltage Terminal voltage DC-link voltage DC-link voltage Rotor current Rotor current Terminal voltage Terminal voltage Terminal voltage DCvoltage (no control) DC voltage (only RSC) DC voltage (only GSC) DC-volt (RSC + GSC) DC voltage (no control) DC volt (only SDBR) DC volt (SDBRvICon) DC-volt (only Converter Interleaved)

35 35 35

10 25 75 30 5 60 3 1 2 10 9 8 1 35 10 10 10 10 20 5 3 5 0 0 45 33 0 5 0 10 20 3 22 5 0 0 5 4 4 5 10 10 10 10

1.20 1.20 1.30 1.20 1.00 1.60 2.10 1.20 0.90 2.10 1.30 1.50 1.00 0.40 0.30 1.00 0.60 0.40 1.00 0.20 0.20 1.00 0.70 0.20 4.80 4.70 0.20 1.90 1.00 1.50 0.50 0.60 0.50 1.50 1.20 1.20 0.25 0.22 0.22 0.20 0.60 0.50 0.40 5.50

8 8 8 8 8

10 10 30 35 60 -10 -10 -10 10 10 10 10 15 0 0 0 10 20

0 0 0 10 3 3 2 10 6 6 4

115

ORIGINAL RESEARCH ARTICLE

Renewable Energy Focus  Volume 19–20

ORIGINAL RESEARCH ARTICLE

Renewable Energy Focus  Volume 19–20

Pmax

ORIGINAL RESEARCH ARTICLE

Turbine Input Power [pu]

1.4 1.2

Pref

1.0 0.8 13m/s

0.6 12m/s

0.4

11m/s

0.2 0.0

6m/s

0.4

0.6

7m/s

0.8

8m/s 9m/s

1.0

1.2

10m/s

1.4

1.6

Turbine Speed[pu] FIGURE A1

DFIG wind turbine characteristics (MPPT).

The reactive power flowing into the grid from the grid side converter can be expressed as: Qg ¼

3 V g igcq 2

(A6)

where Vg is the magnitude of the grid phase voltage and igcq is q-axis current of grid side converter. Therefore, it is seen from (A5) and (A6), that DC-link voltage and Qg can be controlled respectively by adjusting igcd and igcq. References [1] Global Wind Energy Council, The Global wind report annual market update, 2012, 2013. [2] M.R.I. Sheikh, S.M. Muyeen, R. Takahashi, T. Murata, J. Tamura, J. Int. Rev. Autom. Control 3 (2008) 311–320. [3] C.L. Souza, et al. Proc. IEEE Port Power Technol. 2 (2001) 6. [4] J.A. Suul, M. Molinas, T. Undeland, IEEE Trans. Power Electron. 25 (5) (2010) 1240–1250. [5] J. Yu, X. Duan, Y. Tang, P. Yuan, IEEE Trans. Appl. Supercond. 12 (1) (2002) 750– 753. [6] S.M. Muyeen, R. Takahashi, M.H. Ali, T. Murata, J. Tamura, IEEE Trans. Power Syst. 23 (3) (2008) 1179–1187. [7] Singh Rajiv, K.S. Asheesh, K.S. Ashutosh, Int. J. Comput. Electr. Eng. 4 (1) (2012) 55–59. [8] S.D. Patil, Int. J. Electr. Electron. Eng. (IJEEE) 1 (3) (2012) 2231–5284. [9] A. Dezhamkhoy, A. Jamshid, Aust. J. Basic Appl. Sci. 5 (11) (2011) 1601–1608. [10] K.E. Okedu, S.M. Muyeen, R. Takahashi, J. Tamura, 6th IEEE-GCC Conference and Exhibition, Dubai, UAE, (2011), pp. 365–368. [11] J.C. Munoz, C.A. Canizares, Power Systems Conference and Exposition, 2011, 1–7. [12] F. Wu, X.P. Zhang, K. Godfrey, P. Ju, Proc. IET Generation, Transmission, Distribution, 2007, 751–760. [13] L. Sigrist, L. Rouco, Proc. IEEE/PES Power Systems Conference and Exposition, 2009, 1–7. [14] M.V.A. Nunes, J.A.P. Lopes, H.H. Zurn, U.H. Bezerra, R.G. Almeida, IEEE Trans. Energy Convers. 19 (2004) 692. [15] Ch. Eping, J. Stenzel, M. Poller, H. Muller, Impact of Large Scale Wind Power on Power System Stability, DigSILENT GmbH, Germany, April 2005, available at: http//www.digsilent.de/consulting/publications/paperGlasgow_DigSILENT.pdf. [16] D. Gautam, V. Vittal, T. Harbour, IEEE Trans. Power Syst. 24 (2009) 1426–1434. [17] E. Vittal, M. Malley, A. O’Keane, IEEE Trans. Power Syst. 25 (2009) 433–442. [18] H.A. Pulgar-Painemal, P.W. Sauer, Proc., 2010 Bulk Power System Dynamics and Control (iREP) Symposium, 2010, 1–7. [19] J.J. Sanchez-Gasca, N.W. Miller, W.W. Price, Proc. 2004 IEEE PES Power Systems Conf. Expo., 2004, 1148–1152.

116



2017

[20] O. Hasnaoui, M. Allagui, J. Energy South. Afr. 25 (4) (2014) 128–135. [21] T.D. Vrionis, X.I. Koutiva, N.A. Vovos, IEEE Trans. Power Syst. 29 (3) (2014) 1325– 1334. [22] Z. Tasneem, M.R.I. Sheikh, 10th International Conference on Mechanical Engineering, ICME, vol. 90, 2014, 698–703. [23] G. Pannell, D.J. Atkinson, B. Zahawi, IEEE Trans. Energy Conver. 25 (September (3)) (2010) 750–759. [24] M. Rodriguez, G. Abad, I. Sarasola, A. Gilabert, Crowbar control algorithms for doubly fed induction generator during voltage dips, Presented at the 11th Eur. Conf. Power Electronics Application, Dresden, Germany, September 11–14, 2005. [25] J. Yao, H. Li, Y. Liao, Z. Chen, IEEE Trans. Power Electron. 23 (3) (2008) 1205–1213. [26] R. Takahashi, J. Tamura, M. Futami, M. Kimura, K. Idle, IEEJ Trans. Power Energy 126 (2) (2006) 225–235. [27] M.B.C. Salles, J.R. Cardoso, A.P. Grilo, C. Rahman, K. Hameyer, in: Proceedings of IEEE International Electric Machines and Drives Conference—IEMDC, Miami, FL, USA, 2009. [28] K.E. Okedu, S.M. Muyeen, R. Takahashi, J. Tamura, International Conference on Electrical Machines and Systems (ICEMS), Seoul, South Korea, 2010. [29] H. Awad, J. Svensson, M. Bollen, IEEE Trans. Power Electron. 19 (2004) 837–846. [30] R.G. de Almeida, J.A. Lopez, J.A.L. Barreiros, IEEE Trans. Power Syst. 19 (4) (2004) 1942–1950. [31] A.O. Ibrahim, T.H. Nguyen, D. Lee, S. Kim, in: Proceedings IEEE-ECCE (Energy Conversion Congress and Exposition), California, USA, 2009. [32] Q. Wei, G.K. Venayagamorthy, R.G. Harley, IEEE Trans. Ind. Appl. 45 (1) (2009) 98–107. [33] A. Causebrook, D.J. Atkinson, A.G. Jack, IEEE Trans. Power Syst. 22 (3) (2007) 966– 975. [34] J. Yang, E. Fletcher, J. O’Reilly, IEEE Trans. Energy Convers. 25 (2) (2010) 422–432. [35] K.E. Okedu, S.M. Muyeen, R. Takahashi, J. Tamura, IEEE Trans. Sustain. Energy 3 (2) (2012) 242–254. [36] W. Park, B.C. Sung, J.W. Park, IEEE Trans. Appl. Supercond. 20 (3) (2010) 1177– 1181. [37] X. Yan, G. Venkataramanan, Y. Wang, in: Proceedings IEEE-ECCE (Energy Conversion Congress and Exposition), California, USA, 2009. [38] A. Petersson, S. Lundberg, T. Thiringer, Wind Energy J. 8 (2005) 251–263. [39] Y. Ling, X. Cai, Int. J. Electr. Power Energy Syst. 44 (1) (2013) 17–24. [40] M. Mohseni, M.A.S. Masoum, S.M. Islam, Electr. Power Syst. Res. 81 (7) (2011) 1456–1465. [41] C. Feltes, S. Engelhardt, J. Kretschmann, J. Fortmann, F. Koch, I. Erlich, in: Proceedings of the IEEE Power and Energy Society General Meeting—Conversion and Delivery of Electrical Energy in the 21st Century, IEEE, Pittsburgh, PA, USA, (July 2008), pp. 1–8. [42] Y. Wang, Q. Wu, H. Xu, Q. Guo, H. Sun, Energies 7 (7) (2014) 4140–4156. [43] R. Li, H. Geng, G. Yang, in: Proceedings of the International Electronics and Application Conference and Exposition (PEAC’14), Shanghai, China, (November 2014), pp. 496–501. [44] Z. Zheng, G. Yang, H. Geng, in: Proceedings of the 39th Annual Conference of the IEEE Industrial Electronics Society (IECON’13), IEEE, Vienna, Austria, (November 2013), pp. 5282–5287. [45] D.V.N. Ananth, G.V. Nagesh Kumar, ISA Trans. (2015). [46] A.V.N. Duggirala, V.G. Nagesh Kumar, J. Renew. Energy (2015) 1–14. , http:// dx.doi.org/10.1155/2015/490178, Article ID 490178. [47] S. Li, C. Deng, L. Chen, Z. Wu, F. Zheng, Int. Trans. Electr. Energy Syst. 25 (12) (2015) 3643–3650. [48] W. Guo, L. Xiao, S. Dai, et al. IEEE Trans. Ind. Electron. 62 (1) (2015) 332–342. [49] M.A. Chowdhury, W.X. Shen, N. Hosseinzadeh, H.R. Pota, IET Renew. Power Gen. 9 (2) (2015) 184–194. , http://dx.doi.org/10.1049/iet-rpg.2014.0035. [50] M.A. Chowdhury, W.X. Shen, N. Hosseinzadeh, H.R. Pota, Int. J. Sustain. Eng. (2015) 1–12. , http://dx.doi.org/10.1080/19397038.2015.1050480. [51] A.D. Hassan, G. Michalke, Renew. Energy 32 (2007) 1594–1610. [52] A. Petersson, Analysis, modeling and control of doubly-fed induction generators for wind turbines, (PhD Thesis), Division of Electric Power Engineering, Department of Energy and Environment, Chalmers University of Technology, Sweden, 2005. [53] G.G. Dimitrios, N.S. Athanasios, Machines 3 (2015) 2–26. [54] Manitoba HVDC Research Center, PSCAD/EMTDC Manual, 1994. [55] L. Huchel, M.S. El Moursi, H.H. Zeineldin, IEEE Trans. Sustain. Energy 6 (2) (2015) 303–312. [56] P.S. Flannery, G. Venkataramanan, IEEE Trans. Power Electron. 23 (3) (2008) 1126–1135. [57] M. Yamamoto, O. Motoyoshi, IEEE Trans. Power Electron. 6 (4) (1991) 624–629.