- Email: [email protected]
A stator voltage switching strategy for efficient low speed operation of DFIG using fractional rated converters
Renewable Energy 81 (2015) 389e399
Contents lists available at ScienceDirect
Renewable Energy journal homepage: www.elsevier.com/locate/renene
A stator voltage switching strategy for efficient low speed operation of DFIG using fractional rated converters Venkata Rama Raju Rudraraju, Nagamani Chilakapati, G. Saravana Ilango* Department of Electrical and Electronics Engineering, National Institute of Technology, Tiruchirappalli, 620015, Tamil Nadu, India
a r t i c l e i n f o
a b s t r a c t
Article history: Received 14 July 2014 Accepted 23 February 2015 Available online 3 April 2015
This paper presents a scheme for extending low speed operation of a wind driven Doubly Fed Induction Generator (DFIG) with limited rating converters (30%). In this scheme, a low voltage is applied to the stator at rotor speeds below the normal range (0.7 p.u.) while the nominal voltage is applied when the rotor speed is above this range. The switch-over to a lower stator voltage below a threshold speed facilitates to maintain the operational efficiency with the same fractional rated converters. Simulations results of a typical 250 kW DFIG in Matlab/Simulink environment reveal that, a lower stator voltage increases the generator efficiency at lower speeds without exceeding (limited) rating of the rotor side power converter. Typically the speed range is extended by 20% when reduction in voltage is 42%. With a 250 kW generator the energy harvested over one year is increased by 61.1 MWh with the proposed scheme. The scheme is simple, economical and can be implemented through speed sensor and contactors. Experiments on a laboratory test set up demonstrate the efficacy of the proposition. © 2015 Elsevier Ltd. All rights reserved.
Keywords: Wind energy conversion system Variable speed generator Doubly fed induction generator Extending speed range
1. Introduction Large scale Wind Energy Conversion Systems (WECS) using variable-speed generators are popular due to many advantages, including maximized power capture, reduced mechanical stress on the turbine, and reduced acoustic noise [1]. Asynchronous generators are more common for variable speed WECS because of smaller size, lower cost and low maintenance [3]. Among the variable speed asynchronous generators [4,5], the variable speed concept with partial scale converters is economically viable due to the versatile four quadrant operation of DFIG with decoupled power control in sub-synchronous and supersynchronous speed ranges [1], reduced rating of power converters, DC bus capacitor and the line side inductor [2]. The steady state analysis of DFIG in both sub and super-synchronous speeds has been widely reported [6e10]. Various schemes reported for DFIG control include direct torque control, direct power control, sensorless operation [11], vector control with a position sensor [12], stator flux oriented vector control of DFIG with and without position sensor [13], use of hysteresis
* Corresponding author. Tel.: þ91 04312503259. E-mail addresses: [email protected] (V.R.R. Rudraraju), [email protected] (N. Chilakapati), [email protected] (G. Saravana Ilango). http://dx.doi.org/10.1016/j.renene.2015.02.056 0960-1481/© 2015 Elsevier Ltd. All rights reserved.
controllers without position sensor [14], and DPC strategy [15,16] in the grid flux reference frame without position sensors. The Maximum Power Point Tracking (MPPT) method is the key to improve efficiency and energy extraction in the wind turbine system [17e19]. In Ref. [17] the MPPT method through the characteristic power curve is presented and performance of wind turbine with MPPT power control is explained. An analytical approach was suggested [19] to determine the rotor current commands for maximum mechanical power and minimum loss based on generator speed. Efforts were made to improve the efficiency of induction motor under light loads by changing the stator winding connection [20e22]. Most of the over sized induction motors operate with low efficiency and power factor under light loads [20]. The most common method for improving the efficiency and power factor of induction motors at light loads is to reduce the stator voltage by switching the stator winding connection from delta to star. In Ref. [21] a multi-step air gap flux regulation method is implemented with a star-delta switchable stator winding connection, thereby improving the efficiency and power factor over an extended speed range. In Refs. [23,24] the pole changing method is employed to extend the speed range of the induction motor. A few studies focused on improving energy extraction by extending the speed range of generator. Pole changing method is employed for speed extension in standalone wind driven self
390
V.R.R. Rudraraju et al. / Renewable Energy 81 (2015) 389e399
Nomenclature vs
stator voltage space vector
is vr
stator current space vector rotor voltage space vector
ir 4s Ps,Qs Rs,Rr Ls,Lr Lm
rotor current space vector stator magnetizing flux space vector stator active and reactive power stator and rotor phase winding resistances stator and rotor phase winding inductances magnetizing inductance angular frequency of stator flux angular speed of rotor (electrical) mechanical power wind velocity
ue um Pm vw
Subscripts d,q synchronous reference frame Superscripts e synchronous reference frame s stationary reference frame * reference value
excited induction generator [30]. In Ref. [25] the rotor is supplied through a cyclo-converter and the range of speed is extended by changing the number of pairs of poles in stator. Although the speed range is widened, the effects of this on the power output and the generator efficiency have not been explained. Complexity in implementation using a cyclo-converter, aspects of smooth transition between two sets of pole pairs, also need consideration. While most studies in literature considered zero to 30% slip range, wider speed operation is also reported, although with an increased rating of the rotor side converters [4]. Since the reactive power demand of the DFIG and also the rating of power converters increase proportionately with slip, it is a common practice to restrict the maximum slip to about 30%. To facilitate larger slip operation with the same 30% rated converters, a technique involving slackening of the zero reactive power constraint on stator side was proposed [26]. However the issue of decaying generator efficiency at low speeds was not looked into. Another study [27] reported an operating speed range of 50e90% of the synchronous speed with improved efficiencies up to 95%. Conventional DFIG operation for normal wind conditions and stator short circuited mode at lower wind speed are considered. In Ref. [28] the DFIG is operated at rated V/f using the power converters rated for 5% of machine rating, but the disadvantage is that number of converters required is higher. The aspect of extended low speed operation at moderate efficiencies with limited rating converters has not been addressed adequately by researchers so far. This study presents a technique for extending the speed range while maintaining a reasonable efficiency of the generator with the same limited rating converters in the rotor circuit. In general, the DFIG operation below 70% of the synchronous speed is marked by low efficiency due to increased reactive power demand of the generator accompanied by increased losses. Also the converters need to be rated higher. Even with higher rated converters the efficiency is poor below a certain speed. It is possible to improve the efficiency at low speeds by reducing the stator voltage thereby decreasing the excitation requirement. This fixed, lower voltage can be provided through a tapping on the
stator side transformer. At normal and high speeds (i.e. higher mechanical power) nominal stator voltage operation can be pursued, giving the nominal efficiencies. Broadly speaking, the choice of lower stator voltage depends on the new maximum slip under consideration. Since most transformers have the provision of tappings, such a voltage switch-over does not warrant any design modifications or additional cost. Further, since the speed is a slowly changing variable, a speed sensor along with a contactor based voltage switching is adequate for real time implementation. The scheme is simple and economical in implementing and yields greater annual returns through increased energy harvesting. The proposed scheme is verified through MATLAB simulations on a 250 kW wind turbine and also through simulations and experiments on a 2.3 kW laboratory test bench using Altera cyclone II FPGA controller. Results demonstrate the feasibility of moderately efficient extended speed operation with the limited rating converters. This paper is organized as follows. Detailed mathematical modeling of a doubly fed induction machine including iron losses is described in section 2. Further the decoupled control of DFIG with RSC and GSC are explained in section 2. Maximum Power Point Tracking from the wind is explained in section 3. Section 4 describes the proposed scheme of extended low speed operation of DFIG. In the first case study, the simulation results of a 250 kW wind driven DFIG are presented and in the second case, the simulation and test results for a 2.3 kW laboratory DFIG are illustrated in section 5. 2. Mathematical modeling Fig. 1 shows the schematic diagram of DFIG system. Rotor Side Converter (RSC) is used to control active and reactive power of the generator and also to track the maximum power from the wind. The Grid Side Converter (GSC) is used to maintain DC link constant. Transformers are used to step down the voltage at stator side and rotor side. 2.1. Modeling of DFIM Fig. 2a shows the spatial distribution of stator voltage, stator flux and rotor currents in different reference frames. Stator voltage and stator flux are considered along q and d-axes in the synchronous reference frame respectively. Fig. 2b shows the per phase equivalent circuit of induction machine. The stator and rotor voltages in synchronous reference frame, can be expressed as e
ves ¼ Rs is þ e
ver ¼ Rr ir þ
d4es þ ju4es ; dt
(1)
d4er þ jðue um Þ4er ; dt
(2)
For higher rated machines the iron loss component can be neglected. But in smaller rating machines it cannot be neglected. So iron loss component is considered for modeling (laboratory machine). The iron loss component of the machine is modeled as a resistance in parallel to the mutual magnetizing inductance as shown in Fig. 2b. The resistance Rfe provokes an active current consumption (ife) but does not contribute to flux. The flux is created by the current flow through the magnetizing and leakage inductances. Hence, in the synchronous reference frame the stator and rotor fluxes are given by
V.R.R. Rudraraju et al. / Renewable Energy 81 (2015) 389e399
391
Fig. 1. Schematic diagram of DFIG.
e
e
e
e
4es ¼ Lls is þ Lm im ; 4er ¼ Lm is þ Lls ir ;
þ
e ir
¼
e ife
þ
e im
i*qr
(5)
This component is responsible for regulating the active power output of the machine. For regulating the reactive power input of the machine, the d-component of rotor current is controlled using the following relationship:
(6)
" # 2 Qs* i*dr ¼ þ idfe þ idm 3 vqs
Voltage across shunt branch is e
Rfe ife ¼ Lm
e
dim e þ jue im dt
" # 2 Ps* ¼ þ iqfe þ iqm 3 vqs
(4)
and applying KCL e is
The reference for q-axis rotor current is set based on (5) That is
(3)
(9)
(10)
2.2. Control of rotor side converter The q-axis current loop controls the active power flow, while the d-axis current loop regulates the flux by controlling the reactive power flow. The dynamic equations governing the rotor currents in the synchronous rotating frame are as follows:
didr 1 di ¼ vdr Rr iqr þ ðue um Þ Lm iqm þ Llr iqr Lm dm Llr dt dt
(7)
diqr diqm 1 ¼ vqr Rr iqr ðue um ÞðLm idm þ Llr idr Þ Lm Llr dt dt
(8)
The schematic block diagram for control of RSC is shown in Fig. 3a. The active loop (q-axis) controls the torque of the machine, while the reactive loop (d-axis) controls the flux of the machine.
2.3. Grid side converter control The equations that describe the dynamics of the GSC are summarized together as,
Lf Rf Lf Rf
!
!
vqf vqs d i þ iqf ¼ þ dt qf Rf Rf vdf d i þ idf ¼ þ dt df Rfe
! Lf ue idf Rf
! Lf ue iqf Rf
(11)
(12)
where, if and vf denote the current and voltage of GSC respectively. The dynamic equation representing the dc link voltage is given by
Fig. 2. DFIM modeling. (a) Spatial vectors in different reference frames, (b) Per phase equivalent circuit.
392
V.R.R. Rudraraju et al. / Renewable Energy 81 (2015) 389e399
Fig. 3. Control block diagram for (a) RSC, (b) GSC.
C
d ðv Þ ¼ dt dc
vqs i iL vdc qf
(13)
where, iL ¼ ðvqr iqr þ vdr idr =vdc Þ The control of GSC (Fig. 3b) contains two loops, one for active power and the other for reactive power control respectively. The active loop functions to keep dc-link voltage constant by controlling the active power flow in the circuit. The voltage controller output gives the current reference for q-axis current loop such that the there is an active power balance between ac side and dc side of the power circuit. Reference for the d-axis current loop is taken directly depending on the desired power factor. 3. Wind turbine
dum 1 ¼ ½Tm TL Bum J dt
(15)
where, J is the turbine inertia, B is friction coefficient, Tm is torque developed by the turbine, TL is the torque due to load and um is the rotor speed. The optimum power [18] from wind turbine is given by
Popt ¼ Kopt umopt
(16)
where, Kopt ¼ 0:5prCpmax R5 =l3opt
Power produced by a wind turbine is given by
Pm ¼ 0:5prCp R2 v3w
(14)
where, R is the radius of the turbine radius, r is air density, Cp is power coefficient. The dynamic equation of the wind turbine is given by
umopt ¼
lopt Vw R
(17)
umopt is the optimum rotor speed at which power for certain wind speed is maximum and lopt is the optimum tip speed ratio.
V.R.R. Rudraraju et al. / Renewable Energy 81 (2015) 389e399
4. Proposed control strategy for extending the speed range with fractional rating converters Commonly the operating slip range of DFIG systems [1e3] is from zero to 30%. To operate at higher slips (lower speeds), the converters need to be rated higher. In the proposed method, the lower end of speed range can be extended using the same fractional rated converters while not compromising on the efficiency. Applying a lower voltage to the stator at lower speeds has the effect of reducing the reactive power needed and also restricting the VA of the rotor converters. In addition, this reduces the losses in the generator also. The selection of the lower stator voltage to be applied is based on the choice of the minimum speed and the voltage (which is typically 30% of machine rating) and current ratings of the RSC. Further, the choice of threshold speed for voltage change over is based on the efficiency Vs. speed characteristics of the DFIG system and the voltage and current ratings of the RSC. The rotor voltage is given by
Vr ¼ sVs
(18)
Stator, rotor copper losses and iron losses are given by
Pcus ¼ 3Is2 Rs
(20)
Pcur ¼ 3Ir2 Rr
(21)
and 2 Pfe ¼ 3Ife Rfe
Ptotal ¼ Pcus þ Pcur þ Pfe
(23)
Stator and rotor active powers are given by
Ps ¼ vds ids þ vqs iqs
(24)
Pr ¼ vdr idr þ vqr iqr
(25)
The efficiency of the generator is given by
efficiency ¼
x ¼ 0:3*k 0:3
h¼
where k ¼ VNV/VLV Fig. 4 shows the flow chart for predetermination of generator performance for a given speed. The machine and turbine parameters are initialized and the stator voltage is selected. The wind speed is initialized and the corresponding optimum rotational speed is calculated from (17) and the maximum power is calculated using (16). Further, the machine losses and the efficiency are calculated. The generator performance predetermination is carried out till wind speed is 1 m/s.
(22)
Total losses are given by
As the stator voltage (Vs) is reduced at lower speeds, rotor voltage (Vr) also reduces (for same slip). Hence slip range can be extended till Vr reaches its limit (0.3 p.u) with lower stator voltage. Therefore the slip range is extended by
(19)
393
Stator active power þ rotor active power mechanical power
Ps þ Pr Pmech
(26)
5. Results and discussion Based on the analysis presented in the previous section, computer simulations are carried out to investigate the performance of a grid connected DFIG. Two cases are considered. One is a 250 kW machine and the other is a 2.3 kW laboratory machine. Higher rated wind turbines starting from 1 MW are in practice. However, for simulation study 250 kW rated machine is considered. Moreover,
Fig. 4. Flow chart for predetermination of generator performance for different wind speeds.
394
V.R.R. Rudraraju et al. / Renewable Energy 81 (2015) 389e399
the same proposed strategy can be implemented for wind turbine having higher power. Figs. 5e6 present the simulation results for 250 kW DFIG. 5.1. Case study-1: a 250 kW wind turbine The parameters of 250 kW wind turbine are given in appendix A. Three different stator line voltages are considered for study. These are; a) Low Voltage (LV ¼ 230 V or 0.577 p.u.), b) Medium Voltage (MV ¼ 300 V or 0.75 p.u.) and c) Nominal Voltage (NV ¼ 400 V or 1.00 p.u.). Fig. 5 shows the performance of DFIG system for different voltages. The maximum power from the wind is obtained using the MPPT algorithm [17]. The stator active power (Fig. 5a) is nearly the same for all voltages (NV, MV, or LV) except for a marginal difference. At a given speed the stator current with LV is higher than that with MV or NV since a higher stator current is required to maintain the same stator power output with a lower voltage (Fig. 5b). Fig. 5c shows the rotor voltage and current variation with speed. Considering the rotor voltage limit to be 30% of stator voltage, the boundary for the rotor voltage limit is
also shown. Since rotor voltage is proportional to slip times stator voltage, the rotor voltage for LV is less than that for MV and the rotor voltage for MV is less than that for NV. As seen from Fig. 5c using LV or MV at lower speeds results in lower VA for the power converters. For the 250 kW DFIG, the speed threshold for voltage change is found to be 0.72 p.u. The selection of the lower stator voltage to be applied is determined based on the choice of minimum speed (or maximum slip) and the voltage and current limits of the rotor circuit power converters. From Fig. 5c, it is notable that the voltage and current of RSC are within limits for LV case. Hence the LV (i.e., 0.577 p.u. or 230 V) is selected as the second voltage for stator beyond the slip of 0.3. Since the generator excitation (reactive power) is provided through the RSC, the rotor current is high at low speeds for NV (Fig. 5c and d). At lower speeds the LV connection gives higher efficiency compared to MV or NV (Fig. 6b) since the total copper loss with LV is less (Fig. 6a). It is also observed that the performance of DFIG system deteriorates below 0.6 p.u. speed (6 m/s wind speed) even with increased rating of converters (Fig. 6a). The vector diagrams for DFIG for NV and LV at 0.5 slip for the same mechanical power are
Fig. 5. Performance of DFIG system with different voltages. (a) Stator active power vs. rotor speed, (b) Stator current vs. rotor speed, (c) Rotor voltage, current vs. rotor speed and (d) Rotor Reactive power vs. speed.
V.R.R. Rudraraju et al. / Renewable Energy 81 (2015) 389e399
395
Fig. 6. Efficiency of the DFIG system and vector diagram with different voltages. (a) Total copper losses vs. rotor speed, (b) Efficiency vs. rotor speed, (c) Vector diagram for NV at 0.5 slip and (d) Vector diagram for LV at 0.5 slip.
shown in Fig. 6c and d respectively. With LV the rotor current is reduced significantly though the stator current rises slightly. Thus, with LV the total copper losses are reduced at lower speeds leading to efficiency improvement (iron losses are neglected for 250 kW
machine). Hence, switching to a lower stator voltage at lower speeds improves generator efficiency. The point of switching can be either at the instant when the rotor voltage exceeds the rating of RSC or when the efficiency decreases below a desired threshold.
Fig. 7. Comparison of energy production with conventional and proposed scheme. (a) Wind speed data for one year, (b) Energy harvested in one year.
396
V.R.R. Rudraraju et al. / Renewable Energy 81 (2015) 389e399
Table 1 Energy harvested for one year.
Conventional Proposed
Energy (MWh)
Increase in energy (MWh)
681.87 742.97
61.1
5.2. Annual energy harvested To evaluate the performance of proposed configuration, wind speed data registered in Tiruchirappalli (Tamil Nadu, India) from August 2013 to july 2014 is considered [29] (Fig. 7a). It is assumed that the generator is not in operation when the rotor voltage limits of RSC are violated. Energy is computed for the proposed and conventional configuration. The base value for energy is 2190 MWh. In the proposed scheme for wind speeds between 4.8 and 6.6 m/s DFIG is operated with LV and above 6.6 m/s the DFIG is
operated with NV. Whereas, in conventional scheme the DFIG is operated with NV above 6.6 m/s. Fig. 7b and Table 1 show the details of energy harvested (p.u.) in one year. Thus with the 250 kW generator the energy harvested over one year is increased by 61.1 MWh with the proposed scheme. This is an increase of about 2.8% with a single generator. Thus the proposed scheme has the potential to increase the energy output significantly if implemented with more number of generators, such as all those connected in one wind farm through common switchgear and transformer. 5.3. Experimentation: case study-2 Experimentation is carried out on 2.3 kW slip ring induction machine for the verification of the proposed scheme. Fig. 8a shows the schematic diagram of the experimental setup. The machine parameters are given in appendix A. A 5-hp DC machine is used to
Fig. 8. (a) Schematic diagram of experimental setup, (b) Laboratory setup.
V.R.R. Rudraraju et al. / Renewable Energy 81 (2015) 389e399
397
Fig. 9. Performance of DFIG system with different voltages. (a) Mechanical power vs. wind velocity, (b) Stator real power vs. rotor speed, (c) Stator current vs. rotor speed and (d) Rotor current vs. rotor speed.
simulate the wind turbine characteristics using armature controlled dc drive. Altera cyclone II FPGA is used to generate firing pulses for RSC and GSC. To verify the increase in efficiency at lower speeds with limited rated converters the simulation and experimental tests are carried out for two different cases of stator line voltage a) Low Voltage (LV ¼ 240 V) and b) Nominal Voltage (NV ¼ 415 V). The rotor power circuit consists of two power converters (MD B6CI 800/415-30 F) of Semikron make. The Rotor side PWM inverter is controlled using a frequency of 5 kHz. The rotor position is found using sensorless position estimation. Photograph of laboratory setup is shown in Fig. 8b. Fig. 9a shows the mechanical power Vs. wind speed with two different voltages. Since the wind turbine mechanical characteristics considered are same for both NV and LV, the predicted characteristics for NV and LV overlap. It is notable that the experimental results are close to predicted results. The implementation of MPPT in the experimental set up enables tracking the maximum power, as indicated by the markers in Fig. 9a. Fig. 9b shows the variation in stator real power as speed of machine is varied for NV and LV settings. The power with LV is higher compared to NV at lower speeds. The stator reactive power reference is set to zero. Fig. 9c and d show the stator and rotor currents respectively as rotor speed is varied. Stator current is higher for LV for the same stator power.
With NV, since the magnetizing current is high at lower speeds, the rotor current is higher till 1375 rpm and above 1375 rpm the rotor current is higher with LV. Since, the efficiency below 775 rpm is nearly zero, only speeds higher than 775 rpm are considered. Fig. 10a shows the rotor voltage for LV and NV and the rotor voltage limit is indicated. The rotor voltage is high at higher slip. Since magnetizing current is supplied from rotor side, rotor reactive power is less with LV compared to that with NV as shown in Fig. 10b. Since reactive power is less with LV, rotor side VA is also less and is shown in Fig. 10c. Up to 1722 rpm, the efficiency is higher for LV and above 1722 rpm efficiency is high for NV. The mechanical power corresponding to 1722 rpm is 3138 W. It is seen that the efficiency at lower speeds is high for LV, this is due to the reduction in magnetizing current, rotor current and total losses (Fig. 10d). Table 2 shows the performance comparison of the DFIG system with different stator voltages. Data in bold italics correspond to efficiency above 0.9, rotor voltage & VA below 0.3 p.u. It is observed that with 0.3 p.u. rotor side converter rating, the speed is extended down to 0.5 p.u. with efficiency greater than 0.9. For speeds below 0.7 p.u., when the stator voltage is switched to LV the rotor voltage does not exceed the RSC rating of 0.3 p.u. Thus, with the proposed method the speed range is extended by 0.2 p.u. without violating the rating of the rotor side converter.
398
V.R.R. Rudraraju et al. / Renewable Energy 81 (2015) 389e399
Fig. 10. Performance of DFIG system with different voltages. (a) Rotor voltage vs. rotor speed, (b) Rotor active power vs. rotor speed, (c) Rotor reactive power vs. rotor speed and (d) Efficiency vs. rotor speed.
increased by 61.1 MWh with the proposed scheme. Overall, the results demonstrate the viability of the scheme.
6. Conclusions The study demonstrates extended low speed operation of a grid connected DFIG without exceeding the ratings of the rotor side power converters. The generator efficiency is shown to be maintained with the same limited rating converters (rated 30% of the generator). The laboratory machine is found to be less efficient (due to high stator and rotor resistances), being much smaller in rating. For the 250 kW DFIG, low speed operation down to 0.5 p.u. can be achieved with 30% rated power converters at generator efficiencies better than 90%. Also the energy harvested over one year is
Table 2 Performance comparison of the DFIG with different voltages. Case study
1
Speed (p.u.)
0.5 0.6 0.7
Rotor VA (p.u.)
Efficiency
Rotor voltage (p.u.)
NV
LV
NV
LV
NV
LV
0.86 0.91 0.93
0.91 0.93 0.92
0.51 0.42 0.34
0.31 0.25 0.20
0.25 0.22 0.17
0.11 0.11 0.10
Appendix A
Parameters Machine rating Stator voltage (V) Stator current (A) Rotor current (A) Stator resistance (U) Rotor resistance (U) Iron loss resistance (U) Stator inductance (mH) Rotor inductance (mH) Stator turns/rotor turns
250 kW 400 370 370 0.02 0.02 e 4.4 4.4 1
Power Converters: 415 V, 30 A SEMIKRON Inverter. DC-Link Voltage and Capacitor: 800 V, 2000 mF. DC Machine: 5hp, 1500 r/min. Armature: 220 V, 19 A; Field: 220 V, 1 A.
2.3 kW 415 4.7 7.5 3.678 5.26 270 0.30682 0.30682 1.92
V.R.R. Rudraraju et al. / Renewable Energy 81 (2015) 389e399
References [1] Muller S, Deicke M, De Doncker RW. Doubly fed induction generator systems for wind turbines. IEEE Ind Appl Mag 2002;8(3):26e33. [2] Datta R, Ranganathan VT. Variable-speed wind power generation using doublyfed wound rotor induction machine-a comparison with alternative schemes. IEEE Trans Energy Convers 2002;17(3):414e21. [3] Zou Yu, Elbuluk ME, Sozer Y. Simulation comparisons and implementation of induction generator wind power systems. IEEE Trans Ind Appl 2013;49(30): 1119e28. [4] Li H, Chen Z. Overview of different wind generator systems and their comparisons. IET Renew Power Gener 2008;2(2):123e38. [5] Mittal Rajveer, Sandhu KS, Jain DK. An overview of some important issues related to wind energy conversion system (WECS). Int J Environ Sci Devel 2010;1(4):351e63. [6] Cadirci I, Ermis M. Double-output induction generator operating at subsynchronous and supersynchronous speeds: steady-state performance optimization and wind-energy recovery. IEE Proc Elect Power Appl 1992;139(5): 429e42. [7] Alkandari Ahmad M, Soliman Soliman Abd-Elhady, Abdel-Rahman Mansour H. Steady state analysis of a doubly fed induction generator. Energy Power Eng 2011;3:393e400. [8] Sediki H, Ould Abdeslam Dj, Otmane-cherif T, Bechouche A, Mesbah K. Steadystate analysis and control of double fed induction motor. World Acad Sci Eng Tech 2012;61:153e61. [9] Vicatos MS, Tegopoulos JA. Steady state analysis of doubly- fed induction generator under synchronous operation. IEEE Trans Energy Convers 1989;4(3):495e501. [10] Miller TJE. Theory of the doubly-fed induction machine in the steady state. In: Int. Conf. on Elect. Machines, Rome; 2010. p. 1e6. [11] Cardenas R, Pena R, Alepuz S, Asher G. Overview of control systems for the operation of DFIGs in wind energy applications. IEEE Trans Ind Elect 2013;60(7):2776e98. [12] Pena R, Clare JC, Asher GM. Doubly fed induction generator using back-toback PWM converters and its application to variable speed wind-energy generation. IEE Proc Elect Power Appl 1996;143(3):231e41. [13] Hopfensperger B, Atkinson DJ, Lakin RA. Stator-flux-oriented control of a doubly-fed induction machine with and without position encoder. IEE Proc Elect Power Appl 2000;147(43):241e50. [14] Datta R, Ranganathan VT. Direct power control of grid-connected wound rotor induction machine without rotor position sensors. IEEE Trans Power Elect 2001;16(3):390e9. [15] Amiri N, Madani SM, Lipo TA, Zarchi HA. An improved direct decoupled power control of doubly fed induction machine without rotor position sensor and
[16] [17]
[18]
[19]
[20]
[21]
[22]
[23]
[24] [25]
[26]
[27] [28]
[29] [30]
399
with robustness to parameter variation. IEEE Trans Energy Convers 2012;27(4):873e84. Qu Liyan, Qiao Wei. Constant power control of DFIG wind turbines with supercapacitor energy storage. IEEE Trans Ind Appl 2011;47(1):359e67. Zou Yu, Elbuluk ME, Sozer Y. Stability analysis of maximum power point tracking (MPPT) method in wind power systems. IEEE Trans Ind Appl 2013;49(3):1129e36. Zhao Y, Zou XD, Xu YN, Kang Y, Chen J. Maximal power Point tracking under speedmode control for wind energy generation system with doubly fed introduction generator. In: IEEE Int. Conf. Power Elect. and Motion Control1; 2006. p. 1e5. Chen Bo-An, Lu Tien-Kei, Hsu Yuan-Yih, Chen Woei-Luen, Lee Zheng-Cong. An analytical approach to maximum power tracking and loss minimization of a doubly fed induction generator considering core loss. IEEE Trans Energy Convers 2012;27(2):449e56. Ferreira FJTE, de Almeida AT, Baoming Ge, Faria SP, Marques JM. Automatic change of the stator-winding connection of variable-load three-phase induction motors to improve the efficiency and power factor. In: IEEE Int. Conf. Ind. Tech; 2005. p. 1331e6. Cistelecan MV, Ferreira FJTE, Popescu M. Adjustable flux three-phase ac machines with combined multiple-step staredelta winding connections. IEEE Trans Energy Convers 2010;25(2):348e55. Lei Yi, Zhao Zhengming, Wang Shuping, Dorrell DG, Xu Wei. Design and analysis of staredelta hybrid windings for high-voltage induction motors. IEEE Trans Ind Elect 2011;58(9):3758e67. Osama M, Lipo TA. Modeling and analysis of a wide-speed-range induction motor drive based on electronic pole changing. IEEE Trans Ind Appl 1997;33(5):1177e84. Bird BM, Ridge J. Induction-motor speed-control schemes employing statorvoltage variation and pole-change windings. In: PROC. IEE; 1971. p. 537e42. Djurovic M, Joksimovic G, Saveljic R, Maricic I. Double fed induction generator with two pair of poles. In: Int. Conf. Electrical Machines and Drives; Sept. 1995. p. 449e52. Park JW, Lee KW, Lee HJ. Control of active power in a doubly-fed induction generator taking into account the rotor side apparent power. In: Power Electronics Specialists Conf., 3; 2004. p. 2060e4. Smith AC, Todd R, Barnes M, Tavner PJ. Improved energy conversion for doubly fed wind generators. IEEE Trans Ind Appl 2006;42(6):1421e8. Muljadi E, Singh M, Gevorgian V. Doubly fed induction generator in an offshore wind power plant operated at rated V/Hz. IEEE Trans Ind Appl 2013;49(5):2197e205. Wind data, http://weatherspark.com/. Ammasaigounden N, Subbiah M, Krishnamurthy MR. Wind-driven selfexcited pole-changing induction generators. IEE Proc B Electr Power Appl 1986;133(5):315e21.
Contents lists available at ScienceDirect
Renewable Energy journal homepage: www.elsevier.com/locate/renene
A stator voltage switching strategy for efficient low speed operation of DFIG using fractional rated converters Venkata Rama Raju Rudraraju, Nagamani Chilakapati, G. Saravana Ilango* Department of Electrical and Electronics Engineering, National Institute of Technology, Tiruchirappalli, 620015, Tamil Nadu, India
a r t i c l e i n f o
a b s t r a c t
Article history: Received 14 July 2014 Accepted 23 February 2015 Available online 3 April 2015
This paper presents a scheme for extending low speed operation of a wind driven Doubly Fed Induction Generator (DFIG) with limited rating converters (30%). In this scheme, a low voltage is applied to the stator at rotor speeds below the normal range (0.7 p.u.) while the nominal voltage is applied when the rotor speed is above this range. The switch-over to a lower stator voltage below a threshold speed facilitates to maintain the operational efficiency with the same fractional rated converters. Simulations results of a typical 250 kW DFIG in Matlab/Simulink environment reveal that, a lower stator voltage increases the generator efficiency at lower speeds without exceeding (limited) rating of the rotor side power converter. Typically the speed range is extended by 20% when reduction in voltage is 42%. With a 250 kW generator the energy harvested over one year is increased by 61.1 MWh with the proposed scheme. The scheme is simple, economical and can be implemented through speed sensor and contactors. Experiments on a laboratory test set up demonstrate the efficacy of the proposition. © 2015 Elsevier Ltd. All rights reserved.
Keywords: Wind energy conversion system Variable speed generator Doubly fed induction generator Extending speed range
1. Introduction Large scale Wind Energy Conversion Systems (WECS) using variable-speed generators are popular due to many advantages, including maximized power capture, reduced mechanical stress on the turbine, and reduced acoustic noise [1]. Asynchronous generators are more common for variable speed WECS because of smaller size, lower cost and low maintenance [3]. Among the variable speed asynchronous generators [4,5], the variable speed concept with partial scale converters is economically viable due to the versatile four quadrant operation of DFIG with decoupled power control in sub-synchronous and supersynchronous speed ranges [1], reduced rating of power converters, DC bus capacitor and the line side inductor [2]. The steady state analysis of DFIG in both sub and super-synchronous speeds has been widely reported [6e10]. Various schemes reported for DFIG control include direct torque control, direct power control, sensorless operation [11], vector control with a position sensor [12], stator flux oriented vector control of DFIG with and without position sensor [13], use of hysteresis
* Corresponding author. Tel.: þ91 04312503259. E-mail addresses: [email protected] (V.R.R. Rudraraju), [email protected] (N. Chilakapati), [email protected] (G. Saravana Ilango). http://dx.doi.org/10.1016/j.renene.2015.02.056 0960-1481/© 2015 Elsevier Ltd. All rights reserved.
controllers without position sensor [14], and DPC strategy [15,16] in the grid flux reference frame without position sensors. The Maximum Power Point Tracking (MPPT) method is the key to improve efficiency and energy extraction in the wind turbine system [17e19]. In Ref. [17] the MPPT method through the characteristic power curve is presented and performance of wind turbine with MPPT power control is explained. An analytical approach was suggested [19] to determine the rotor current commands for maximum mechanical power and minimum loss based on generator speed. Efforts were made to improve the efficiency of induction motor under light loads by changing the stator winding connection [20e22]. Most of the over sized induction motors operate with low efficiency and power factor under light loads [20]. The most common method for improving the efficiency and power factor of induction motors at light loads is to reduce the stator voltage by switching the stator winding connection from delta to star. In Ref. [21] a multi-step air gap flux regulation method is implemented with a star-delta switchable stator winding connection, thereby improving the efficiency and power factor over an extended speed range. In Refs. [23,24] the pole changing method is employed to extend the speed range of the induction motor. A few studies focused on improving energy extraction by extending the speed range of generator. Pole changing method is employed for speed extension in standalone wind driven self
390
V.R.R. Rudraraju et al. / Renewable Energy 81 (2015) 389e399
Nomenclature vs
stator voltage space vector
is vr
stator current space vector rotor voltage space vector
ir 4s Ps,Qs Rs,Rr Ls,Lr Lm
rotor current space vector stator magnetizing flux space vector stator active and reactive power stator and rotor phase winding resistances stator and rotor phase winding inductances magnetizing inductance angular frequency of stator flux angular speed of rotor (electrical) mechanical power wind velocity
ue um Pm vw
Subscripts d,q synchronous reference frame Superscripts e synchronous reference frame s stationary reference frame * reference value
excited induction generator [30]. In Ref. [25] the rotor is supplied through a cyclo-converter and the range of speed is extended by changing the number of pairs of poles in stator. Although the speed range is widened, the effects of this on the power output and the generator efficiency have not been explained. Complexity in implementation using a cyclo-converter, aspects of smooth transition between two sets of pole pairs, also need consideration. While most studies in literature considered zero to 30% slip range, wider speed operation is also reported, although with an increased rating of the rotor side converters [4]. Since the reactive power demand of the DFIG and also the rating of power converters increase proportionately with slip, it is a common practice to restrict the maximum slip to about 30%. To facilitate larger slip operation with the same 30% rated converters, a technique involving slackening of the zero reactive power constraint on stator side was proposed [26]. However the issue of decaying generator efficiency at low speeds was not looked into. Another study [27] reported an operating speed range of 50e90% of the synchronous speed with improved efficiencies up to 95%. Conventional DFIG operation for normal wind conditions and stator short circuited mode at lower wind speed are considered. In Ref. [28] the DFIG is operated at rated V/f using the power converters rated for 5% of machine rating, but the disadvantage is that number of converters required is higher. The aspect of extended low speed operation at moderate efficiencies with limited rating converters has not been addressed adequately by researchers so far. This study presents a technique for extending the speed range while maintaining a reasonable efficiency of the generator with the same limited rating converters in the rotor circuit. In general, the DFIG operation below 70% of the synchronous speed is marked by low efficiency due to increased reactive power demand of the generator accompanied by increased losses. Also the converters need to be rated higher. Even with higher rated converters the efficiency is poor below a certain speed. It is possible to improve the efficiency at low speeds by reducing the stator voltage thereby decreasing the excitation requirement. This fixed, lower voltage can be provided through a tapping on the
stator side transformer. At normal and high speeds (i.e. higher mechanical power) nominal stator voltage operation can be pursued, giving the nominal efficiencies. Broadly speaking, the choice of lower stator voltage depends on the new maximum slip under consideration. Since most transformers have the provision of tappings, such a voltage switch-over does not warrant any design modifications or additional cost. Further, since the speed is a slowly changing variable, a speed sensor along with a contactor based voltage switching is adequate for real time implementation. The scheme is simple and economical in implementing and yields greater annual returns through increased energy harvesting. The proposed scheme is verified through MATLAB simulations on a 250 kW wind turbine and also through simulations and experiments on a 2.3 kW laboratory test bench using Altera cyclone II FPGA controller. Results demonstrate the feasibility of moderately efficient extended speed operation with the limited rating converters. This paper is organized as follows. Detailed mathematical modeling of a doubly fed induction machine including iron losses is described in section 2. Further the decoupled control of DFIG with RSC and GSC are explained in section 2. Maximum Power Point Tracking from the wind is explained in section 3. Section 4 describes the proposed scheme of extended low speed operation of DFIG. In the first case study, the simulation results of a 250 kW wind driven DFIG are presented and in the second case, the simulation and test results for a 2.3 kW laboratory DFIG are illustrated in section 5. 2. Mathematical modeling Fig. 1 shows the schematic diagram of DFIG system. Rotor Side Converter (RSC) is used to control active and reactive power of the generator and also to track the maximum power from the wind. The Grid Side Converter (GSC) is used to maintain DC link constant. Transformers are used to step down the voltage at stator side and rotor side. 2.1. Modeling of DFIM Fig. 2a shows the spatial distribution of stator voltage, stator flux and rotor currents in different reference frames. Stator voltage and stator flux are considered along q and d-axes in the synchronous reference frame respectively. Fig. 2b shows the per phase equivalent circuit of induction machine. The stator and rotor voltages in synchronous reference frame, can be expressed as e
ves ¼ Rs is þ e
ver ¼ Rr ir þ
d4es þ ju4es ; dt
(1)
d4er þ jðue um Þ4er ; dt
(2)
For higher rated machines the iron loss component can be neglected. But in smaller rating machines it cannot be neglected. So iron loss component is considered for modeling (laboratory machine). The iron loss component of the machine is modeled as a resistance in parallel to the mutual magnetizing inductance as shown in Fig. 2b. The resistance Rfe provokes an active current consumption (ife) but does not contribute to flux. The flux is created by the current flow through the magnetizing and leakage inductances. Hence, in the synchronous reference frame the stator and rotor fluxes are given by
V.R.R. Rudraraju et al. / Renewable Energy 81 (2015) 389e399
391
Fig. 1. Schematic diagram of DFIG.
e
e
e
e
4es ¼ Lls is þ Lm im ; 4er ¼ Lm is þ Lls ir ;
þ
e ir
¼
e ife
þ
e im
i*qr
(5)
This component is responsible for regulating the active power output of the machine. For regulating the reactive power input of the machine, the d-component of rotor current is controlled using the following relationship:
(6)
" # 2 Qs* i*dr ¼ þ idfe þ idm 3 vqs
Voltage across shunt branch is e
Rfe ife ¼ Lm
e
dim e þ jue im dt
" # 2 Ps* ¼ þ iqfe þ iqm 3 vqs
(4)
and applying KCL e is
The reference for q-axis rotor current is set based on (5) That is
(3)
(9)
(10)
2.2. Control of rotor side converter The q-axis current loop controls the active power flow, while the d-axis current loop regulates the flux by controlling the reactive power flow. The dynamic equations governing the rotor currents in the synchronous rotating frame are as follows:
didr 1 di ¼ vdr Rr iqr þ ðue um Þ Lm iqm þ Llr iqr Lm dm Llr dt dt
(7)
diqr diqm 1 ¼ vqr Rr iqr ðue um ÞðLm idm þ Llr idr Þ Lm Llr dt dt
(8)
The schematic block diagram for control of RSC is shown in Fig. 3a. The active loop (q-axis) controls the torque of the machine, while the reactive loop (d-axis) controls the flux of the machine.
2.3. Grid side converter control The equations that describe the dynamics of the GSC are summarized together as,
Lf Rf Lf Rf
!
!
vqf vqs d i þ iqf ¼ þ dt qf Rf Rf vdf d i þ idf ¼ þ dt df Rfe
! Lf ue idf Rf
! Lf ue iqf Rf
(11)
(12)
where, if and vf denote the current and voltage of GSC respectively. The dynamic equation representing the dc link voltage is given by
Fig. 2. DFIM modeling. (a) Spatial vectors in different reference frames, (b) Per phase equivalent circuit.
392
V.R.R. Rudraraju et al. / Renewable Energy 81 (2015) 389e399
Fig. 3. Control block diagram for (a) RSC, (b) GSC.
C
d ðv Þ ¼ dt dc
vqs i iL vdc qf
(13)
where, iL ¼ ðvqr iqr þ vdr idr =vdc Þ The control of GSC (Fig. 3b) contains two loops, one for active power and the other for reactive power control respectively. The active loop functions to keep dc-link voltage constant by controlling the active power flow in the circuit. The voltage controller output gives the current reference for q-axis current loop such that the there is an active power balance between ac side and dc side of the power circuit. Reference for the d-axis current loop is taken directly depending on the desired power factor. 3. Wind turbine
dum 1 ¼ ½Tm TL Bum J dt
(15)
where, J is the turbine inertia, B is friction coefficient, Tm is torque developed by the turbine, TL is the torque due to load and um is the rotor speed. The optimum power [18] from wind turbine is given by
Popt ¼ Kopt umopt
(16)
where, Kopt ¼ 0:5prCpmax R5 =l3opt
Power produced by a wind turbine is given by
Pm ¼ 0:5prCp R2 v3w
(14)
where, R is the radius of the turbine radius, r is air density, Cp is power coefficient. The dynamic equation of the wind turbine is given by
umopt ¼
lopt Vw R
(17)
umopt is the optimum rotor speed at which power for certain wind speed is maximum and lopt is the optimum tip speed ratio.
V.R.R. Rudraraju et al. / Renewable Energy 81 (2015) 389e399
4. Proposed control strategy for extending the speed range with fractional rating converters Commonly the operating slip range of DFIG systems [1e3] is from zero to 30%. To operate at higher slips (lower speeds), the converters need to be rated higher. In the proposed method, the lower end of speed range can be extended using the same fractional rated converters while not compromising on the efficiency. Applying a lower voltage to the stator at lower speeds has the effect of reducing the reactive power needed and also restricting the VA of the rotor converters. In addition, this reduces the losses in the generator also. The selection of the lower stator voltage to be applied is based on the choice of the minimum speed and the voltage (which is typically 30% of machine rating) and current ratings of the RSC. Further, the choice of threshold speed for voltage change over is based on the efficiency Vs. speed characteristics of the DFIG system and the voltage and current ratings of the RSC. The rotor voltage is given by
Vr ¼ sVs
(18)
Stator, rotor copper losses and iron losses are given by
Pcus ¼ 3Is2 Rs
(20)
Pcur ¼ 3Ir2 Rr
(21)
and 2 Pfe ¼ 3Ife Rfe
Ptotal ¼ Pcus þ Pcur þ Pfe
(23)
Stator and rotor active powers are given by
Ps ¼ vds ids þ vqs iqs
(24)
Pr ¼ vdr idr þ vqr iqr
(25)
The efficiency of the generator is given by
efficiency ¼
x ¼ 0:3*k 0:3
h¼
where k ¼ VNV/VLV Fig. 4 shows the flow chart for predetermination of generator performance for a given speed. The machine and turbine parameters are initialized and the stator voltage is selected. The wind speed is initialized and the corresponding optimum rotational speed is calculated from (17) and the maximum power is calculated using (16). Further, the machine losses and the efficiency are calculated. The generator performance predetermination is carried out till wind speed is 1 m/s.
(22)
Total losses are given by
As the stator voltage (Vs) is reduced at lower speeds, rotor voltage (Vr) also reduces (for same slip). Hence slip range can be extended till Vr reaches its limit (0.3 p.u) with lower stator voltage. Therefore the slip range is extended by
(19)
393
Stator active power þ rotor active power mechanical power
Ps þ Pr Pmech
(26)
5. Results and discussion Based on the analysis presented in the previous section, computer simulations are carried out to investigate the performance of a grid connected DFIG. Two cases are considered. One is a 250 kW machine and the other is a 2.3 kW laboratory machine. Higher rated wind turbines starting from 1 MW are in practice. However, for simulation study 250 kW rated machine is considered. Moreover,
Fig. 4. Flow chart for predetermination of generator performance for different wind speeds.
394
V.R.R. Rudraraju et al. / Renewable Energy 81 (2015) 389e399
the same proposed strategy can be implemented for wind turbine having higher power. Figs. 5e6 present the simulation results for 250 kW DFIG. 5.1. Case study-1: a 250 kW wind turbine The parameters of 250 kW wind turbine are given in appendix A. Three different stator line voltages are considered for study. These are; a) Low Voltage (LV ¼ 230 V or 0.577 p.u.), b) Medium Voltage (MV ¼ 300 V or 0.75 p.u.) and c) Nominal Voltage (NV ¼ 400 V or 1.00 p.u.). Fig. 5 shows the performance of DFIG system for different voltages. The maximum power from the wind is obtained using the MPPT algorithm [17]. The stator active power (Fig. 5a) is nearly the same for all voltages (NV, MV, or LV) except for a marginal difference. At a given speed the stator current with LV is higher than that with MV or NV since a higher stator current is required to maintain the same stator power output with a lower voltage (Fig. 5b). Fig. 5c shows the rotor voltage and current variation with speed. Considering the rotor voltage limit to be 30% of stator voltage, the boundary for the rotor voltage limit is
also shown. Since rotor voltage is proportional to slip times stator voltage, the rotor voltage for LV is less than that for MV and the rotor voltage for MV is less than that for NV. As seen from Fig. 5c using LV or MV at lower speeds results in lower VA for the power converters. For the 250 kW DFIG, the speed threshold for voltage change is found to be 0.72 p.u. The selection of the lower stator voltage to be applied is determined based on the choice of minimum speed (or maximum slip) and the voltage and current limits of the rotor circuit power converters. From Fig. 5c, it is notable that the voltage and current of RSC are within limits for LV case. Hence the LV (i.e., 0.577 p.u. or 230 V) is selected as the second voltage for stator beyond the slip of 0.3. Since the generator excitation (reactive power) is provided through the RSC, the rotor current is high at low speeds for NV (Fig. 5c and d). At lower speeds the LV connection gives higher efficiency compared to MV or NV (Fig. 6b) since the total copper loss with LV is less (Fig. 6a). It is also observed that the performance of DFIG system deteriorates below 0.6 p.u. speed (6 m/s wind speed) even with increased rating of converters (Fig. 6a). The vector diagrams for DFIG for NV and LV at 0.5 slip for the same mechanical power are
Fig. 5. Performance of DFIG system with different voltages. (a) Stator active power vs. rotor speed, (b) Stator current vs. rotor speed, (c) Rotor voltage, current vs. rotor speed and (d) Rotor Reactive power vs. speed.
V.R.R. Rudraraju et al. / Renewable Energy 81 (2015) 389e399
395
Fig. 6. Efficiency of the DFIG system and vector diagram with different voltages. (a) Total copper losses vs. rotor speed, (b) Efficiency vs. rotor speed, (c) Vector diagram for NV at 0.5 slip and (d) Vector diagram for LV at 0.5 slip.
shown in Fig. 6c and d respectively. With LV the rotor current is reduced significantly though the stator current rises slightly. Thus, with LV the total copper losses are reduced at lower speeds leading to efficiency improvement (iron losses are neglected for 250 kW
machine). Hence, switching to a lower stator voltage at lower speeds improves generator efficiency. The point of switching can be either at the instant when the rotor voltage exceeds the rating of RSC or when the efficiency decreases below a desired threshold.
Fig. 7. Comparison of energy production with conventional and proposed scheme. (a) Wind speed data for one year, (b) Energy harvested in one year.
396
V.R.R. Rudraraju et al. / Renewable Energy 81 (2015) 389e399
Table 1 Energy harvested for one year.
Conventional Proposed
Energy (MWh)
Increase in energy (MWh)
681.87 742.97
61.1
5.2. Annual energy harvested To evaluate the performance of proposed configuration, wind speed data registered in Tiruchirappalli (Tamil Nadu, India) from August 2013 to july 2014 is considered [29] (Fig. 7a). It is assumed that the generator is not in operation when the rotor voltage limits of RSC are violated. Energy is computed for the proposed and conventional configuration. The base value for energy is 2190 MWh. In the proposed scheme for wind speeds between 4.8 and 6.6 m/s DFIG is operated with LV and above 6.6 m/s the DFIG is
operated with NV. Whereas, in conventional scheme the DFIG is operated with NV above 6.6 m/s. Fig. 7b and Table 1 show the details of energy harvested (p.u.) in one year. Thus with the 250 kW generator the energy harvested over one year is increased by 61.1 MWh with the proposed scheme. This is an increase of about 2.8% with a single generator. Thus the proposed scheme has the potential to increase the energy output significantly if implemented with more number of generators, such as all those connected in one wind farm through common switchgear and transformer. 5.3. Experimentation: case study-2 Experimentation is carried out on 2.3 kW slip ring induction machine for the verification of the proposed scheme. Fig. 8a shows the schematic diagram of the experimental setup. The machine parameters are given in appendix A. A 5-hp DC machine is used to
Fig. 8. (a) Schematic diagram of experimental setup, (b) Laboratory setup.
V.R.R. Rudraraju et al. / Renewable Energy 81 (2015) 389e399
397
Fig. 9. Performance of DFIG system with different voltages. (a) Mechanical power vs. wind velocity, (b) Stator real power vs. rotor speed, (c) Stator current vs. rotor speed and (d) Rotor current vs. rotor speed.
simulate the wind turbine characteristics using armature controlled dc drive. Altera cyclone II FPGA is used to generate firing pulses for RSC and GSC. To verify the increase in efficiency at lower speeds with limited rated converters the simulation and experimental tests are carried out for two different cases of stator line voltage a) Low Voltage (LV ¼ 240 V) and b) Nominal Voltage (NV ¼ 415 V). The rotor power circuit consists of two power converters (MD B6CI 800/415-30 F) of Semikron make. The Rotor side PWM inverter is controlled using a frequency of 5 kHz. The rotor position is found using sensorless position estimation. Photograph of laboratory setup is shown in Fig. 8b. Fig. 9a shows the mechanical power Vs. wind speed with two different voltages. Since the wind turbine mechanical characteristics considered are same for both NV and LV, the predicted characteristics for NV and LV overlap. It is notable that the experimental results are close to predicted results. The implementation of MPPT in the experimental set up enables tracking the maximum power, as indicated by the markers in Fig. 9a. Fig. 9b shows the variation in stator real power as speed of machine is varied for NV and LV settings. The power with LV is higher compared to NV at lower speeds. The stator reactive power reference is set to zero. Fig. 9c and d show the stator and rotor currents respectively as rotor speed is varied. Stator current is higher for LV for the same stator power.
With NV, since the magnetizing current is high at lower speeds, the rotor current is higher till 1375 rpm and above 1375 rpm the rotor current is higher with LV. Since, the efficiency below 775 rpm is nearly zero, only speeds higher than 775 rpm are considered. Fig. 10a shows the rotor voltage for LV and NV and the rotor voltage limit is indicated. The rotor voltage is high at higher slip. Since magnetizing current is supplied from rotor side, rotor reactive power is less with LV compared to that with NV as shown in Fig. 10b. Since reactive power is less with LV, rotor side VA is also less and is shown in Fig. 10c. Up to 1722 rpm, the efficiency is higher for LV and above 1722 rpm efficiency is high for NV. The mechanical power corresponding to 1722 rpm is 3138 W. It is seen that the efficiency at lower speeds is high for LV, this is due to the reduction in magnetizing current, rotor current and total losses (Fig. 10d). Table 2 shows the performance comparison of the DFIG system with different stator voltages. Data in bold italics correspond to efficiency above 0.9, rotor voltage & VA below 0.3 p.u. It is observed that with 0.3 p.u. rotor side converter rating, the speed is extended down to 0.5 p.u. with efficiency greater than 0.9. For speeds below 0.7 p.u., when the stator voltage is switched to LV the rotor voltage does not exceed the RSC rating of 0.3 p.u. Thus, with the proposed method the speed range is extended by 0.2 p.u. without violating the rating of the rotor side converter.
398
V.R.R. Rudraraju et al. / Renewable Energy 81 (2015) 389e399
Fig. 10. Performance of DFIG system with different voltages. (a) Rotor voltage vs. rotor speed, (b) Rotor active power vs. rotor speed, (c) Rotor reactive power vs. rotor speed and (d) Efficiency vs. rotor speed.
increased by 61.1 MWh with the proposed scheme. Overall, the results demonstrate the viability of the scheme.
6. Conclusions The study demonstrates extended low speed operation of a grid connected DFIG without exceeding the ratings of the rotor side power converters. The generator efficiency is shown to be maintained with the same limited rating converters (rated 30% of the generator). The laboratory machine is found to be less efficient (due to high stator and rotor resistances), being much smaller in rating. For the 250 kW DFIG, low speed operation down to 0.5 p.u. can be achieved with 30% rated power converters at generator efficiencies better than 90%. Also the energy harvested over one year is
Table 2 Performance comparison of the DFIG with different voltages. Case study
1
Speed (p.u.)
0.5 0.6 0.7
Rotor VA (p.u.)
Efficiency
Rotor voltage (p.u.)
NV
LV
NV
LV
NV
LV
0.86 0.91 0.93
0.91 0.93 0.92
0.51 0.42 0.34
0.31 0.25 0.20
0.25 0.22 0.17
0.11 0.11 0.10
Appendix A
Parameters Machine rating Stator voltage (V) Stator current (A) Rotor current (A) Stator resistance (U) Rotor resistance (U) Iron loss resistance (U) Stator inductance (mH) Rotor inductance (mH) Stator turns/rotor turns
250 kW 400 370 370 0.02 0.02 e 4.4 4.4 1
Power Converters: 415 V, 30 A SEMIKRON Inverter. DC-Link Voltage and Capacitor: 800 V, 2000 mF. DC Machine: 5hp, 1500 r/min. Armature: 220 V, 19 A; Field: 220 V, 1 A.
2.3 kW 415 4.7 7.5 3.678 5.26 270 0.30682 0.30682 1.92
V.R.R. Rudraraju et al. / Renewable Energy 81 (2015) 389e399
References [1] Muller S, Deicke M, De Doncker RW. Doubly fed induction generator systems for wind turbines. IEEE Ind Appl Mag 2002;8(3):26e33. [2] Datta R, Ranganathan VT. Variable-speed wind power generation using doublyfed wound rotor induction machine-a comparison with alternative schemes. IEEE Trans Energy Convers 2002;17(3):414e21. [3] Zou Yu, Elbuluk ME, Sozer Y. Simulation comparisons and implementation of induction generator wind power systems. IEEE Trans Ind Appl 2013;49(30): 1119e28. [4] Li H, Chen Z. Overview of different wind generator systems and their comparisons. IET Renew Power Gener 2008;2(2):123e38. [5] Mittal Rajveer, Sandhu KS, Jain DK. An overview of some important issues related to wind energy conversion system (WECS). Int J Environ Sci Devel 2010;1(4):351e63. [6] Cadirci I, Ermis M. Double-output induction generator operating at subsynchronous and supersynchronous speeds: steady-state performance optimization and wind-energy recovery. IEE Proc Elect Power Appl 1992;139(5): 429e42. [7] Alkandari Ahmad M, Soliman Soliman Abd-Elhady, Abdel-Rahman Mansour H. Steady state analysis of a doubly fed induction generator. Energy Power Eng 2011;3:393e400. [8] Sediki H, Ould Abdeslam Dj, Otmane-cherif T, Bechouche A, Mesbah K. Steadystate analysis and control of double fed induction motor. World Acad Sci Eng Tech 2012;61:153e61. [9] Vicatos MS, Tegopoulos JA. Steady state analysis of doubly- fed induction generator under synchronous operation. IEEE Trans Energy Convers 1989;4(3):495e501. [10] Miller TJE. Theory of the doubly-fed induction machine in the steady state. In: Int. Conf. on Elect. Machines, Rome; 2010. p. 1e6. [11] Cardenas R, Pena R, Alepuz S, Asher G. Overview of control systems for the operation of DFIGs in wind energy applications. IEEE Trans Ind Elect 2013;60(7):2776e98. [12] Pena R, Clare JC, Asher GM. Doubly fed induction generator using back-toback PWM converters and its application to variable speed wind-energy generation. IEE Proc Elect Power Appl 1996;143(3):231e41. [13] Hopfensperger B, Atkinson DJ, Lakin RA. Stator-flux-oriented control of a doubly-fed induction machine with and without position encoder. IEE Proc Elect Power Appl 2000;147(43):241e50. [14] Datta R, Ranganathan VT. Direct power control of grid-connected wound rotor induction machine without rotor position sensors. IEEE Trans Power Elect 2001;16(3):390e9. [15] Amiri N, Madani SM, Lipo TA, Zarchi HA. An improved direct decoupled power control of doubly fed induction machine without rotor position sensor and
[16] [17]
[18]
[19]
[20]
[21]
[22]
[23]
[24] [25]
[26]
[27] [28]
[29] [30]
399
with robustness to parameter variation. IEEE Trans Energy Convers 2012;27(4):873e84. Qu Liyan, Qiao Wei. Constant power control of DFIG wind turbines with supercapacitor energy storage. IEEE Trans Ind Appl 2011;47(1):359e67. Zou Yu, Elbuluk ME, Sozer Y. Stability analysis of maximum power point tracking (MPPT) method in wind power systems. IEEE Trans Ind Appl 2013;49(3):1129e36. Zhao Y, Zou XD, Xu YN, Kang Y, Chen J. Maximal power Point tracking under speedmode control for wind energy generation system with doubly fed introduction generator. In: IEEE Int. Conf. Power Elect. and Motion Control1; 2006. p. 1e5. Chen Bo-An, Lu Tien-Kei, Hsu Yuan-Yih, Chen Woei-Luen, Lee Zheng-Cong. An analytical approach to maximum power tracking and loss minimization of a doubly fed induction generator considering core loss. IEEE Trans Energy Convers 2012;27(2):449e56. Ferreira FJTE, de Almeida AT, Baoming Ge, Faria SP, Marques JM. Automatic change of the stator-winding connection of variable-load three-phase induction motors to improve the efficiency and power factor. In: IEEE Int. Conf. Ind. Tech; 2005. p. 1331e6. Cistelecan MV, Ferreira FJTE, Popescu M. Adjustable flux three-phase ac machines with combined multiple-step staredelta winding connections. IEEE Trans Energy Convers 2010;25(2):348e55. Lei Yi, Zhao Zhengming, Wang Shuping, Dorrell DG, Xu Wei. Design and analysis of staredelta hybrid windings for high-voltage induction motors. IEEE Trans Ind Elect 2011;58(9):3758e67. Osama M, Lipo TA. Modeling and analysis of a wide-speed-range induction motor drive based on electronic pole changing. IEEE Trans Ind Appl 1997;33(5):1177e84. Bird BM, Ridge J. Induction-motor speed-control schemes employing statorvoltage variation and pole-change windings. In: PROC. IEE; 1971. p. 537e42. Djurovic M, Joksimovic G, Saveljic R, Maricic I. Double fed induction generator with two pair of poles. In: Int. Conf. Electrical Machines and Drives; Sept. 1995. p. 449e52. Park JW, Lee KW, Lee HJ. Control of active power in a doubly-fed induction generator taking into account the rotor side apparent power. In: Power Electronics Specialists Conf., 3; 2004. p. 2060e4. Smith AC, Todd R, Barnes M, Tavner PJ. Improved energy conversion for doubly fed wind generators. IEEE Trans Ind Appl 2006;42(6):1421e8. Muljadi E, Singh M, Gevorgian V. Doubly fed induction generator in an offshore wind power plant operated at rated V/Hz. IEEE Trans Ind Appl 2013;49(5):2197e205. Wind data, http://weatherspark.com/. Ammasaigounden N, Subbiah M, Krishnamurthy MR. Wind-driven selfexcited pole-changing induction generators. IEE Proc B Electr Power Appl 1986;133(5):315e21.