Characterization of Measured Voltage Dips in Wind Farms in the Light of the New Grid Codes

Emilio G omez-L azaro and Miguel Ca nas

Renewable Energy Research Institute Dept. of Electrical, Electronic and Control Eng. EPSA Universidad de Castilla-La Mancha 02071 Albacete (SPAIN) Email: emilio.gomez@uclm.es, miguel.canas@uclm.es

Juan Alvaro Fuentes and Angel Molina-Garc a

Dept. of Electrical Eng. Universidad Polit ecnica de Cartagena 30202 Cartagena (SPAIN) Email: juanalvaro.fuentes@upct.es, angel.molina@upct.es

Abstract Voltage dips are a major concern for some transmission system operators. With increasing wind farm penetration in power systems, many national grid codes are imposing requirements for uninterrupted generation throughout power system disturbances. The aim of this paper is to present the results of measured voltage dips using two power quality analyzers installed for months in a Spanish wind farm. From the measured voltage waveforms, voltage dips are characterized by extracting dip type and characteristics.

I. I NTRODUCTION Voltage dips are short duration reductions in rms voltage (between 10% and 90%) of the voltage at a point in the electrical system, which lasts for half a cycle to 1 min, [1]. The duration of voltage dips is described as the total time interval between the point on wave of sag initiation and recovery, [2]. They are caused by faults in the electric supply system and the starting of large loads, such as motors. Voltage dips are generally considered a power quality problem of equal importance as long and short interruptions in the supply. The interests in voltage dips are increasing because they cause the detrimental effects on several sensitive equipments such as adjustable-speed drives, process-control equipments, and computers. Such equipments can be tripped when the voltage drops below 90% of the rated voltage for longer than a few cycles, [3], [4]. In accordance with the practice in wind farms, wind turbines are disconnected from the grid when the terminal voltage fell below 80-90%, [5], [6]. Even relays and contactors in motor starters can be sensitive to voltage dips, resulting in shutdown of a process when they drop out, [7]. Voltage dips are mainly due to short-circuits and earth faults in the grid, [8]. These faults in the power system, even far away from the location of the wind farm, or any other power installation, can generate a voltage dip at the connection point of the wind turbines. Factors governing the magnitude and duration of voltage dips include the fault impedance and location, the conguration of the power network, and the system protective relay design and operation, [7]. This last aspect is important since voltage dip condition lasts until the fault is cleared by a protective device. Solutions to the voltage

dip effects must be implemented in the customer facility, since although it is possible for the utility to reduce the number of faults through design practices and specic equipment, it is impossible to avoid faults on the power system, [7]. Therefore, in wind farms, the solution must be installed at the wind farm interconnection point with the electrical grid or at the wind turbine level. The installed capacity of wind power generation has grown very fast in the past years, increasing dramatically the level of wind power generation into existing utilities network. On the other hand, the rating of large wind farms reaches rating of conventional generating units and this development has carried out to requirements of how to connect wind farms to the grid. Until now, these grid codes specied by the transmission system operators mainly dealed with how a wind farm should operate in steady state while requirements recently imposed in some countries dealed with how a wind installation response to grid faults must be addressed, [9]. These are commonly referred to as the fault ride-through specications. Specically, national grid codes are requiring uninterrupted generation throughout power system disturbances supporting the network voltage and frequency, and therefore, extending characteristics such as low voltage ride through, or reactive and active power capabilities, [9], [10]. In [9][13] grid connection requirements in Spain, Denmark, Germany, Ireland, Sweden and Scotland are studied. When the wind power penetration level is high, the protective disconnection of a large amount of wind power is an unacceptable consequence that may threaten the power system stability. This is due to conventional power plants will be not be able to support the voltage and the frequency of the grid during and immediately following the grid failure. In Spanish case, REE the transmission system operator Red El ectrica de Espa na grid code, recently approved, species that the wind farm must support voltage dips, at the point of interconnection with the transmission network, without tripping. The voltage-time curve that limits the magnitude and duration of the voltage dips, produced by single-phase-toground, two-phase-to-ground and three-phase short-circuits, is shown in gure 1. For non-earthed two-phase short-circuits,

Voltage (pu) 1 0.8

Beginning of perturbation 0.95 pu

Fault clearance 0.2 Duration of fault 0

Fig. 1.

0.5

15

Time (sec.)

Voltage-time curve that the generation facility must support

the voltage limit is chosen at 0.6 p.u. instead of 0.2 p.u. II. VOLTAGE DIP CHARACTERIZATION Voltage-dip characterization concerns the quantication of voltage-dip events through a limited number of parameters, [14]. Most methods for voltage dip characterization use two parameters to quantify the severity of a voltage dip magnitude (or remaining voltage) and duration, [3], [15] [17]. However, voltage dips can be far more complicated than this type of characterization can show: Usually, methods are based on the lowest remaining voltage and the longest duration of all the three voltages. This causes some problems. For example, a voltage drop in one phase is not distinguished from other in three phases, or voltage dips due to an earth fault in a highimpedance grounded network will be seen as equally severe or even more severe than the dip due to a short-circuit fault, [15]. Therefore, it is an appropriate approximation for balanced dips, whereas the majority of dips are unbalanced, [16]. Dip depth and phase-angle jump information can be required along with start and end times, [15]. Therefore, some authors are researching other methods to characterize unbalanced dips, being based on measurements or applying the basic circuit theory in the faults, [14]. It is usually assumed that the voltage prole during voltage dip is rectangular, failing in the characterization of non-rectangular dip, overestimating it, [3], [14]. This can be important to many industrial customers with large induction motor loads or in the case or wind farms. This methodology is not adequate to multistage voltage dips events, in which the fault can evolve to a different type. Basically, the input data to the characterization methods can be tted in: Monitoring the lowest remaining voltage and the longest duration of all the three voltages.

Monitoring all the voltages waveforms with an adequate sample rate. Vd2 + Vq2 in a vector controller, [4]. Monitoring of This method is derived from space vector control usually employed in the control of induction machinery. The three-phase voltages are converted into one phasor with two orthogonal components and synchronous reference frame, which is locked via a phase-locked loop PLL In [15] two methods to obtain three-phase voltage dip characterization ABC classication and symmetrical components classication were exposed and compared. It is concluded that ABC classication due to its simplicity is more used than the symmetrical components, being also more intuitive and giving a good approximation about the evolution of the dips along the different voltage levels of the network. The ABC classication should not be viewed as a different classication, since it can be considered as a special case of the symmetrical components classication. However, this classication is based on a simplied model of the network, being not recommended in [15] for the classication of voltage dips obtained from measured instantaneous voltages. ABC classication considers seven types of voltage dips, by dening complex voltages and phasor diagrams. In [18], theoretical relations between the minimum phaseto-neutral voltage VP N and the minimum phase-to-phase voltage VP P are presented for the seven types of dips, dened in [15]: Type A: all phases experience the same retained voltage and phase-angle jump.

VP P = VP N

(1)

Type B: It is not common because it is seen only when a line to ground fault occurs at the same voltage level or at a location connected by star-star transformer grounded at both sides. VP P =
1 2

+ VP N 3

3 4

(2)

Type C: It is a reduction of the voltage in two phases. It is caused by a line to line fault or by a propagation of dip type B through a delta-star connected transformer. 4 2 1 VP N (3) 3 3 Type D: It is caused by a propagation of a type C dip through a delta-star windings connected transformer. It is a voltage drop in one phase.
2 VP P =

1 3 2 + V (4) 4 4 PN Type E: It shows a symmetrical relation between PP and PN voltage like type A. This dip is rare as type B by the same reasons. Type F: It is a reduction of the voltage in one phase, caused by the propagation of a line to line to ground
2 VP P =

GPRS Network GPRS Ethernet Modem

Internet

GPS Time Synchronization

Wifi Router

Wifi Router

Power quality analyzer

Power quality analyzer

Transformer 690/20 kV Dfig

Transformer 20kV/66 kV

Crowbar

Power converter

Fig. 2.

Power quality analyzer installation scheme

fault through a delta-star connected transformer. 1 2 1 1 2 3VP (5) P = (2 + )VP N + VP N + 3 3 3 Type G: These dips are obtained from the propagation of a dip type F through a delta-star connected transformer. VP P = 0.0707 +

III. R ESULTS Two power quality analyzers fullling IEC 61000-4-30 class A accuracy, frequency synchronization, and absolute time requirements have been installed in a Spanish wind farm. These analyzers, with a 10MHz sample rate, are useful to capture detailed voltage and current waveforms during the voltage dip and the clearance of the fault, together with powerful trigger options to obtain the entire transient. Figure 2 presents the power quality analyzer installation scheme, showing a power quality analyzer connected in the nacelle Gamesa G90 2,0 MW between the doubly fed induction

2 0.327 3.112 VP N (6) 1.556 In this classication, voltage dip types D and F C and G, are similar, making difcult their distinction from the measurements without knowing the fault types that caused them, [19].

generator and the 690/20.000V power transformer. This power quality analyzer measures three stator voltages and currents, together with a rotor line current and the DC bus voltage in the rotor power converter. The other power quality analyzer measures line voltages and currents at the wind farm electrical substation. Both power quality analyzers are linked via wi to a UMTS/GPRS modem, allowing the remote accessing to the power quality analyzer conguration and their recorded data. Figures 3 and 4 show the three-phase voltage waveforms, the amplitude and the angle of the voltage space vector, [9], of two measured voltage dips. The voltage space vector is a complex magnitude obtained by combining the threephase instantaneous voltages va , vb and vc , representing therefore the voltage dip uniquely through an instantaneous (t). Once this complex voltage is given, complex vector, V the three instantaneous phase voltages can be obtained by the 2 (t) along the three axis: ej 0 , ej 23 projection of V and ej 3 . These two voltage dips together with others measured for months in a Spanish wind farm have been characterized according to the methodology explained in section II, [18]. The characteristics of voltage dips at these equipment terminals vary according to the transformer winding connections where they are located. So, in general, the difference in voltage dips performance magnitude and phase angle is a result of the propagation of voltage dips from the fault location to the terminals of wind farms and wind turbines, taking into account the transformers and the lines located between these two electrical points, [19]. In [2] the importance of the winding transformer connections have been deeply studied, summarizing that the Yd transformer winding connection has the strongest inuence, reducing the number of severe voltage sags and increasing the number of medium voltage sags. On the other hand, the Yd1, Yd11, or Yd5 transformers maintain the voltage dip magnitude, but the associated phase shifts are completely different from each other. Therefore, transformer winding connections in the substation and wind turbines must be taken into account, being usual a Dy11 transformer in the wind turbine 690/20000 V and Yd11 transformer in the wind farm electrical substation 20/66 kV, as it is shown in the gure 2. Figure 5 shows the classication of measured voltage dip according to the characterization established in [18]. It can be seen that voltage dips are undoubtedly classied. Moreover, real data measured results are not over the theoretical curves on the contrary of simulation based results but in any way very near to them. This is due to factors such as fault impedance neglected in the theoretical curves and simulations and load effects. On the other hand, this characterization and classication method presents some disadvantages such as, obviously it is not adequate to multistage voltage dips, in which the fault and therefore, the voltage dip itself evolves to a different type. Moreover, during the evolution in time of the fault, the voltage dip can not be correctly characterized, since it is based on the theoretical relationship between the minimum phase-to-

1.5

0.5 Voltage (pu)

0.5

1.5 0 0.1 0.2 0.3 0.4 Time (s) 0.5 0.6 0.7 0.8

(a) Voltage waveforms

1.1

1.05

1 Voltage (pu)

0.95

0.9

0.85

0.8

0.75 0 0.1 0.2 0.3 0.4 Time (s) 0.5 0.6 0.7 0.8

(b) Space vector voltage (amplitude)

Voltage space vector angle 200

150

100

50 Angle[Deg]

50

100

150

200 0.7

0.8

0.9

1.1 Time [s]

1.2

1.3

1.4

1.5

(c) Space vector voltage (angle) Fig. 3. Example of voltage dip measured

1.5

0.5 Voltage [(pu)

neutral voltage and the minimum phase-to-phase voltage. In the application of this method to measured voltage dips, as the presented here, other considerations must be taken into account, as the selected point reects the voltage dip type. So, minimum density of measured data around zones must be established, and only these zones accessed for consideration in the classication of dips. Finally, as it was commented above, it is usually assumed that the voltage prole during voltage dip is rectangular, failing in the characterization of non-rectangular dip, overestimating it. IV. C ONCLUSIONS

0. 5

1. 5 0 0.5 Time (s) 1 1.5

(a) Voltage waveforms

1.06 1.04 1.02 1 Voltage (pu) 0.98 0.96 0.94 0.92 0.9 0.88 0 0.1 0.2 0.3 0.4 Time (s) 0.5 0.6 0.7 0.8

Voltage dips are a major concern in wind energy industry, due to the imposition by some national grid codes of requiring uninterrupted generation throughout power system disturbances, such as the voltage dips. Two power quality analyzers have been installed in a wind farm, being the rst one located inside a multi-megawatt wind turbine, and the second one in the wind farm substation. Voltage dip records from instantaneous voltages measured using the power quality trigger abilities have been characterized taking into account their waveforms along the dip. Basically, the presented method previously applied by its authors to simulated voltage dips obtains a value for each voltage dip according to the relationship between the minimum phase-to-neutral and the minimum phase-to-phase voltages, being classied according to the distance of that point to six theoretical curves. Disadvantages of the used method have also been highlighted, such as being this point representative of the voltage dip without taking into account measurement errors, for example or the errors in the characterization of multistage voltage dips. ACKNOWLEDGMENT The nancial support provided by the Ministerio de Educaci on y Ciencia ENE2006-15422-C02-01/ALT and ENE2006-15422-C02-02/ALT is gratefully acknowledged. The authors also would like to thank to Mr. Juan Manuel Abell an from Dea y Energ as Renovables, the Moralejo wind farm people located in Alpera, Albacete (Spain) and Gamesa technicians for their support in the measurements. R EFERENCES

(b) Space vector voltage (amplitude)

200

150

100

50 Angle (deg)

50

100

150

200 0.5

0.6

0.7

0.8

0.9 Time (s)

1.1

1.2

1.3

(c) Space vector voltage (angle) Fig. 4. Example of voltage dip

[1] Recommended practices and requirements for harmonic control in electrical power systems, IEEE Standard 519-1992. [2] M. T. Aung and J. V. Milanovic, The inuence of transformer winding connections on the propagation of voltage sags, IEEE Transactions on Power Delivery, vol. 21, no. 1, pp. 262269, January 2006. [3] D. J. Won, S. J. Ahn, and S. I. Moon, A modied sag characterization using voltage tolerance curve for power quality diagnosis, IEEE Transactions on Power Delivery, vol. 20, no. 4, pp. 26382643, October 2005. [4] C. Fitzer, M. Barnes, and P. Green, Voltage sag detection technique for a dynamic voltage restorer, IEEE Transactions on Industry Applications, vol. 40, no. 1, pp. 203212, January/February 2004. [5] I. Erlich and U. Bachmann, Grid code requirements concerning connection and operation of wind turbines in germany, in IEEE Power Engineering Society General Meeting, San Francisco, California, June 2005.

0.9

0.8

0.7

0.6

PP voltage (pu)

0.5

0.4
F

0.3

0.2 A,E 0.1

G C

0 0

0.1

0.2

0.3

0.4 0.5 PN voltage (pu)

0.6

0.7

0.8

0.9

Fig. 5.

Relation between phase-to-phase and phase-to-neutral measured dip magnitudes

[6] A. Dahlgren, C. Klippel, and A. Juntti, Recorded fault ride through capability for two types of wind turbine generators, in Nordic wind power conference, Espoo, Finland, May 2006. [7] M. F. McGranaghan, D. R. Mueller, and M. J. Samotyj, Voltage sags in industrial systems, IEEE Transactions on Industry Applications, vol. 29, no. 2, pp. 391403, 1993. [8] M. H. J. Bollen, G. Olguin, and M. Martins, Voltage dips at the terminals of wind power installations, Wind Energy, vol. 8, pp. 307 318, 2005. [9] E. G omez, J. A. Fuentes, A. Molina, F. Ruz, and F. Jim enez, Results using different reactive power denitions for wind turbines submitted to voltage dips: Application to the spanish grid code, in 2006 IEEE Power Systems Conference and Expo, October-November 2006. [10] E. G omez, J. Fuentes, A. Molina, F. Ruz, and F. Jim enez, Field tests of wind turbines submitted to real voltage dips under the new spanish grid code requirements, Wind Energy, 2007, accepted for publication. [11] C. Jauch, J. Matevosyan, T. Ackermann, and S. Bolik, International comparison of requirements for connection of wind turbines to power systems, Wind Energy, vol. 8, pp. 295306, 2005. [12] C. Jauch, P. Sorensen, and B. Bak-Jensen, International review of grid connection requirements for wind turbines, in Nordic wind power conference, Goteborg, March 2004. [13] J. Matevosyan, T. Ackermann, S. Bolik, and L. Sder, Comparison of international regulations for connection of wind turbines to the network, in Nordic wind power conference, Goteborg, March 2004. [14] M. H. J. Bollen, Algorithms for characterizing measured three-phase unbalanced voltage dips, IEEE Transactions on Power Delivery, vol. 18, no. 3, pp. 937944, July 2003. [15] M. H. J. Bollen and L. D. Zhang, Different methods for classication of three-phase unbalanced voltage dips due to faults, Electric Power Systems Research, vol. 66, pp. 5969, 2003.

[16] L. Zhang and M. H. J. Bollen, Characteristic of voltage dips (sags) in power systems, IEEE Transactions on Power Delivery, vol. 15, no. 2, pp. 827832, April 2000. [17] R. Leborgne and D. Karlsson, Phasor based voltagge sag monitoring and characterisation, in International Conference on Electricity Distribution, June 2005. [18] R. Leborgne, G. Olguin, and M. H. J. Bollen, The inuence of pqmonitor connection on voltage dip measurements, in IEE MedPower Conference, 2004. [19] S. Z. Djokic, J. V. Milanovic, D. J. Chapman, and M. F. McGranaghan, Shortfalls of existing methods for classication and presentation of voltage reduction events, IEEE Transactions on Power Delivery, vol. 20, no. 2, April 2005.