Profile of voltage in grids with wind farms

Joaqun Mur ngel A. Bayod Jess Salln University of Zaragoza CIRCE Foundation Electrical Engineering Department

EWEC 2003 (European Wind Energy Conference & Exhibition)

Probability Density Function

to now, distributed generation in Spain must inject power with unity power factor. But directive is going to change and some feasible regulations for reactive power are studied, attending specially to voltages across the grid. The range of reactive power depends on generating technology and the wind parameters. The proposed method can help to evaluate the affection to the profile of voltage and the availability of reactive power injection or absortion. I. INTRODUCTION Distribution networks are usually limited for the limits in voltage variation (specially at the end of the line). Therefore, distribution companies prefer distributed generation to be connected to higher voltage levels, where its impact in voltage profile is smaller. If reactive power is available, voltage fluctuations can be limited with lower cost. In addition, wind energy could reduce the losses in the grid if the zone has generation shortage (of active or reactive power). II.
WIND FARM MODEL

AbstractUp

Grids Thevenin equivalent

Farm equivalent transmisin matrix

Power Output Hp.u.L

the medium voltage network that connects the substation to the wind turbines. the wind turbines (including the MV to LV transformer)
ZSC grid -Igrid PCC
PCC, point of common coupling

Simplified power curve, defined by output of 25 % and 75% of nominal power

Power 1 curve of the farm or turbina

0.8

0.6

Ugrid PCC

Uturbine

(average)

+ U0

A C

B D Farm

-Iturbine (average) Pturbine Qturbine

0.4

P=

Pnominal 2

0.2

0 0 5

CDF@ActivePower < pD = 2 p yShapey v25 + v75 Hv25 - v75L 1- ^i- i + ArcTanhA1 E Log@3D Scale Pn { k k 2Scale {
10 Wind 15 Speed
H m s L

w (w 25% + w 75% ) / 2 1 Tanh Ln(3) w 25% w 75%

20

25

30

Figure 1: Model of the farm with fixed tap transformer using its transmission matrix. Final representation of the Wind Farm

III. DISTRIBUTION OF ACTIVE POWER OUTPUT

(computed using Weibull Distribution and power curve )
Probability 2 1.75 1.5 1.25 1 0.75 0.5 0.25 0 0.2 0.4 Power output
H p.u .L

Density

Farm with tap-changing transformer Substation voltage at the MV side of the transformer is almost the commanded value. A.1 Added model of turbines using the model of moment
a)
Substation Usubstation Distance substation 1st turbine Distance 1 st turbine - last Isubstation Nturb Distance substation 1st turbine Dist. 1 st turbine - last 3 Equivalent model with the turbines concentrated at 1/3 distance from the first one Nturb

Solid Red -> mean wind speed = 5 Pink -> mean wind speed = 5,5 Dark blue -> mean wind speed = 6 Light blue -> mean wind speed = 6,5 Green -> mean wind speed = 7

0.6

0.8

Distribution of active power output of wind farms (real data)

H istograma de las potencias 3500 3000

b)
Substation Usubstation

Isubstation

N de horas en el intervalo

Figure 2: Concentrated model of a MV circuit in a park.

2500

2000

A.2 Model of the MV circuit branches

Substation Averaged model of MV net I substation MV A net MV

1500

1000

In this model, the farm is divided into the following parts: substation (including the parks substation and the portion of the line that goes to the PCC)

Cnet MV

Bnet MT 1 0 . Dnet MT 0 N

I generator average Ugenerator average Pturbine Qturbine

5 0 0

Figure 3: Simplified scheme of the medium voltage network of the park with added generator model.

Usubstation

0 -0.2

0.2

0.4 0.6 0.8 Potencia aparente en p.u.

1.2

1.4

EWEC 2003 (European Wind Energy Conference & Exhibition)

IV.

DISTRIBUTION OF VOLTAGE

C2 =

"

4 H1 + k2L U2 HUo2 - U2 Abs@AD2L Abs@BD2 - U4 HB H + kL Conjugate@ AD - A H- + kL Conjugate@ BDL2

Active power corresponding to a certain voltage C3 - C2 C2 + C3 ; P2 = ; 2 H1 + k2L Abs@BD2 2 H1 + k2L Abs@ BD2 CDF@ U < uD = 1+ CDF@ActivePower < P1D - CDF@ActivePower < P2D;

C3 = U2 HHB - B kL Conjugate@AD + A H1 + kL Conjugate@BDL

P1 =

In the analyzed cases, power and voltage are linearly related by short circuit impedance:

Reactive power can be controlled in a range, depending on the wind. Some turbines can generate reactive power even when they are still. The ability to generate reactive power is best characterized using the statistical distribution of the feasible reactive output. The objective of the control can be stabilize voltage or diminish losses. The availability must be computed taking into account if reactive generation is possible at stand-by, possible voltage limit violations.
FQmax@Qturbine qD = FP@P mn p < P mxD = FP@p < P mxD - FP@p PmnD

Availability p.u.

Cumulative Distribution Function of voltage can be computed through Cumulative Distribution function of power and supposing a relationship between active and reactive power. For example, if power factor is fixed, Q = k P, where k = tan()

Note: Near collapse, voltage decreases quickly and this equation is no longer valid, but this is not the case of wind farms in Spain (in normal operation), due to the limit:

Reactive limited by apparent power.

CDF@Q < qD = CDF@P < P1D - CDF@ w < wcutinD
Availability 1 0.8 0.6 0.4 0.2 of Reactive Power Regulation

Pshort circuit > 20 Pwind farm

MinAPcutout, MaxAPcutin,

Sn2 - Q2 EE

V.

AVAILABILITY OF REACTIVE POWER

REGULATION

-1

-0.5 0 0.5 Reactive Power Q

U Uo +

PR + QX Uo

Therefore, voltage and power distribution have almost the same shape (minor influence in voltage when k = -Rsc / Xsc). Study case: Simple impedance connecting the farm to the network (sc=60, Zsc=1/20 p.u.)
Probability Voltage
< v

Induction generator with capacitor banks Reactive power in a farm with induction generators follow approximately a quadratic relationship. Q Qcond - Q0 - Qmgen - b S - cS2
Availability 1 of Reactive Power Regulation

Availability p.u.

Cumulative Probability U < u

1 0.8 0.6 0.4 0.2

Reactive limited by maximum capacitive and inductive power factor.

Pind = Min@Pcutout, Max@Pcutin, -Q kindDD Pcap = Min@Pcutout, Max@Pcutin, Q kcapDD CDF@Q < qD = If@Q > 0, 1 - CDF@P < PcapD, 1 - CDF@P < PindDD
Availability 1 of Reactive Power Regulation

0.8 0.6 0.4 0.2

Reactive generation with the generator connected.

Total availavility Capacitor banks connected with generator disconnected

Solid Red -> mean wind speed = 5 Pink -> mean wind speed = 5,5 Dark blue -> mean wind speed = 6 Light blue -> mean wind speed = 6,5 Green -> mean wind speed = 7
1.005 1.01 Voltage 1.015 p. u. 1.02

Availability p.u.

-0.4
0.8 0.6 0.4 0.2

-0.2 Reactive

0 Power Q

0.2

Voltage 40

profile

REFERENCES
[1] UNE-EN 50160 Caractersticas de la tensin suministrada por las redes generales de distribucin [2] J. Mur, J. Salln, A. A. Bayod, Statistical model of wind farms for power flow, International Conference On Renewable Energies And Power Quality (ICREPQ03), Vigo 2003. [3] A. E. Feijoo, J. Cidrs, J.L.G. Drenlas, Wind speed simulation in wind farms for steady-state security assessment of electrical power systems, IEEE Trans, on Energy Conv., Vol. 14, No 4, Dec. 1999, pp 1582-1588. [4] C.L. Masters, J. Mutale, G. Strbac, S. Curcic and N. Jenkins Statistical evalualtion of voltages in distribution systems with embedded wind generation, IEE Proc.-Gener. Transm. Distrib. Vol. 147, N 4, July 2000, pp. 207-212. [5] R. Allan, R. Billinton, Probabilistic Assessment of Power Systems, Proceedings of the IEEE, Vol. 88, No 2, February 2000, pp 140-162.

Probability density

30

20

10

-0.4
1.005 1.01 Voltage 1.015 p. u. 1.02

-0.2 Reactive

0 Power Q

0.2

Farm with

kmax cap = 0,3 = tan(max cap); kmax ind = 0,5 = tan(max ind)