AckermannTh. (ed) Wind Power in Power Systems

Подождите немного. Документ загружается.

//INTEGRAS/KCG/PAGINATION/WILEY/WPS/FINALS_14-12-04/0470855088_07_CHA06.3D – 101 – [97–114/18]

20.12.2004 7:35PM

Reactive power and power factor

The reactive power is measured as a 10-minute average value (IEC guideline) or as a

1-minute average value (German guideline) over the whole power range of the wind

turbine. The reactive power is given for each 10 % step of the active power from 0 % to

100 % of rated power. Instead of reactive power, the German guideline requires the

measurement of the power factor.

Harmonics

According to the guidelines, harmonic measurements are not required for fixed-speed

wind turbines (Type A), where the induction generator is directly connected to the

grid.

(1)

Harmonic measurements are required only for variable-speed turbines with

electronic power converters (Types C and D).

Wind turbines with electronic power converters produce harmonic currents. The three

guidelines include different requirements concerning harmonic measurements. The IEC

guideline only requires the measurement of integer harmonic currents up to the 50th

order. In general, the power converters of wind turbines are pulse-width modulated

(PWM) invert ers, though, which have clock frequencies in the range of 2–3 kHz and

produce mainly interharmonic currents. Thus the requirements of the IEC guideline do

not really reflect the harmonic emission of the wind turbine.

The MEASNET guideline and also the Germ an guideline require measurements of

interharmonic currents of up to 2 kHz and of current distortions in the higher frequency

range of between 2–9 kHz. That means that these two guidelines reflect the harmonic

emission of wind turbines. Even though both guidelines require measur ements in the

same frequency range, the results of such measurements are not comparable. The

MEASNET guideline requires measuring intervals of 10 minutes. The relevant value

is the maximum 10-minute value of each frequency. Integer harmonic currents and

interharmonic currents are grouped according to IEC 61000-4-7 Ed.2.0 (IEC, 2002). In

this context, grouping means that all interharmonic currents between two neighbouring

integer harmonics are combined into a single interharmonic current, for example. Current

distortions in the higher frequency range are grouped into bandwidths of 200 Hz.

The German guidel ine requires measuring intervals of 8 line periods. For each

frequency, a 99 % value of the harmoni c (or interharmonic) current is determined (i.e.

the value that is not exceeded in 99 % of all cases). These 99 % values are relevant.

Grouping for interharmonics or integer harmonics is not used. Thus, interharmonics are

given in steps of 6.25 Hz. Similar to the MEASNET guideline, the current distortions in

the higher frequency range are grouped into bandwidths of 200 Hz.

In practice, the main difference between the German and the MEASNET guideline lies

in the different averaging intervals. In general, the harmonic emission of wind turbines

with power converters is not steady. The harmonics, interharmonics and higher frequency

distortions behave stochastically. There is an averaging effect, which leads to lower values

in the measurements according to the MEASNET guideline. Owing to the very short

measurement intervals of the German guideline, the corresponding values are higher.

(1)

For definitions of turbine Types A to D see Section 4.2.3.

Wind Power in Power Systems 101

//INTEGRAS/KCG/PAGINATION/WILEY/WPS/FINALS_14-12-04/0470855088_07_CHA06.3D – 102 – [97–114/18]

20.12.2004 7:35PM

One of the main problems of the harmonic measurements at wind turbines is the

influence of the already existing harmonic voltages in the grid. The voltage waveshape of

the grid is, of course, not sinusoidal. There are always already harmonic voltages in the

grid, such as integer harmonics of 5th and 7th order, which affect the measurements. In

many cases, the wind turbine behaves like a consumer of such harmonics. It makes no

sense to evaluate such harmonics, because they originate from the grid. The problem is

to identify and separate the harmonic emission originating from the turbine and that

originating from the grid. A first indicator may be the phase angle of the harmonic and

the time or power dependency. We also have to know whether there are wind turbines

with power converters in the vicinity of the turbine that is measured. The harmonic

emission of the neighbouring turbines can influence the harmonic measurement, par-

ticularly as the harmonic emission of the neighbouring turbines may also be correlated

to the wind speed.

Flicker

The term ‘flicker’ means the flickering of light caused by fluctuations of the mains

voltage, which can cause distortions or inconvenience to people as well as other elec-

trical consumers. Flicker is defined as the fluctuation of voltage in a frequency range of

up to 35 Hz. Flicker evaluation is based on IEC 61000-3-7 (IEC, 1996). The basis for the

evaluation is the curve given in Figure 6.1, with a threshold value for the short-term

flicker disturbance factor, P

,ofP

¼ 1. This curve shows the level where flicker is a

visible disturbing factor to people. The most sensitive frequency is 8.8 Hz.

IEC 61000-4-15 (IEC, 2003) specifies a flickermeter that can be used to measure

flicker directly. In general, flicker is the result of the flicker that is already present in

the grid and of the emissions to be measured. A direct measurement will require an

undisturbed constant-impedance power supply. However, this is not feasible for wind

turbines because of their size. The flicker measurement is therefore based on measure-

ments of three instantaneous phase voltages and currents, which are followed by an

analytical determination of P

for different grid impedance angles. The measured

0.1

0.1 1 10 100 1000 10 000

Rectangular voltage changes per minute

Voltage change, d (%)

Figure 6.1 Curve of the regular rectangular voltage charge (as a percentage of the nominal value)

against the number of changes per minute, for a short-term flicker disturbance, P

,of1

102 Power Quality Measurements

//INTEGRAS/KCG/PAGINATION/WILEY/WPS/FINALS_14-12-04/0470855088_07_CHA06.3D – 103 – [97–114/18]

20.12.2004 7:35PM

voltage and current time series of the wind turbines are the input to a simple grid model

(see Figure 6.2). For different grid impedances, especially grid impedance angles, time

series of voltage fluctuations are calculated for this fictitious grid. The simulated time

series of instantaneous voltage fluctuations are the input to the standard voltage flicker

algorithm in order to generate the flicker emission values P

. The algorithm is descri bed

in IEC 61000-4-15. Such P

values are calculated for a larger number of measured time

series over the whole power range of the turbine. Weighted with standard wind-speed

distributions, a 99 percentile of the P

values is calculated, which ensures that for 99 %

of the time, the flicker of the wind turbine will be within this 99 percentile. Last, a flicker

coefficient is calculated from this 99 percentile of the P

values. The flicker coefficient

gives a normalised, dimensionless measure of the flicker, independent of the network

situation and thus independent of the selected short-circuit apparent power of the

fictitious grid. It determines the ratio between short-circuit power and generator-rated

apparent power, which is necessary to achieve a long-term flicker level, P

, of 1. The

flicker coefficient c is defined as:

cð

; V

Þ¼P

ð6:1Þ

where

c(

, V

) is the flicker coefficient, dependent on the grid impedance angle 

and on

the annual average wind speed V

is the short-circuit power of the grid at the point of common coupling (PCC).

is the apparent power of the wind turbine at rated power .

is the long-term flicker emission.

This flicker coefficient can be used by utilities to calculate the flicker emission of a wind

turbine or a wind farm at a specific site.

The procedures of IEC 61400-21 (IEC, 2001) and the German guideline are similar.

The only difference is that the German guideline requires 1-minute time inter vals and

thus 1-minute P

values, and the IEC guideline requires 10-minute intervals. This leads

to different results, with the flicker coefficients of the German guideline in general

slightly exceeding those of the IEC guideline.

Wind

turbine

Grid impedances

Ideal voltage supply

Figure 6.2 Simulation model for flicker measurements

Wind Power in Power Systems 103

//INTEGRAS/KCG/PAGINATION/WILEY/WPS/FINALS_14-12-04/0470855088_07_CHA06.3D – 104 – [97–114/18]

20.12.2004 7:35PM

Switching operations

Switching operations can cause voltage fluctuations due to in-rush currents and can

cause flicker. Therefore, the German guideline and the IEC guideline as well as the

MEASNET guideline require measurements concerning flicker and voltage fluctuations

during switching operations. Owing to the already existing flicker and voltage fluctu-

ations in the grid, the measur ement procedure is similar to the procedure for flicker

measurements during normal operation. This means that the measured time series

during switching operations is inserted into a grid model with specific grid impedances

(see Figure 6.2). The model produces the time series of voltages, from which the voltage

fluctuations are calculated. The normalised voltage fluctuations lead to a voltage change

factor, which is a dimensionless value of the voltage fluctuations during switching

operations. In addition, the standard flicker algorithm is used to calculate a flicker step

factor from the voltage fluctuations that the grid model generates. This flicker step

factor is the basis for the calculation of the flicker impression that the switching

operations of a wind turbine or farm at a specific site will cause.

The German guideline requires calculation methods that are similar to those of the

IEC guideline. In the German guideline, though, the flicker impression and the voltage

fluctuation are combined into a single grid-related switching factor. This factor is very

easy for the utilities to apply. It is, however, less accurate than the procedure of the IEC

and MEASNET guidelines.

The following switching operations have to be measured:

cut-in at cut-in wind speed;

cut-in at rated wind speed;

switching operations between generator stages;

service cut-out at rated power (German guideline only).

6.2.3 Future aspects

In some countries, such as Germany and Denmark, there is a high penetration of

electricity generated from wind energy. Grid operators will have to take precautions

in order to ensure a safe and stable operation of the grid, taking into account the

expected increase in production from wind energy. Therefore, some utilities have

established new additional requirements (EON, 2003; VDN, 2004; Eltra, 2000; NGC,

2004 for a discussion of the new grid requirement s, see Chapter 7).

Until now, the philosophy regarding wind farms connected to the grid has been to

switch off the farm as soon as there is a fault on the grid. The new guidelines are a

radical change to this philosophy. They require wind farms to remain connected in the

case of grid faults (e.g. short- term voltage drops). The wind farms are expected to

support the grid. Therefore, wind farms have to control reactive power over a wide

range, deliver reactive power in the case of voltage drops and remain connected during

short-term voltage drops. And they must be able to operate over a wide frequency range.

In order to ensure that wind turbines and wind farms will fulf il these new require-

ments, in Germany, an additio nal measurement guideline has been developed (FGW,

2003). This guideline includes methods to check the compatibility of the turbine or farm

104 Power Quality Measurements

//INTEGRAS/KCG/PAGINATION/WILEY/WPS/FINALS_14-12-04/0470855088_07_CHA06.3D – 105 – [97–114/18]

20.12.2004 7:35PM

with the requirements of EON (2003) and VDN (2004). The guideline includes the

following measurements and checks:

range of reactive power (capacitive as well as inductive);

power gradient after grid losses;

operation at underfrequency and overfrequency;

power reduction through set-point signal;

protection system concerning frequency and voltage changes;

behaviour of the wind turbine in the case of short voltage drops (ride-through capability).

6.3 Power Quality Characteristics of Wind Turbines and Wind Farms

The grid interferences of wind turbines or wind farms have different causes, which are

mostly turbine-specific. The relevant parameters are listed in Table 6.2. Average power

production, turbulence intensity and wind shear refer to causes that are determined by

meteorological and geographical conditions. All the other causes are attributable to the

technical performance of the wind turbine. This performance is determined not only by

the characteristics of the electrical components, such as generators, transformers and so

on, but also by the aerodynamic and mechanical behaviour of rotor and drive train. The

turbine type (i.e. variable versus fixed speed stall versus pitched controlled) is of major

importance to the power quality characteristics of wind turbines and wind farms.

6.3.1 Power peaks

Variable-speed wind turbines (Types C and D) can control the power output of the

inverter system by pitch c ontrol, thus smoothing power fluctuatio ns as well as power

peaks. Thus the power peaks lie within the range of the rated power. Instantaneous

power peaks of fixed-speed wind turbines (Type A) often exceed rated power by 30 % or

more, even in the case of pitch-controlled fixed-speed turbines. The pitch control is not

Table 6.2 Grid interferences caused by wind turbines and wind farms

Parameter Cause

Voltage rise Power production

Voltage fluctuations and flicker Switching operations

Tower shadow effect

Blade pitching error

Yaw error

Wind shear

Fluctuations of wind speed

Harmonics Frequency inverter

Thyristor controller

Reactive power consumption Inductive components or generating systems

(asynchronous generator)

Voltage peaks and drops Switching operations

Wind Power in Power Systems 105

//INTEGRAS/KCG/PAGINATION/WILEY/WPS/FINALS_14-12-04/0470855088_07_CHA06.3D – 106 – [97–114/18]

20.12.2004 7:35PM

fast enough to control fast power peaks. Slower power peaks, such as 1-minute and 10-

minute power peaks, can be reduced by the pitch control. These slower power peaks are

therefore also close to rated power, also for pitch-controlled fixed-speed turbines (Type

A1). The power peaks of a stall-controlled fixed-speed turbine (Type A0) depend on air

pressure and air temperature, among other things. This has the effect that 1-minute and

10-minute power peaks often exceed rated power by about 10 % to 20 %.

The number of wind turbines on a wind farm is important for smoothing power

peaks. In particular, fast power peaks at the individual wind turbines of a wind farm are

in general uncorrelated and thus smoothed throughout the wind farm. IEC 61400-21

includes a formula for this smoothing effect, which in principal adds up the power peaks

of the single wind turbines geometrically.

6.3.2 Reactive power

The reactive power demand of the asynchronous g enerator of fixed-speed wind turbines

(Type A) is partly compensat ed by capacitor banks. Thus the power factor (i.e. the ratio

of active power and apparent power) lies, in general, at about 0.96. Variable-speed wind

turbines with PWM inverter systems (Types C and D) have an inverter to control the

reactive power. Thus, these wind turbi nes have, in general, a power factor of 1.00. These

turbines can control the reactive power over a wide range (inductive and capacitive). It is

therefore possible to control the voltage and to keep it more stable at the grid connec-

tion point of the wind farm or wind turbine (see also Chapter 19).

6.3.3 Harmonics

Today’s variable-speed turbines (Types C and D) are equipped with self-commutated

inverter systems, which are mainly PWM inverters, using an insulated gate bipolar

transistor (IGBT; see also Chapter 4). This type of inverter has the advantage that both

the active power and the reactive power can be controlled. It has the disadvantage,

though, that it produces harmonic currents. In general, the inverter generates harmonics

in the range of some kilohertz. Therefore, filters are necessary to reduce the harmonics.

Two types of PWM inverters are used: those with a fixed clock frequency and those

with a variable clock frequency. Figure 6.3 gives examples of the harmonic emission of

wind turbines with such inverter systems. The main difference between each type is that

the inverter with a fixed clock frequency [Figure 6.3(a)] produces single interharmonics

in the range of the clock frequency and multiples of the clock frequency. Inverters with

variable clock frequency [Figure 6.3(b)] have a wide band of interharmonics and integer

harmonics. Resonances of the grid are excited by this wide band of interharmoni cs and

integer harmonics. The result are interharmonic and harmonic cu rrents, as Figure 6.3(b)

illustrates, where the specific maximum occurs at the resonance frequency of the grid.

The measurement of the harmonic currents poses one of the biggest challenges to the

measurement of power quality. Harmonic current measurements require great accuracy,

even for high frequencies, because the measurements refer to interharmonics that are in the

range of 0.1 % of the rated current for frequencies of up to 9 kHz (for MEASNET and the

German guideline). Therefore, up to 9 kHz, current clamps need to have a linear ratio.

106 Power Quality Measurements

//INTEGRAS/KCG/PAGINATION/WILEY/WPS/FINALS_14-12-04/0470855088_07_CHA06.3D – 107 – [97–114/18]

20.12.2004 7:35PM

Harmonic measurements at the low-voltage side of the wind turbine may give results

that are different from measurements at the medium-voltage side. Investigations have

shown that in some cases the transformer of the turbine has an influence on the

harmonics (Santjer, 2003). The transformers of wind turbines are often connected at

the medium-voltage side in delta operation and at the low-voltage side in star operation.

Owing to this connection, a single phase at the medium-voltage side is influenced by two

phases at the low-voltage side. This can lead to a smoothing effect that reduces some of

the harmonics. In particu lar, the so-called ‘zero current’ may be eliminated. Thus

harmonic measurements at the medium-voltage side of the transformer often produce

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0 250 500 750 1000 1250 1500 1750 2000 2250

Frequency (Hz)

2500

Current /rated current (%)

(

)

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0 250 500 750 1000 1250 1500 1750 2000 2250 2500

Frequency (Hz)

Current/rated current (%)

(b)

Figure 6.3 Harmonic current emission of a variable-speed wind turbine where the inverter has

(a) a fixed clock frequency and (b) a variable clock frequency

Wind Power in Power Systems 107

//INTEGRAS/KCG/PAGINATION/WILEY/WPS/FINALS_14-12-04/0470855088_07_CHA06.3D – 108 – [97–114/18]

20.12.2004 7:35PM

lower values than at the low-voltage side, but, for practical reasons, measurements at the

medium-voltage side are often more difficult than at the low-voltage side.

6.3.4 Flicker

Fluctuations of active and/or reactive power of wind turbi nes cause flicker. Active

power fluctuations of wind turbines may be caused by the effect of the wake of the

tower, by yaw errors, by wind shear, by wind turbulences or by fluctuations in the

control system. The main reason for flicker in fixed-speed turbines (Type A) is the wake

of the tower. Each time a rotor blade passes the tower, the power output of the turbine is

reduced. This effect causes periodical power fluctuations with a frequency of about

1 Hz. Figure 6.4 gives an example of such power fluctuations in a fixed-speed wind

turbine. Power fluctuations due to wind-speed fluctuations hav e lower frequencies and

thus are less critical for flicker. In general, the flicker of fixed-speed turbines reaches its

maximum at high wind speeds. Owing to smoot hing effects, large wind turbines gen-

erally produce lower flicker than small wind turbines, in relation to their size. An

example of the flicker behaviour of a fixed-speed turbine (Type A) is given in Figure 6.5.

For variable-speed turbines (Types C and D), fast power fluctuations are smoothed

and the wake of the tower does not affect power output. Therefore, the flicker of

variable-speed turbines (Types C and D) is in general lower than the flicker of fixed-

speed turbines (Type A). Figure 6.6 gives an example of the power output of a variable-

speed wind turbine.

In wind farms, power fluctuations are smoothed because of the fact that the power

fluctuations of the single wind turbines are uncorrelated. The flicker of a wind farm is

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0.45

0.5

0 5 10 15 20 25 30 35

Time (s)

Active power/rated power

Figure 6.4 Active power output of a fixed-speed wind turbine (Type A); periodical fluctuations

are due to the wake of the tower

108 Power Quality Measurements

//INTEGRAS/KCG/PAGINATION/WILEY/WPS/FINALS_14-12-04/0470855088_07_CHA06.3D – 109 – [97–114/18]

20.12.2004 7:35PM

the geometrical sum of the flicker of all single turbines in the farm. This means that for a

wind farm consisting of n single turbines of the same type, the flicker of the farm is

ﬃﬃﬃ

times the flicker of a single wind turbine.

6.3.5 Switching operations

Voltage changes during switching operations are due to in-rush current s and the

respective changes in the active and reactive power of a wind turbine. For fixed-speed

turbines, a soft-starter limits the in-rush current of the asynchronous generat or. The

0.1

0.2

0.3

0.4

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2

Active power/rated power

Flicker, P

Switchings

Figure 6.5 Flicker behaviour of a fixed-speed wind turbine as a function of active power output

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.1

90 95 100 105 110 115 120 125 130

Time (s)

Active power/rated power

Reactive power/rated power (kvar/kW)

Active power

Reactive power

Figure 6.6 Power output of a variable-speed wind turbine

Wind Power in Power Systems 109

//INTEGRAS/KCG/PAGINATION/WILEY/WPS/FINALS_14-12-04/0470855088_07_CHA06.3D – 110 – [97–114/18]

20.12.2004 7:35PM

soft-starter is generally based on thyristor technology and limits the highest root mean

square (RMS) value of the in-rush current to a level below twice the rated current of the

generator. Figure 6.7(a) gives an example of the cut-in of a fixed-speed turbine (Type A).

Before cut-in, the rotational speed of the rotor and that of the asynchronous generator

increase, driven by the wind. Once the synchronous speed is reached, the generator will

be connected to the grid. The soft-starter is in operation for about 1 or 2 seconds and

limits in-rush current. During this period, the generator needs reactive power to become

magnetised. A few seconds after the generator is connected, the capacitors are switched

on to minimis e the demand of reactive power. The fast power changes during the

switching operation cause flicker. The (large) changes in active and reactive power cause

voltage fluctuations.

–2

–1.5

–1

–0.5

0.5

1.5

0 5 10 15 20 25

Time (s)

Active power/rated power,

Reactive power/rated power (kvar/kw)

Rotational speed/synchronous speed

Cut-in of capacitors

Active power

Rotational speed

Reactive power

(a)

0.2

0.4

0.6

0.8

1.2

Time (s)

Active power/rated power

Rotational speed/synchronous speed

Rotational speed

Active power

0 102030405060

(b)

Figure 6.7 Cut-in of (a) a fixed-speed wind turbine (Type A) and (b) a variable-speed wind

turbine (Types C and D)

110 Power Quality Measurements