Influence of Transformers on Harmonics

Fritz Santjer
Deutsches Windenergie- Institut DEWI, Ebertstrae 96, D-26382 Wilhelmshaven
Phone.: 04421-4808-0; FAX.: 04421-4808-43; email: dewi @dewi.de; http //www. dewi. de
Rainer Klosse, formerly DEWI, now working at
University of Oldenburg, Energy and Semiconductor Research Laboratory
Phone: +49-441-798-3926; email: [email protected]; www.physik.uni-oldenburg.de/ehf

Summary
When wind turbines equipped with inverter systems are connected to the grid, harmonic currents can cause additional loads on the electrical components or additional costs if the harmonic filters have to meet the criteria of the
guidelines of the utilities. The transformer of the wind turbine influences the harmonics. In general the harmonics
at the medium voltage side of the transformer are lower than on low voltage side. Power quality measurements
are often performed at the low voltage side of the transformer due to practical reasons. The German Wind Energy
Institute found a method, how to calculate the harmonics at the medium voltage side from the harmonics, measured at the low voltage side. Measurements of harmonics with a high-precision measurement system demonstrate how measured values can be converted from the low voltage side to the medium voltage side by means of
the symmetric components in the complex three-phase system.

1. Theory

For a transformer of Dyn11 vector group the winding

connection diagram will be as shown in Diagram 1.
The currents of the secondary side IS of the transformer are calculated from the currents of primary
side IP, ignoring the transformer impedances, being the transmission ratio of the transformer, see
equation 1 - 3.
Two specific features become evident when looking
at the equations:
1. The three currents of the secondary side can
be depicted by means of only two currents of
the primary side. The measuring tolerance on
the secondary current calculated is doubled. In
the same way, only the positive and the negative phase-sequence currents are required.
2. The zero current of the low-voltage side is
short-circuited on the delta winding of the medium-voltage side, thus there is no zero current
on the medium-voltage side.
By way of illustration, the time characteristic of the
zero current can be calculated on the primary side
according to equation (3). The value indicated for
the zero current has the same error tolerance as the
mean value of all three currents.
wind turbine

grid

U1yp

U12s

I1p 1
U12p

I1s1
U31p
U2yp

U23s

I2p 2

I2s

U23p
U3yp
I3p 3

U31s
I3s 3

Diagram 1: Transformer of the vector group Dyn11

and connected wind turbine

I1S = ((1 a ) ImP + (1 a) IgP)/

2
2
I2S = ((a a) ImP + (a a ) IgP)/
2
I3S = ((a 1) ImP + (a 1) IgP)/
equation (1) Dyn5
2

I1S = ((1 a) ImP + (1 a ) IgP)/

2
I2S = ((a 1) ImP + (a 1) IgP)/
2
2
I3S = ((a a ) ImP + (a a) IgP)/
equation (2) Dyn11
i0(t) = (i1(t) + i2(t) + i3(t))/3
equation (3) zero current

2. Verification by measurement
For the verification of the calculation method the
three currents and voltages on the low voltage side
of a transformer of a wind turbine with IGBT converter, as well as the three currents on the medium
voltage side were measured synchronously. Measurements were carried out over the whole range of
the power output of the wind turbine. One spectrum
measured at a time when the wind turbines had
caused the highest load on the grid at the critical
frequencies, was evaluated separately. The transformer of the wind turbine is connected according to
vector group Dyn 11.
The wind turbine has been measured in the generator reference-arrow system (the energy flow runs
from the wind turbine into the grid). In this paper,
primary side of the transformer means low voltage
side.
Each measuring channel was recorded synchronously by a separate digital-analog converter in order to avoid a phase shift, as it is likely to occur
when using the usual multiplexer. This characteristic
is of crucial importance in the investigations. Recording the channels one after the other would soon
result in phase shifts of the harmonics, rendering
any further investigation invalid. The currents were
measured with different instrument transformers of
classes 0.2 to 1. The measurement accuracy, re-

cludes, in diagram 2, bottom left, the zero current

which can be calculated from the three currents of
the low-voltage side. Bottom centre of the diagram
shows the calculated current for the medium-voltage
side of Dyn11 transformer, taking into account only
the vector group, but not the transformer impedances. As already mentioned, this calculation includes only the positive sequence and the negative
sequence current. By comparison, the current
measured on the medium-voltage side is shown
bottom right. All the currents, also in the spectra
shown in the following diagrams, are referred to the
rated current of the turbine and indicated in %, taking into account the transformation ratio for 50Hz.
The maximum deviation of a sample (a single
measurement) between measured and calculated
current is 2% referred to the rated current of the
wind turbine.
For better clarity, the fundamental-frequency current
of 50 Hz was subtracted from diagram 2. The resulting distortion currents are given in diagram 3.
This shows even more clearly how the harmonic
currents are transmitted by a transformer.

ferred to a harmonic current of 1% of the rated current, is stated to be better than 7% referred to the
measured value. The harmonic spectra of the currents and voltages were defined over 8 grid cycles
each, in accordance with [1], which results in a band
width resolution of 6.25 Hz..
The equations discussed above were compared
with a measurement in three different ways. First,
the time characteristic of a grid cycle, as generated
by the wind turbine and recorded by the measurement system, was compared. Then this grid cycle
and 7 other grid cycles were used to determine the
spectres of a phase and again were compared with
each other. Finally, in order to prove the suitability of
the data for an evaluation according to Technische
Richtlinien Teil 3 [1], a complete data record of
12000 spectra over 8 grid cycles was compared
once more between measurement on the low voltage side and calculation or measurement on the
medium voltage side.
Time characteristic of the current.
Diagram 2, top, shows the 3-phase current characteristic measured on the low voltage side. It inStrom/Nennstorm [%]

60
40
20
0
-20
-40
-60
0

0,005

0,01

0,015

0,02
Zeit t[s]

60

40

40

20
0
-20

20
0
-20

-40

-40

-60

-60
0

0,005

0,01

0,015

Strom/Nennstrom [%]

60

40

Strom/Nennstorm [%]

Nullstrom/Nennstorm [%]

measured current at low voltage

60

0,02
Zeit t[s]

20
0
-20
-40
-60

0,005

0,01

0,015

0,02

0,005

0,01

0,015

Zeit t[s]

Zeit t[s]

0,02

Verzerrungsstrom/Nennstorm [%]

zero current
calculated current at medium voltage measured current at medium voltage
Diagram 2: time characteristics of the currents.
4
3
2
1
0
-1
-2
-3
-4
0

0,002

0,004

0,006

0,008
0,01
Zeit t[s]

measured current at low voltage (without fundamental current)

3
2
1
0
-1
-2
-3
-4

4
Verzerrungsstrom/Nennstrom [%]

Verzerrungsstrom/Nennstorm [%]

Nullstrom/Nennstorm [%]

3
2
1
0
-1
-2
-3
-4

0,002

zero current

0,004

0,006

0,008

0,01
Zeit t[s]

3
2
1
0
-1
-2
-3
-4

0,002

0,004

0,006

0,008
0,01
Zeit t[s]

calculated current at medium voltage

0,002

0,004

0,006

0,008
Zeit t[s]

measured current at medium

voltage
Diagram 3: time characteristics of the distortion currents (fundamental current was eliminated)

Spectra of a single measurement.

1

Strom / Nennstrom [%]

0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0

500

1000
1500
Frequenz [Hz]

2000

2500

1
0.9

0,8

0,8

0.8

0,7
0,6
0,5
0,4
0,3
0,2
0,1

Strom / Nennstrom [%]

1
0,9
Strom / Nennstrom [%]

Nullstrom / Nennstrom [%]

measured current at low voltage

1
0,9

0,7
0,6
0,5
0,4
0,3
0,2
0,1

0
500

1000
1500
Frequenz [Hz]

2000

2500

0.6
0.5
0.4
0.3
0.2
0.1

0
0

0.7

0
0

500

1000
1500
Frequenz [Hz]

2000

2500

500

1000
1500
Frequenz [Hz]

2000

2500

zero current
calculated current at medium voltage measured current at medium voltage
Diagram 4: Spectra of the currents (up to 2.5 kHz).
1

Strom / Nennstrom [%]

0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
2000

3000

4000

5000
6000
Frequenz [Hz]

7000

8000

9000

0,9

0.9

0,8

0,8

0.8

0,7
0,6
0,5
0,4
0,3
0,2
0,1

Strom / Nennstrom [%]

0,9
Strom / Nennstrom [%]

Nullstrom / Nennstrom [%]

measured current at low voltage

1

0,7
0,6
0,5
0,4
0,3
0,2
0,1

0
2000

3000

4000

5000
6000
Frequenz [Hz]

7000

8000

0.6
0.5
0.4
0.3
0.2
0.1

0
2000

9000

0.7

3000

4000

5000
6000
Frequenz [Hz]

7000

8000

0
2000

9000

3000

4000

5000
6000
Frequenz [Hz]

7000

8000

9000

zero current
calculated current at medium voltage measured current at medium voltage
Diagram 5: Spectra of the current in the frequency range 2kHz up to 9kHz, bandwidth 200 Hz.
1

Strom / Nennstrom [%]

0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0

500

1000
1500
Frequenz [Hz]

2000

2500

0.9

0.9

0,8

0.8

0.8

0,7
0,6
0,5
0,4
0,3
0,2

0.7
0.6
0.5
0.4
0.3
0.2

0,1

0.1

0
0

500

1000
1500
Frequenz [Hz]

2000

2500

Strom / Nennstrom [%]

0,9
Strom / Nennstrom [%]

Nullstrom / Nennstrom [%]

measured current at low voltage

1

0.7
0.6
0.5
0.4
0.3
0.2
0.1
0

500

1000
1500
Frequenz [Hz]

2000

2500

500

1000
1500
Frequenz [Hz]

2000

2500

zero current
calculated current at medium voltage measured current at medium voltage
Diagram 6: Spectra of the current of a complete data set of 12000 single measurements.

Strom / Nennstrom [%]

0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
2000

3000

4000

5000
6000
Frequenz [Hz]

7000

8000

9000

0.9

0.9

0,8

0.8

0.8

0,7
0,6
0,5
0,4
0,3
0,2
0,1
0
2000

Strom / Nennstrom [%]

0,9
Strom / Nennstrom [%]

Nullstrom / Nennstrom [%]

measured current at low voltage

1

0.7
0.6
0.5
0.4
0.3
0.2
0.1

3000

4000

5000
6000
Frequenz [Hz]

7000

8000

9000

0
2000

0.7
0.6
0.5
0.4
0.3
0.2
0.1

3000

4000

5000
6000
Frequenz [Hz]

7000

8000

9000

0
2000

3000

4000

5000
6000
Frequenz [Hz]

7000

8000

9000

zero current
calculated current at medium voltage measured current at medium voltage
Diagramm 7: Spectra of the current of a complete data set of 12000 single measurements, frequency range 2 kHz
to 9kHz, bandwidth 200 Hz.
Statistical evaluation
The spectra of diagrams 6 and 7 depict the resulting
analysis of 12000 single measurements. The calculated spectrum (up to 2.5 kHz) is rated not more
than 0.1% higher than the measured spectrum at
the medium voltage side (exception: 0,25% at
250Hz), and at least 0.001% referred to the rated
value of the wind turbine. For the frequency spectrum from 2kHz to 9kHz differences are not more
than 0.04%.

Richtlinie Teil 3 [1], but it may also lead to an

amendment of the Measnet recommendation [3], as
applied in Europe. An acknowledgment of this procedure would considerably simplify the measurements and lead to a cost reduction. A measurement
on the medium-voltage side, critical because of the
electrical insulation, is no longer necessary. Measurements carried out on the low-voltage side will get
more validity because they are able to give a good
approximate description of the relevant behaviour
on the medium-voltage side.

Further measurements

5. References

A number of 6 further measurements on the lowvoltage side have already been investigated. Another comparative measurement with a Dyn5
transformer has confirmed the above-mentioned results.

[1] Technische Richtlinie fr Windenergieanlagen

Teil 3: Bestimmung der Elektrischen Eigenschaften;
Rev. 15; Stand: 1.09.2002; Herausgeber: Frdergesellschaft Windenergie e.V.

3. Conclusion
The grid quality of a wind turbine is often rated
much more negatively by measurements carried out
on the low-voltage side than by measurements on
the medium-voltage side. We have demonstrated
here that the calculations specified allow to transpose spectra from the low-voltage side to the medium-voltage side. Here in particular the vector
group of the transformer has to be taken into account. With transformers of vector groups Dyn5 and
Dyn11, the zero current is not transmitted from the
low-voltage side to the medium-voltage side. The
transformer impedances only have a small influence
on the current and are normally within the measurement tolerances required according to [1].
4. Prospects
As a result of the investigations presented here, experts recommend that the relevant guidelines
should allow the transformation of harmonics from
the low voltage side to the medium voltage. This
change would not only concern the Technische

[2] Eigenerzeugungsanlagen am
Mittelspannungsnetz. Richtlinie fr Anschluss und Parallelbetrieb von Eigenerzeugungsanlagen am Mittelspannungsnetz. 2. Ausgabe 1998, Hrsg.: VDEW e.V.;
VWEW-Verlag, Frankfurt
[3] Power quality of wind turbines MEASNET measurement procedure, 02.11.2000