# I graduated from NTNU, Department of Electrical Power Engineering in December 1997. After wor- king a few months in a project for Jernbaneverket I started my studies towards a PhD in april 1998. The thesis is planned to be finished during the spring in 2002.

Ηλεκτρονική - Συσκευές

24 Νοε 2013 (πριν από 4 χρόνια και 7 μήνες)

73 εμφανίσεις

28
,QLWLDWLRQ
I graduated from NTNU, Department of Electrical
Power Engineering in December 1997. After wor-
king a few months in a project for Jernbaneverket
I started my studies towards a PhD in april 1998.
The thesis is planned to be finished during the
spring in 2002.
My PhD is a part of the EFFEKT research pro-
gram, and is funded by the Research Council of
Norway. One year is funded by NTNU, Depart-
ment of Electrical Power Engineering.
My supervisor is Professor Tore M. Undeland.
%DFNJURXQG
The HVDC-converters in the planned HVDC-ter-
minals to the continent represent a risk of harmo-
and maybe other FACTS-components will be
introduced in the grid. It is therefore important to
map potensial problems connected to harmonics
and to find actions which prevent harmonic pene-
tration in the main grid.
,QWURGXFWLRQ
A HVDC-converter has typically a 12-pulse thy-
ristor bridge. The commutation voltages for this
bridge are the voltages at the filter busbar at the
AC-side of the converter. Normally these voltages
are reasonably sinusoidal. Nevertheless, the
HVDC-converter generates substantial harmonic
currents which are feeded into the AC-grid if not
filtered. Under idealized conditions (which means
symmetric voltages and impedances in the AC-
system, constant DC-current and no error in the
firing of the thyristors), the order n of the harmo-
nic currents generated on the AC-side are given
by:
(1)
k is a positive integer.
The magnitude of a harmonic AC-current of order
n is:
(2)
I
1
is the magnitude of the fundamental component
of the AC-current.
To prevent harmonic currents to enter the AC-grid
filtering is necessary. This is traditionally done by
using RLC- branches (passive filters) which crea-
tes a low-impedance path to ground for the expec-
ted harmonic currents. Under idealized conditions
as described before, the passive filters will operate
satisfactory, and also in many cases in real life.
Nevertheless there are certain drawbacks when
using passive filters. A very important drawback
is that they may cause resonance with the AC-
network impedance. This may cause substantial
distortion in the commutation voltages, and thus
even more harmonic currents are generated. It
may also cause large harmonic currents. Changes
in the components of the passive filters due to age-
ing and temperature variations are also a major
drawback with passive filters. These changes
result in new impedance caracteristics of the pas-
sive filters, and this causes more of the harmonic
currents to enter the AC-grid. In many cases pas-
sive filters are unsatisfactory when the HVDC-
converter generates non-characteristic harmonic
currents (i.e harmonic currents with orders diffe-
rent than those given by Eq.(1)). Non-characte-
ristic harmonics origins in that the idealized
conditions described earlier are not met. When
using passive filters the presence of non-characte-
ristic harmonics can cause resonance conditions
between the passive filters and the AC-network
impedance, and reinforce harmonic instability
n 12k 1±=
I
n
I
1
n
----
=
3RZHUHOHFWURQLFVLQWKHSRZHUV\VWHP
by Ian Norheim
29
conditions due to interaction between harmonics
on the DC-side and the AC-side of the HVDC-
converter. An example of this is core saturation
instability which can shortly be described as in the
following paragraph.
A small 2nd harmonic positive sequence voltage
component on the AC-side produce a fundamental
voltage component on the DC-side. If there is a
series-resonance at fundamental frequency on the
DC-side a relatively large fundamental current
will be present. Through the control system of the
HVDC-scheme this might cause unequal firing
pulse spacing and this may reinforce the funda-
mental DC-side current. Due to the coupling bet-
ween the AC-side and the DC-side harmonics
through the converter a 2nd harmonic positive
sequence and a negative sequence DC-current (the
sum of the DC-currents in the tree phases are zero)
will flow on the AC-side. A paralell resonance
near the 2nd harmonic on the AC-side may rein-
force the 2nd harmonic voltages. In addition, the
negative sequence DC-current will skew magni-
tize the converter transformers, and this way cause
harmonics in all sequences. Under certain opera-
ting condition this phenomenon might develop
until the zero-sequence protection causes discon-
nection of the AC-line connected to the HVDC-
converter. Other consequences might be over-
heating of filters due to additional harmonics pro-
duced because of the DC-magnetization of the
converter transformers.
This type of harmonic instability (core saturation
instability) is probably the reason for several line
disconnections at the existing HVDC-converters
in Kristiansand. Statnett has through simulations
in EMTDC tried to recreate this instability, but
without success. This may have several causes. It
can be that the model they use is not accurate
enough, and/or that they include to many compo-
nents in their simulations and lose the overview of
the system, and/or that the component models
they used in EMTDC were not accurate enough.
In my work I seek to understand the mechanisms
that causes 2nd harmonic instability through the-
ory and simulations. It is also in this context
important to understand the models used in the
simulation program (EMTDC). When the mecha-
nisms for core-saturation instability and the
EMTDC-models are understood it is natural to
simulate this type of instability. In a report I’ve
given a basic theoretical background for core satu-
ration instability and demonstrated it through
simulations on a modified CIGRE HVDC bench-
mark model. After finishing this report I’ve tested
out preventive actions in the control system of the
HVDC-scheme. Both Fourier-filtering and band-
pass filtering has been used in this context. It
seems like the Fourier-filter is most successful.
Other preventive actions like online filter tuning,
air gap in the converter transformers, and active/
hybrid filtering has not been tested yet, but might
be interesting topics in the work ahead.
During my work I receive useful supervision from
Prof. Tore Undeland (NTNU), Magnar Hernes
(SEfAS), and Thor Henriksen (SEfAS).
5HIHUHQFHV
[1] AC-DC power system analysis. J. Arrillaga
and Bruce Smith. The Institution of
Electrical Engineers, 1998.
[2] Working paper on Core-saturation instability.
Ian Norheim. NTNU, 2000.