Note : Les descriptions sont présentées dans la langue officielle dans laquelle elles ont été soumises.
~14~:)988
The invention relates to controlling the quenching angle of an
inverter connecting a direct-current line to an alternating-voltage network.
In high-voltage direct-current transmission CHVDC~, on the inverter
side, the smallest possihle quenching angle y, with adequate stability, is
required. In order to meet this requirement, there are two ways of control-
ling the quenching angle, either to a predetermined rated value, or taking
into account the foreseeable commutating angle ~.
One method of quenching-angle con~rol is described in "Wissen-
schaftliche Berichte AEG-Telefunken", Vol. 44 (1971), issue 1, pages 6 to
12. In stationary operation, higher accuracy can be obtained by regulation
as compared with control, but only provided the quenching angle, used as
the regulating factor, can be obtained with the highest possible accuracy
from the voltage at the inverter rectifiers. This, however, requires a
considerable expenditure.
In contrast to this, the use of quenching-angle control offers
advantages, in that the control can react more quickly to changes in
lnfluencing factors, namely direct current and alternating voltage. This
requlres predeterminatlon of these factors, which is not very expensive.
According to the publication "Thristoren" by Heumann/Stumpe,
Verlag B.G. Teubner, Stuttgart, 1969, page 93, the following applies for
control-angle ~, which must be presented as reference input for the control:
cos ~ = cos YN - kl u
wherein YN is the rated quenching angle, i is the value of the direct
current related to the rated value, and u is the value of the alter-
nating voltage related to the rated value, taking into account the
transformation ratio between the voltage at the voltmeter and the voltage
at the inverter-rectifier, while kl takes into account the commutating
reactance. Control-angle ~ is also known as the lead-angle and is the sum
-1-
988
of quenching angle y and commutating angle w (~ =y ~ w). The commutating
reactance, which causes the înductive direct-voltage drop, is taken here
as being known and constant. If it changes, the effect upon the forming of
kl must be taken into account.
It is the purpose of the invention, in the case of a control of
the type mentioned at the beginning hereof, to make it possible to pre-
determine the reference input for control-angle ~, in the case of inverter
installations with stationary inverter operation, with the necessary
accuracy, i.e. without a cos-function generator.
According to the invention, this purpose is achieved as indicated
in that the control-angle ~, corresponding to the firing time of the
inverter rectifiers, numbered backwards from zero passage to semi-
oscillation of the commutating voltage, is predetermined in accordance
with the relationship:
~2 ~ - 2 (1 - cos YN) K u
as the reference input, wherein:
YN ls the rsted value of the quenching angle,
i ls the value of the direct current related to the rated value,
u ia the value of the alternating voltage related to the rated value,
taking into account the ratio between the voltage at the voltmeter and
the voltage at the inverter rectifier,
K is four times the relative direct-current change.
The invention makes use of the knowledge that the cosine relation-
ships in the above-mentioned known equation for control-angle ~ are large
as compared with the actual factor kl i , and that, in the case of small
angles, any dependency upon the angle disappears completely, so that the
cosine function is preferably replaced by a trigonometrical series. This
makes it possible to subtract the number 1 from both sides of the above-
988
mentioned equation, thus reducing the numerical value, and therefore the
error, to a ~raction. The following apply:
1 - - + ~4 - ... = 1 _ rN ... _ kl - or converted
~2 + ... = 2 (1 - cos YN) + K 1 , with K = 2 kl
The cos-series could be carried as far as required, but in view
of component tolerances, there is no point in going beyond the ~4 member.
This produces advantageously, and at no great cost, the following relation-
ship for predetermining the reference input:
12 2 (1 - cos ~N) ~ K i = 0
wherein cos ~N i9 presented as a constant for rated quenching angle ~N.
A further reduction in cost and complexity may be obtained in
that the control is not com6ined with a bridge or a group on the inverter
side of the HVDC, so that it need be present only once per half-station
on the inverter side. The voltage-measurement component for alternating
voltage u i8, in fact, necessary only once per station. The quenching-
angle control comes lnto action rapidlg at all changes in voltage and
current, except voltage changes during commutating, the results of which
cannot, however, be captured by the quenching-angle control.
The control assumes that the alternating voltages are free of
harmonics, equal in value, and in phase, and that there are no changes in
current or in the alternating-voltage mains.
The alternating current may be attenuated in such a manner that
voltage drops are picked up without delay, whereas the picking up of volt-
age increases is greatly delayed.
According to another advantageous development, only positive
changes in current are superimposed in such a manner as to compensate for
the effect of changes in current.
114V~88
Phase changes in the alternating current may ~e picked up by a
phase-detector and be added to the reference input. This is based upon
the assumption that a temporarlly excessive quenching angle is acceptable,
so that in restricting or eliminating the said sources of faults, an
evaluation of the sign of the quenchîng-angle fault will be made.
An advantageous example of apparatus for carrying out the control
method according to the invention, with only a few circuit elements, is
characterized in that:
for each phase of the alternating voltage, a first multiplier is provided
and is acted upon at the input end, through both connections, by the
value of the respective phase-voltage related to the rated value, the said
multiplier being connected, at the output end, to the input or a first
amplifier comprising in the reverse feedback branch, a second multiplier,
both inputs to which are supplied with the output voltage from the said
first amplifler;
a third multiplier i8 provided, one input to which carries a voltage
corresponding to the direct current (i) related to the rated value,
while the other input carries a voltage corresponding to the conetant (K);
a second amplifier i9 provided, which i8 connected to the output from
the third multiplier, and in the feedback-branch of which is arranged a
fourth.multiplier, one input to which is connected to the output from the
first amplifIer, while the other input i8 connected to the output from
the second amplifier;
a differential amplifier is provided, from the output of which the control-
angle ~ is taken and to which are also connected two inputs to a fifth
multiplier, the output from which passes, through a monovalent ohmic
resistor, to the negative input to the differential amplifier;
a sixth multiplier is provided, having two inputs to which are connected
88
the output from the fifth multiplier? while the output th~refrom is
connected, t'nrough a dodecavalent ohmic resistor, to the positive input
to the differential amplifier;
the posit;ve input to the differential amplifier carries, in addition to
the output voltage from the second amplifier, a voltage corresponding to
the factor 2(:1-cos yN2.
An example of embodiment of the invention is explained hereinafter
in conjunction with the drawing attached hereto, showing a schematic circuit
diagram for establishing control-angle ~ as the reference input for
controlling the quenching angle for an inverter (not shown). The said
inverter is to connect a direct-current line, carrying current i related
to the rated value, to a three-phase network having phase-voltages UR, Us,
UT related to the rated value. A rated quenching angle YN is provided for
the inverter drive. K is the input voltage for the circuit shown and
corresponds to four times the change in the relative inductive direct
voltage resulting from the commutating processes in the inverter. Input
voltage K therefore takes into account the commutating reactances (which
are variable under certain circumstances).
In the circuit shown, three input voltages corresponding to phase-
voltages UR, Us, UT related to the rated value, and also multiplied by the
ratio between the voltage at the voltmeter for the said phase-voltages and
the voltage at the inverter rectifier, are fed to both inputs of the three
first multipliers 11,12,13. The outputs from the first multipliers 11,12,13
are connected, through ohmlc wiring resistors 7,8,9, to the input to a
first amplifier 22, in the negative feedback circuit oE which is located
a second multiplier 16 with an ohmic resistor 6. The two inputs to second
multiplier 16 carry the output voltage from the said first amplifier 22.
1140~88
A voltage corresponding to current i and a voltage corresponding
to factor K are fed to a third multiplier 14 wh~ch îs connected, through
an ohmic resistor 3, to the înput of a second amplifier 21. Arranged in
the feedback branch of second amplifîer 21, together with an ohmic
resistor 4, is a fourth multiplier 15, to one input of which is fed output
voltage u of first amplîfier 22, and, to the other input, the output volt-
age from second amplifi~r 21. Second amplifier 21 delivers, through an
ohmic resistor 5, a voltage K . i to the positive input of a differential
amplifier 23.
The desired value R may be taken from the output of this differ-
ential amplifier 23, the said output being also connected to both inputs of
a fifth.multiplier 17. The output from the latter is passed, through a
monovalent ohmic resistor 1 to the negative input of differential amplifier
23 and also to the two inputs of a sixth multiplier 18, the output from
whlch is connected through a dodecavelent ohmic resistor 2, as compared
with ohmic resistor 1, to the positive input to differential amplifier 23.
In addition to the voltage K , i , this connection to differential amplifier
23 carries a voltage corresponding to the factor 2 (1 - cos YN).
The circuit described delivers, at the output from differential
amplifier 23, at low cost and in a simple manner the desired magnitude
for control-angle ~, in accordance with the following relationship
achieved in the circuit:
12 2 (1 - cos YN) - K i = 0
