U.S. patent application number 09/785674 was filed with the patent office on 2001-11-08 for control system adapted to control operation of an ac/dc converter.
Invention is credited to Ainsworth, John Desmond.
Application Number | 20010038544 09/785674 |
Document ID | / |
Family ID | 9885824 |
Filed Date | 2001-11-08 |
United States Patent
Application |
20010038544 |
Kind Code |
A1 |
Ainsworth, John Desmond |
November 8, 2001 |
Control system adapted to control operation of an ac/dc
converter
Abstract
A control system to control at least one of the thyristor
converters (1) or (2) of an ac system (7,8) or (9,10) connected by
busbars (5,6) and transformers (3,4) to the converters (1,2) which
have their dc sides connected by a dc link, for example inductor
(11). Each thyristor converter receives firing pulses from a firing
control (12) or (13), including a firing angle control (18) and a
summing junction (16) receiving a control current signal input (17)
and a current signal (14) which is a function of converter
operation. An electrical power measurement is made of one of the ac
systems, for example the ac system (7,8,5) and in accordance with
that power measurement a signal which is a function thereof is
produced as an output (19) from an integral two-axis ac servo
control having demodulator, integrator and modulator means
receiving the power measurement signal as input. Signal (19) is a
function of the cross-modulation noise component caused by
interaction of electrical noise from the two converters. This
signal (19) is input to at least the firing control (12) to vary
the firing of thyristor converter (1) so as to remove from at least
the ac system (7,8,5) said cross-modulation noise. Signal (19) may
also be input to firing control (13).
Inventors: |
Ainsworth, John Desmond;
(Staffordshire, GB) |
Correspondence
Address: |
Kirschstein, Ottinger, Israel & Schiffmiller, P.C.
489 Fifth Avenue
New York
NY
10017-6105
US
|
Family ID: |
9885824 |
Appl. No.: |
09/785674 |
Filed: |
February 16, 2001 |
Current U.S.
Class: |
363/71 |
Current CPC
Class: |
H02M 7/1623 20130101;
H02M 7/2195 20210501; H02M 1/12 20130101; Y02B 70/10 20130101; Y02E
60/60 20130101; H02J 3/36 20130101 |
Class at
Publication: |
363/71 |
International
Class: |
H02M 001/00 |
Foreign Application Data
Date |
Code |
Application Number |
Feb 18, 2000 |
GB |
0003706.9 |
Claims
1. An auxiliary control system adapted to control at least one of
the converters of an ac/dc system comprising a pair of ac/dc
converters with a dc link extending between them linking two ac
systems, at least one of said converters comprising a plurality of
switches and including a main control means providing converter
operating pulses or signals to said switches, said auxiliary
control system being arranged to take an electrical power or
reactive power measurement of one of said ac systems and in
accordance with said measurement cause said at least one of said
converters to operate so as to substantially remove from said ac
system a cross modulation noise component caused by interaction of
electrical noise from said two converters.
2. An auxiliary control system as claimed in claim 1, in which said
auxiliary control system comprises an integral two-axis ac servo
system comprising demodulating means, integrating means and
modulating means, said measurement of electrical power or reactive
power being input to said demodulating means and an output of the
modulating means modulating said switches of said at least one
converter.
3. A control system as claimed in claim 1, in which at least one or
each of said ac/dc converters is a thyristor converter.
4. An auxiliary control system as claimed in claim 1, in which at
least one or each of said ac/dc converters i a voltage-source
converter.
5. An auxiliary control system according to claim 1 in which gain
and phase settings of said auxiliary control system are
substantially fixed.
6. An auxiliary control system according to claim 1, in which gain
and phase settings of said auxiliary control system are varied in
dependence on at least one measured value of an electrical
phenomenon in said ac/dc converter according to a pre-determined
characteristic.
7. An auxiliary control system as claimed in claim 6, in which said
phenomenon is the dc voltage of said converter.
8. An auxiliary control system as claimed in claim 6, in which said
phenomenon is the dc current.
9. An auxiliary control system as claimed in claim 2, in which gain
and phase settings of said auxiliary control system are set to
values substantially equal to a reciprocal of a measured complex
value of total loop gain of said main control systems and ac/dc
converters and ac systems and auxiliary control system excluding
said integrating means.
10. An auxiliary control system as claimed in claim 9, in which an
automatic measurement of said total loop gain is carried out after
any substantial disturbance in the operation of said ac/dc
converters and in which said measurement remains effective in
determining said gain and phase settings until the next such
disturbance.
11. An auxiliary control system as claimed in claim 10, in which
said automatic measurement of total loop gain is by injecting
pre-determined signals into said modulating means and measuring
resulting outputs of said demodulating means.
12. An auxiliary control system as claimed in claim 11, in which
two said total loop gain measurements are performed in succession
with different injection signals and a result of said gain
measurements is calculated from a difference between the said two
gain measurements.
Description
[0001] This invention relates to a control system adapted to
control the operation of an ac/dc converter, particularly in
applications where a pair of ac systems are linked by two ac/dc
converters connected by a dc link.
[0002] Systems are well known where a pair of ac systems are linked
by two ac/dc converters connected by a dc link. For instance
propulsion systems for ships can include such a link wherein an ac
supply of fixed frequency is fed into an ac/dc converter acting as
a rectifier to produce dc. This dc is then fed into an ac/dc
converter acting as an inverter to drive a motor. The advantage of
this arrangement is that the firing angle of the inverter can be
controlled to adjust the speed of the motor without affecting the
first ac system.
[0003] Dc links are also known to exist between the supply grids of
neighbouring countries.
[0004] However, as is well known the ac/dc converters can produce
electrical noise at harmonic frequencies of the ac supply
frequencies. These harmonic frequencies can pass through the ac/dc
converters and give rise to cross modulation frequencies on the ac
systems. These cross modulation frequencies appear as noise on the
ac systems, that is they can lead to voltage flicker (causing
lamps, etc. to flicker) and can lead to resonances in
electro-mechanical systems. An example of such an
electro-mechanical system is a turbine/generator arrangement
provided in power stations. The turbine and generator are linked by
a shaft which has a natural frequency. The natural frequency of
this system is of the same order as the cross modulation
frequencies and it is theoretically possible that the system (in
particular the linking shaft) could be damaged through resonance of
the system due to excitation by the cross modulation. It is
therefore clearly advantageous to remove this cross modulation.
[0005] According to the invention there is provided an auxiliary
control system adapted to control at least one of the converters of
an ac/dc system comprising a pair of ac/dc converters with a dc
link extending between them linking two ac systems, at least one of
said converters comprising a plurality of switches, a main control
means providing converter operating signals to said switches, said
auxiliary control system being arranged to take an electrical power
or reactive power measurement of one of the ac systems and in
accordance with that measurement cause the controlled converter to
operate so as substantially to remove from the ac system a cross
modulation noise component caused by interaction of electrical
noise from the two converters.
[0006] Conveniently the auxiliary control system is additional to
the main control systems of the converters.
[0007] An advantage of such an auxiliary control system is that the
cross modulation component that had previously caused interference
within the ac electrical systems connected by the dc link can be
removed. This can help to remove or reduce the likelihood of damage
to electromechanical systems within those ac systems or unwanted
effects such as light flicker.
[0008] The controlled converter may be provided with main control
means providing converter operating signals or pulses.
[0009] Preferably the auxiliary control system comprises
demodulating means, integrating means and modulating means, and can
include an input which is a function of said power or reactive
power measurement and an output to modulate said control means of
said main controlled converter.
[0010] In one embodiment the auxiliary control system comprises an
integral two-axis ac servo system comprising aforesaid
demodulating, integrating, and modulating means and said input
being input to said integral two-axis ac servo system arranged to
provide said output.
[0011] At least one or each of said ac/dc converters may be a
thyristor converter or a voltage-source converter.
[0012] In one embodiment gain and phase settings of said auxiliary
control system may be substantially fixed.
[0013] In another embodiment gain and phase settings of said
auxiliary control system can be varied in dependence on at least
one measured value of an electrical phenomenon in said ac/dc
converter according to a pre-determined characteristic. Said
phenomenon may be the dc voltage of said converter and/or said
phenomenon may be the dc current.
[0014] Gain and phase settings of said control system can
alternatively be set to values substantially equal to the
reciprocal of a measured complex value of total loop gain of said
control system and ac/dc converters and ac systems excluding said
integrating means.
[0015] An automatic measurement of said total loop gain may be
carried out after any substantial disturbance in the operation of
said ac/dc converters and in which said measurement remains
effective in determining said gain and phase settings until the
next such disturbance.
[0016] Said automatic measurement of total loop gain may be by
injecting pre-determined signals into said modulating means and
measuring resulting outputs of said de-modulating means.
[0017] Advantageously two said total loop gain measurements may be
performed in succession with different injection signals and a
result of said gain measurements is calculated from a difference
between the said two gain measurements.
[0018] There now follows, by way of example only, description of
the invention with reference to the accompanying drawings in
which:-
[0019] FIG. 1 shows a general arrangement of the main circuit and
main converter controls for realising the present invention;
[0020] FIG. 2 includes a general arrangement of an auxiliary
control system according to the invention;
[0021] FIG. 3 shows the generation of the reference signals in FIG.
2;
[0022] FIG. 4 is a schematic circuit for measuring the loop gain
(G) of the circuit being controlled by the control arrangement of
FIG. 2;
[0023] FIG. 5 is a schematic circuit for inserting the function
(1/G.sub.s) in the circuit of FIG. 2;
[0024] FIG. 6 shows more details of a basic 6-pulse converter
circuit using thyristors;
[0025] FIG. 7 shows details of a voltage-source converter circuit
using transistors;
[0026] FIG. 8 shows details of gain and phase change block; and
[0027] FIG. 9 shows additional controls for the second control
method.
[0028] The control system of this invention is used to control two
ac/dc converters linked by a dc link. FIG. 1 shows a first example
of an arrangement of this, represented as single-line equivalents
for converters 1, 2, transformers 3, 4, ac system busbars 5, 6, ac
system impedances 8, 10, and ac system emf's 7, 9. This is for
thyristor converters. FIG. 6 shows more detail of one of the
converters, in the form of a bridge circuit 1 using six thyristors
in a 6-pulse arrangement. It is to be understood that for high
rated power each thyristor shown may, in reality, consist of many
thyristors in series and/or parallel; in addition an alternative
arrangement, not shown, may be two such bridges connected in
series, connected on their ac sides to separate transformer
secondaries respectively in star and delta configuration, in the
known from of a 12-pulse group. The configuration shown may be used
for example in high-voltage direct-current (HVDC) applications, or
for driving an ac electric motor.
[0029] As shown the converters are connected on their dc sides via
a single inductor 11. However, this connection may alternatively be
direct (no dc inductor) or via an overhead dc line or a dc cable,
or other form of dc system.
[0030] In normal operation each converter will operate with its own
main control system, shown respectively as 12, 13. Typically if
converter 1 is operating as a rectifier its main control may adjust
dc current to an ordered value, and converter 2 will operate as an
inverter, adjusting dc voltage to an ordered value. This is shown
in further detail for converter 1 only (although a similar
arrangement would be applied to converter 2), where a dc current
signal 14 is shown derived from a dc current transformer 15, and
applied to a summing junction 16. The difference between this and a
current order signal 17 is then effectively applied to a firing
angle control 18 to converter 1. A further modulating signal 19
will be described later.
[0031] It has been found that the ripple components caused by each
converter in the dc line current can cross-modulate to form
modulations on the ac side of each converter.
[0032] As a first cross-modulation example, consider the effect of
negative sequence unbalance in the ac system of converter 1. If the
nominal ac system frequency in this is f.sub.1 then the effect is
to generate a component at a frequency of 2 f.sub.1 on dc current;
similarly converter 2 may generate a component at 2 f.sub.2 on dc
current.
[0033] The cross-modulation seen in each ac system is then at
frequencies of (2 f.sub.1+2 f.sub.2). This will be defined as
2.sup.nd harmonic cross-modulation. Normally only the difference
frequency component at (2 f.sub.1-2 f.sub.2) is of practical
importance. As an example, if the system frequencies are each
nominally 50 Hz but with a variation of up to .+-.3 Hz, then the
principal modulation frequency may vary from zero to 12 Hz. If the
nominal system frequencies are different, for example 50 Hz and 60
Hz, also with 13 Hz variations, the modulation frequency may vary
between 8 Hz and 32 Hz.
[0034] As a second example of cross-modulation, assume each
converter to be a 12-pulse thyristor arrangement. Ripple components
will exist on dc current at frequencies of 12 f.sub.1 and 12
f.sub.2 due to the normal harmonics of the converters, giving
12.sup.th harmonic cross-modulation in each ac system at a
frequency of (12 f.sub.1-12 f.sub.2). If each system frequency is
50 Hz nominal with variations up to 3 Hz, the cross-modulation may
be at frequencies from zero to 72 Hz. Cross-modulation may also
occur at other frequencies such as 4.sup.th harmonic, or between
current components at different harmonic orders, but the effects of
these are generally small.
[0035] In each of the examples above the effect will be of
modulation in each ac system. This will generally be small, usually
less than 3% of rated power, but can be sufficient to cause
unacceptable flicker in voltages of supplies to consumers in the ac
systems, and can excite mechanical shaft resonances in machines in
the ac systems, possibly leading to over-stressing of shafts. The
flicker may manifest itself as flickering lights, or other
phenomena.
[0036] A second example of coupled ac/dc converters (not shown) may
be similar to FIG. 1 but based on voltage-source converters. These
differ in that they use switching devices such as insulated gate
bipolar transistors (IGBT) rather than thyristors, and that instead
of the series inductor 11 in FIG. 1 they may share a common
capacitor connected in shunt to each on their dc sides.
Alternatively such converters may each have their own dc capacitor,
coupled via a dc line or cable or single inductor, as for the first
case.
[0037] Cross-modulation may occur in this second example,
particularly at 2.sup.nd harmonic caused by negative sequence
unbalance of the ac systems as before. The mechanism is different
in that it is caused by 2.sup.nd harmonic components of voltage on
the shared dc capacitor or dc system, rather than by shared
2.sup.nd harmonic currents as in the first example. The effect is
however similar in that cross-modulation components may appear in
each ac system at (2 f.sub.1.+-.2 f.sub.2). It is possible in all
cases to reduce the cross-modulation components by adding passive
filters in the dc system, for example by making the inductor 11 in
FIG. 1 very large, or by adding shunt filters on the dc side, but
the cost of these can be large.
[0038] It has been found that cross-modulation components in the ac
systems of two coupled ac/dc converters can be nulled by injection
of a small modulating signal at the frequency of the
cross-modulation, with suitable amplitude and phase, into the main
control system of one or both converters.
[0039] FIG. 2 shows the general arrangement of the auxiliary
control system for the generation of one such signal according to
the invention. The description will initially be for nulling of
12.sup.th harmonic cross-modulation. This description will be in
mathematical terms, based on functions such as add, or multiply, or
divide, but it is to be understood that these represent physical
control equipment operating in real time.
[0040] In the upper part of FIG. 2 is shown a means of generating a
measurement of total converter power. This shows the addition of a
voltage transformer 20 and a current transformer 21 to the ac
system of converter 1. The outputs of these are multiplied in
multiplier 22 to form a power signal 23. Although shown as a single
line diagram, it is to be understood that these components may in
reality be three-phase, including three multipliers and a summing
junction (not shown) so that signal 23 is proportional to total
3-phase power. Providing a measure of power in this manner is
necessary because the cross modulation component that the invention
tries to cancel is not apparent in a suitable form from signals
from one single voltage and a single current sources.
[0041] The auxiliary control system is shown as block 24. This is a
form of integral two-axis ac servo control, having two channels,
respectively to generate the direct and quadrature components of
the desired modulation 19 to the main converter control system 12
(as shown in FIG. 1).
[0042] The input signal to the auxiliary control system 24 is the
ac power signal 23 derived from the multiplier 22 and the voltage
and current transformers 20,21 as described above. This will
contain a component proportional to the power modulation to be
nulled, plus a large dc component proportional to the mean ac
power. The dc component is blocked by a high-pass filter 25. The
signal then passes through a demodulator 26 containing two
multipliers, forming two signals with dc components V.sub.D1 and
V.sub.Q1 respectively proportional to the direct and quadrature
components of measured power modulation.
[0043] The multipliers within the demodulator 26 multiply the
signal passing from the high pass filter to the demodulator by a
reference signal which is derived as will described
hereinafter.
[0044] The signals V.sub.D1 and V.sub.Q1 then pass through a
gain/phase shift unit 27 described later, then an integrator block
28 containing two integrators. From here the signals pass via an
injection block 29 also described later, and then via a modulator
block 30 containing two multipliers and a summing junction which
combines the two signals into a single signal. This single output
signal is the modulation output signal 19, applied to the main
control system of converter 1, as in FIG. 1. It will be appreciated
by the skilled person that the prior art methods of controlling a
pair of ac/dc converters linked by a dc link does not make use of
this control signal 19. The control signal 19 is produced for
putting the invention into effect.
[0045] Reference signals in the form of two ac signals 32 and 33 in
quadrature at the cross-modulation frequency (12 f.sub.1-12
f.sub.2) are required by the demodulator 26 and modulator 30. (It
should be remembered that this control scheme is being described in
relation to a method of cancelling the cross modulation produced by
the 12.sup.th harmonic). These may be generated from the 3-phase
busbar voltages of each converter by first generating 2-phase
signal systems at the 12.sup.th harmonic frequencies (12 f.sub.1
and 12 f.sub.2) of the respective ac systems; this may be by known
frequency multiplication means, such as phase-locked
oscillators.
[0046] The production of these references 32 and 33 is shown in
FIG. 3. The input signals to reference generation means 31 are
derived as described above from the ac bus bars and are v.sub.A,
v.sub.B, v.sub.C, and v.sub.D. These inputs define the
instantaneous values of the resulting ac signals and are defined as
follows:
1 V.sub.A = sin A V.sub.B = cos A V.sub.C = sin B V.sub.D = cos B
where A = 12(2.pi.f.sub.1)t B = 12(2.pi.f.sub.2)t t = time
[0047] The required reference signals v.sub.refD (signal 32) and
v.sub.refQ (signal 33) are therefore obtained via block 31, shown
in FIG. 3. Block 31 effectively solves the trigonometrical
relations:
2 V.sub.refD = sin (A-B) = sinA cosB - cosA sinB V.sub.refQ = cos
(A-B) = cosA cosB + sinA sinB
[0048] These two signals are shown as 32, 33 in FIGS. 2 and 3.
[0049] In order to describe basic operation it is convenient to
initially ignore the gain/phase shift block 27 and the injection
block 29 of FIG. 2 and consider that signals pass straight through
these blocks 27 and 29 without change. Then ideally the auxiliary
control system should settle to a steady state condition in which
the resulting modulation signal 19 has the correct magnitude and
phase to null the 12.sup.th harmonic modulation in the ac system of
converter 1.
[0050] The integrators within block 28 then have mean input
signals, defined as V.sub.D11 and V.sub.Q11 as in FIG. 2, which are
then practically zero. The integrator outputs are finite, and
settle to values such as to give the appropriate values of
V.sub.D2, V.sub.Q2 so that the cross-modulation components in ac
system 5, 8, 7 are reduced to zero or near zero.
[0051] In practice, as in any closed-loop control system, the
settings of loop gain and phase must be within reasonable limits to
obtain a suitable performance and avoid instability. Adjustments to
each of these may be via gain/phase shift unit 27 discussed below,
which modifies the signals V.sub.D1 and V.sub.Q1 to produce
V.sub.D11 and V.sub.Q11.
[0052] A first method of arranging gain and phase control settings
is simply for these to be fixed at compromise values determined by
calculation or experiment. This may be practicable where the
operating conditions of the main converter are relatively constant.
This method may be by control setting signals 34, 35 from fixed
sources applied to gain/phase shift 27. Details of the latter are
shown in the known form of FIG. 7. This contains four multipliers,
two adders, a SINE function and a COSINE function. If the gain and
phase input control signals are respectively defined as G and
.theta. this block generates output signals V.sub.D11 and V.sub.Q11
which are related to inputs V.sub.D1 and V.sub.Q1 by:
3 V.sub.D11 = G(V.sub.D1 cos.theta. - V.sub.Q1 sin.theta.)
V.sub.Q11 = G(V.sub.D1 sin.theta. + V.sub.Q1 cos.theta.)
[0053] These are known transformations, and in terms of two-axis
quantities give a gain change in the ratio G, and a phase change by
.theta. radians. Block 29 is not required in this method. Gain G
may be set at a compromise value between too small a value, giving
slow response, and too large a value, which may give control loop
instability. In addition there is always an optimum phase setting
.theta. giving best stability; a value of phase setting differing
from this by more than 90.degree. will give instability; some
lesser tolerance such as .+-.30.degree. is acceptable.
[0054] A second method of determining settings is possible by
automatic open-loop adjustment if only moderate or well-defined
changes in the converter operating conditions occur. This may be
for example from a measurement of main converter dc voltage and
current, and automatic adjustment of phase and gain from
pre-determined relations to these. This may be suitable for a
medium range of operating conditions. FIG. 9 shows a suitable
arrangement for this, containing four multiplying constants
K.sub.1, K.sub.2, K.sub.3, K.sub.4, and two adders. The input
signals are shown as dc voltage V.sub.DC and dc current I.sub.DC,
obtained from the main dc system respectively from a dc voltage
divider and a dc current transformer, not shown. The constants may
be chosen to give compromise values for gain G and phase .theta.
which are effective over a wider range of converter working
conditions than for the first control method. In this method block
29 is again not required.
[0055] A third method of determining the optimum gain and phase
settings may be used, in which these are calculated from a brief
preliminary test of the main system, with the auxiliary control
loop open. This may be carried out automatically for example when
the main converter is first deblocked or otherwise switched on, or
on the occurrence of any substantial disturbance to the system
being controlled.
[0056] To understand this method, it is first necessary to define a
system gain as G.sub.s, measured from the input of the modulator
block 30 (signals V.sub.D2, V.sub.Q2) through the main converter
control system, the main converter, the whole of the dc and ac main
systems, the power measurement circuit, and the demodulator 26.
(This excludes only blocks 27, 28, 29 of FIG. 2). Since the dc
voltages (V.sub.D1, V.sub.Q1) and (V.sub.D 2,V.sub.Q2) represent
the complex components of ac quantities, G.sub.s will in general be
complex, of the form:
G=G.sub.D+jG.sub.Q
[0057] It can then be shown that if the equivalent transfer
function of block 27 is arranged to be (1/G.sub.s), then the
equivalent phase of this will automatically have exactly the
optimum value for best stability, and a gain which can also be
suitable (independent of the magnitude of G) if a suitable extra
fixed gain is applied also, for example at the integrator
inputs.
[0058] The blocks relevant to this method are 27 (which is
different to that in the first two methods), 29 and a measuring
block 36 (which is shown in FIG. 4). The measuring block is in
addition to the other blocks of the auxiliary control system 24
shown in FIG. 2.
[0059] A measurement of G can be carried out in principle simply by
injecting known signals into the modulator 30 and measuring the
resulting signals V.sub.D1 and V.sub.Q1. The effective frequency of
measurement referred to the main system is that of the reference
source, i.e. (12 f.sub.1-12 f.sub.2) in this example (this is being
described in relation to cancelling the 12.sup.th harmonic). If the
injection signals are for example equivalent as a pair to (1+j0)
per unit of rated converter power, then gain G.sub.s can be shown
to be equivalent to (V.sub.D1+jV.sub.Q1) where V.sub.D1 and
V.sub.Q1 are the measured values. In practice a much smaller
injection signal, typically 0.01 per unit, will be desirable to
keep down system disturbance due to the test; a suitable scale
factor can allow for this. In reality a single measurement as
described will give an incorrect answer, because of the presence of
cross-modulation components at frequency (12 f.sub.1-12 f.sub.2) in
the main circuit, other than those caused by the injection. The
proposed method according to the invention therefore performs two
tests in succession, one with injection as above, the second with
different injection values, preferably (0+j0). It can then be shown
that the true value of system gain G is equal to the difference
between the results for the respective tests, and is independent of
the existing cross-modulation components in the main system.
[0060] A suitable means for carrying out the measurements is shown
in FIG. 4, and may be described as follows:
[0061] 1. Assume the first test starts at time t=0. A strobe pulse
P.sub.1 is generated from suitable timing means, not shown,
starting from t=0 and ending at time t=T where T may be for example
1 second.
[0062] 2. During this time the injection block 29 switches its two
outputs to the desired values, for example respectively 0.01 and
zero, first disconnecting these from the integrators.
[0063] 3. At time t=T strobe pulse P.sub.1 switches off.
Sample-and-hold circuits 37 and 39 in block 36 (FIG. 4) will
thereafter store the values of V.sub.D1 and V.sub.Q1 measured at
time t=T.
[0064] 4. At time t=T a second strobe pulse P.sub.2 then starts,
and continues for a further time of T, i.e. from t=T to t=2T.
[0065] 5. During this time the injection block 29 switches each of
its outputs to zero, i.e. an equivalent injection of (0+j0).
[0066] 6. At time t=2T strobe pulse P.sub.2 turns off, leaving a
second pair of values of V.sub.D1 and V.sub.Q1 on the outputs of
sample-and-hold circuits 38, 40 in block 36 (FIG. 4).
[0067] 7. From t=2T onwards outputs 43, 44 of the summing junctions
41, 42 in block 36 give the difference of the two tests above; as
stated above, this represents the true system gain G in equivalent
complex form.
[0068] The reason for the two time delays T above is to allow the
complete system to settle to a steady condition before sampling.
Smaller or larger times for these may be used, depending on the
actual signal frequency ranges expected.
[0069] The switching or injection block 29 is not shown in detail,
but is relatively simple, containing only switches operated by
pulses P.sub.1 and P.sub.2, and dc signal sources, as described
above.
[0070] It is then required to control the gain/phase block 27 such
that it has an equivalent transfer function of 1/G, i.e. the
equivalent complex reciprocal of the signals from the measuring
block 36. Assume G is expressed as G.sub.D+jG.sub.Q. Then the real
and imaginary components of its reciprocal can be shown to be:
4 V.sub.D11 = (V.sub.D1 G.sub.D + V.sub.Q1 G.sub.Q)/(G.sub.D.sup.2
+ G.sub.Q.sup.2) V.sub.Q11 = (V.sub.Q1 G.sub.D - V.sub.D1
G.sub.Q)/(G.sub.D.sup.2 + G.sub.Q.sup.2)
[0071] A suitable method of implementing these is shown in block 27
in FIG. 5. This contains only multipliers, summing junctions and
one reciprocal. (The alternative of converting 1/G to polar form,
with block 27 also in polar form, is much less convenient,
particularly in on-line digital control systems, which would
require sine, cosine, arctangent, and square-root functions, which
take much computing time).
[0072] This third method of determining control settings will
operate in a wide range of system conditions, for example where
there are resonances between an ac system impedance and local shunt
filters associated with the main converter, particularly where such
resonant frequencies may change substantially with changes in the
ac system or in converter operation.
[0073] The description above refers only to an auxiliary control
system having an input from measured power at one converter, hence
in principle only nulls cross-modulation power in that converter;
in practice the energy storage in the converters and dc system is
relatively low, so that power modulations in the dc circuit and in
the `remote` ac system are also substantially attenuated.
[0074] However it is possible that modulation in reactive power in
each ac system may still be substantial; this may have relatively
little effect on machine torques but can affect ac voltage flicker.
It is then possible to provide two auxiliary control systems
substantially as described above, respectively to null the real and
reactive components of cross-modulation. These can share the ac
reference sources in common, but are otherwise separate. A
preferred method of applying these is for the measurement and
control of real power modulation to be at one converter, and the
measurement and control of reactive power modulation to be at the
other converter, though different conditions of these are
possible.
[0075] The control system description above was for an auxiliary
control system to null 12.sup.th harmonic cross-modulation. The
method may be applied alternatively to a control system for any
other cross-modulation frequency, for example at 2.sup.nd harmonic
as described previously. The changes required are relatively few,
being principally to the generation of reference signals, which may
be by the same method but from appropriately different basic
frequencies.
[0076] It will also be clear to those skilled in control art that
filters (omitted above) may be required in various parts of the
auxiliary control system, to reduce the effects of signal
components at unwanted frequencies, for example those due to other
types of cross-modulation.
[0077] If settings are adjusted by the third method it will also be
required to provide means to automatically re-trigger the test
process described for determining control settings in various
circumstances. This will normally be required on normal start-up of
the main converter by either deblocking or closing a main
circuit-breaker. A re-test should preferably also occur on recovery
of the power system from a major disturbance such as a temporary ac
or dc system fault. Yet another case will be where the working
conditions of the main converter or ac system change relatively
slowly, into a condition where the initial settings for the
auxiliary control are unsuitable, and it may gradually become
unstable. This can be detected by a measurement of the integrator
input voltages V.sub.D11 and V.sub.Q11; when either of these
becomes finite, in excess of a small set value for a set time, for
example 1 second, a re-test can be triggered.
[0078] An example of a system where working conditions of the main
converter or ac system change relatively slowly would be a DC link
between two ac supply grids, which may be the supply grids of
neighbouring countries. In such systems the supply frequencies are
independently set and therefore are likely to drift over time; such
drifting could cause the control functions to become unstable, and
hence can trigger a re-test as said above. An auxiliary control
system with functions as described may be implemented by electronic
means known in control system art, such as by analogue electronic
devices, or digital electronic devices, provided that in the second
case calculations are carried out at an iteration frequency
substantially greater than the relevant harmonic frequencies.
[0079] The description above was particularly relevant to thyristor
converters. However, as mentioned in the introduction the invention
is also applicable where one or each converter is a voltage-source
converter. FIG. 7 shows one type of voltage-source converter in the
form of one 3-phase bridge. The switching elements, shown as plain
circles, may each consist of a transistor connected in reverse
parallel with a diode, or series and/or parallel combinations of
these. Other types of voltage-source converter are also applicable,
such as a single-phase bridge, or two bridges in a "12-pulse"
arrangements.
[0080] In each case one or more shunt capacitors 45 may be
connected across the dc side of the or each bridge. If both
converters are voltage-sourced converters, they may be coupled on
the dc side via an inductor 11, or this may be omitted; in the
latter case one capacitor 45 only is required.
[0081] A characteristic of voltage-sourced converters is normally
that the transistors are switched on and off at a relatively high
frequency, in a switching time pattern which inherently nulls many
harmonics as seen on the ac systems. Thus for example, for a
3-phase converter, assuming balanced conditions, there may be
virtually zero ac harmonics up to 23.sup.rd harmonic or usually
higher, hence 12.sup.th harmonics on the dc current, and 12.sup.th
harmonic cross-modulation, may not be significant. However, the
effect of negative-sequence unbalance in one or both ac systems is
similar to that in thyristor converters, in that it caused 2.sup.nd
harmonic on the dc current, and 2.sup.nd harmonic cross-modulation.
The invention is also applicable in reducing such 2.sup.nd harmonic
cross-modulation effects in voltage-sourced converters to small
values or zero, as for a thyristor converter.
* * * * *