NO158769B - PROCEDURE AND APPROACH TO AA REDUCE A BACKGROUND EFFECT OF A COMMUNITY DIRECT CONVERT. - Google Patents

Description

The invention relates to a method

to reduce the feedback effects that a grid-commuting direct inverter connected to a three-phase alternating current grid exerts on the grid by equipping a grid-connected controllable reactive power compensator. as well as a control link to form the desired value for the compensator's equipment in accordance with the method.

Three-phase networks, especially weak networks, often react to

the reactive power taken up by connected rectifiers, with significant voltage fluctuations that can lead to visible flickering in other consumers, or to voltage distortions at the resonance point of the grid impedance. By means of conden-

sators or filter circuits can be compensated for a more reactive effect. Inductive or capacitive storage can also be connected to the grid with control devices (e.g. controlled direct current regulators), which are controlled using a superimposed control circuit for reactive power and are able to supply the grid with reactive power corresponding to the inverter's reactive power absorption.

Such reactive power regulation with associated measuring devices

is, however, relatively expensive and due to the time constants in the rule are only poorly suited to also compensate for the periodic modulation of the reactive load absorption that usually occurs with such converters.

It is common for power calculations to consider the arithmetic mean value of active current i^ and of reactive current i^

during a half-period of the inverter's output frequency and from it to form the conventional apparent power which is more accurately termed as apparent fundamental oscillation-

effect S^. Decisive for the dimensioning of electrical devices, however, is the real power, which is determined by the rms value of the modulated fundamental harmonic in the mains current during half a period of the output frequency, and which is constantly a certain power value ("modulation power") greater than the apparent power for the fundamental harmonic. Due to the inverter function, the absorption of reactive power in direct inverters is constantly modulated with a multiple of the inverter's output frequency. The result is a modulation of the mains current with overshoot and undershoot currents. Thus generating e.g. one of the 6-pulse sub-inverters built up direct converter with three-phase output, due to the aforementioned blind-

load modulation, in addition to the mains-frequency reactive current also side frequencies that are shifted in relation to the mains frequency by six times the ratio between output frequency and mains frequency. In the case of inverters with single-phase output, both the active power or the wattage and the reactive power for the fundamental harmonic are modulated so that additional side frequencies occur. The corresponding frequency spectrum of side frequencies also occurs for the additional overharmonics of the mains current which are dependent on the pulse number of the inverter connection, but for the most part, these overharmonics cause only small feedback effects on the grid.

The present invention is based on the task of compensating the fundamental oscillation reactive power and/or its modulation without any regulation of the reactive power with associated measuring devices. Thereby, the time constants for the usual regulation of reactive power are also eliminated, so the compensation can be carried out more quickly.

This is achieved by equipping a controllable reactive power compensator connected to the three-phase network, where the desired value for the control of the reactive current to be delivered from the compensator is independent of the measured value of the reactive current formed by the product of the value of the direct inverter output current and one from the inverter control voltage or the direct inverter output voltage (U ) derivative given function of the rectifier steering angle^. In this connection, the direct inverter's control angle can be derived from the direct inverter's output voltage or from its control voltage, e.g. with corresponding proportionality between control voltage and control angle is it possible to use the control voltage as control angle or as the function of the steering angle included in the product.

The basis for the invention is this recognition

that for the equipping of the compensating reactive current to be fed into the grid, it is sufficient to specify a desired value ifi te which, on the basis of theoretical considerations of the fundamental harmonic of the reactive current and idealistic approximations, can be calculated by the formula.

Here, k is a proportionality factor, /iA/ the absolute value of the direct inverter output current and' ty the control angle, when this is chosen so that the full positive (negative) amplitude of the inverter output voltage occurs for ty = ^90°. If the rectifier valves are controlled so that the control voltage is proportional to this control angle, then the control voltage can be used directly instead of the value ty in a corresponding calculation link to realize equation (1).

For mains commutating direct current converters, there is a relationship between the converter output voltage u^, the ideal no-load DC voltage U,. q and the steering angle ty, so that ty can be calculated as

As a rule, a control circuit is arranged for the control of the direct inverter where the control voltage U t for the direct inverter valves is formed from the measured values of the output current ia resp. the output voltage U and a current-desired value iB . The input quantities for equation (1) are thus already available in the form of corresponding desired values iB and ty (resp. U .). However, it may be advantageous to use one of the desired values or both desired values using the corresponding measured values iA and UA using equation (2).

According to the invention, the formation of the desired value for the equipment of the reactive power compensator according to the method can advantageously take place with the help of a control link in the form of a calculation link with a value generator that is incurred by the measured value or the desired value of the direct-inverter output current, a function generator that is incurred by a quantity corresponding to the direct-inverter -the control angle and has a predetermined function characteristic, and a multiplication term which is incurred by the output sizes from the value generator and the function generator and serves to form the product. The output value of the control link can be used directly or via a subsequent average value link which will be discussed later, as a setting value for the compensating reactive current iD to be delivered.

In cases where instead of ty in equation (1) the output voltage uA according to equation (2) is used, the function generator is connected to another function generator, which is the inverter's output

voltage is connected, and whose characteristic corresponds to the relationship between steering angle and output voltage.

It is often the case that the grid feedback caused by the average reactive current (the average value over a period of the output frequency) is not critical for the grid itself and for other consumers connected to the grid, or that the base-oscillation reactive load does not need to be fully compensated or must be compensated for economic reasons, but the side frequencies of the feedback frequencies due to the modulation of the reactive load must be suppressed. In this case, the desired value in * is advantageously not the full value calculated in accordance with equation (1) of the reactive current taken up by the inverter, but only its deviation Ai^ from a corresponding mean value ig^- This mean value can by smoothing is formed directly from the value i^ which has already been calculated from the corresponding equation (1). Since iA and UA change approximately sinusoidally with the inverter output frequency and thus in vector

form can be represented by rotating vectors of length iA and <u>A, the mean value for i ^ according to equation (1) can also

is calculated by the corresponding values iA a and u A resp. = are cos uA according to equation (2) for these rotating vectors can be inserted into equation (1). In this connection, the product of the mean value of the inverter output current over at least one mains frequency half-wave and the function of the corresponding mean value of the inverter control angle is calculated as the product mean value ^l'

Although a complete compensation for the reactive effect, especially in inverters with three-phase output, is often not economically possible, a complete compensation of the modulation effect may be desirable. In this case, for the formation of the desired value in la to the just calculated difference Ai^ which leads to perfect compensation of the modulation load, the value in ^ of the actual product mean value is added with a corresponding weighting. This trade-off can e.g. consist in lower values of i^^ to be switched on, but then with the help of a corresponding threshold value limitation, starting from a certain maximum value i ^ makS only this value is switched on for the basic oscillation reactive load that is to be compensated maximally.

In the case of direct inverters with single-phase output, the wattage also pulsates with double the output frequency. In this case, it is advantageous as a controllable reactive load compensator to use a compensator which can also compensate load variations, e.g. a rotating field magnetized slip ring machine that drives a flywheel. For controlling the watt current to be delivered by the compensator, a calculated value iw is advantageously used which is calculated as proportional to the product iA<sin<J;. The desired value for the compensation current for the watt-power modulation is then formed as the difference between this calculated value iw and a corresponding mean value iw«

These and further advantageous embodiments of the invention are specified in more detail in the claims.

In the following, the invention will be explained in more detail by way of examples with reference to the drawing. Fig. 1 shows a device with a direct converter which is connected to a three-phase alternating current network and has a single-phase output, a controllable compensation device and a control circuit consisting of a calculating circuit and a mean value circuit for forming the desired value for the equipment of the reactive load compensator with input values i<*> and Aug. Fig. 2 shows a calculation link and a mean value link with measured values iA and uA as input quantities. Fig. 3 is a connection for balanced connection of i^ to the difference. Figs. 4 and 5 show control connections for compensation of the watt-power modulation in direct converters with single-phase output. Fig. 6 shows an embodiment of the control connection in the case where a direct converter with three-phase output is connected to the grid. Fig. 7 shows a variant of mean value coupling h in a control coupling according to fig. 6.

According to fig. 1, a direct current converter with a single-phase output is connected to a three-phase network R, S, T, and consisting of two

antiparallel partial rectifiers I, II. At the inverter outputs 1,

2, the output voltage uA and the output current iA are sensed, which are supplied to a regulation device 3 which, according to a predetermined current demand value iA, determines a control voltage Ugt which is supplied to a control link 4. The control link 4 is constructed in such a way that there

there is proportionality between the control voltage U and the control angle ty. With the help of the output pulses from the control link 4, the valves in the rectifier 1 are controlled with an ignition angle ot.j. = tt/2 - ty resp. by changing the sign of the valves in rectifier II with ignition angle

aII = + fr/2.

Between the mains sockets R, S, T there are arranged bearings consisting of coils 5 and capacitors 6 which can be discharged into the mains via inductively loaded direct current regulators 7 to compensate for the reactive current taken up by the direct inverter I, II. A reactive current control device 8 serves to control this reactive current compensator 5, 6, 7, which supplies the ignition pulses for the valves of the direct current regulators 7.

To generate the desired value iD to which the reactive current control device 8 controls the reactive current to be fed in from the reactive current compensator 5, 6, 7, a calculation link 9 is arranged which simulates equation (1). This control link contains a value generator (rectifier) 10 which is affected by an electrical quantity corresponding to the output current iA. For this, one can use the measurement<y>erdi iA which appears at the output 1. With a sufficiently fast regulation device 3, one can, however, as shown in fig. 1, advantageously also use the corresponding desired value iA. Furthermore, there is arranged a function encoder 11 which is applied with the control voltage Ugt (in this case equivalent to the control angle ty), and whose output magnitude corresponds to cos ty. The output quantities from the two elements 10, 11 are supplied to a multiplication element 12 at whose output 13, where according to equation (1) a theoretically determined quantity for the basic oscillation reactive current ig^ can be extracted.

This value i^ can be supplied directly to the control element 8 as the desired value for the reactive current i to be delivered.

In the design example in fig. 1, however, the bearing 5, 6 is not designed for compensation of the total grid feedback from the direct inverter, as it is instead only the modulation effect that is to be compensated, while the fundamental oscillation blind effect remains uncompensated. For that reason, the difference Aig^ between the theoretical value i^ for the instantaneous reactive power current and the mean value iol for the fundamental harmonic of the reactive current is formed as the desired value ifi. To this, where the output 13 is connected downstream, a differential element 14 is applied from the output 13 with the value i^ and, with the opposite sign,

of the mean value 1^ which is taken from the output 13 via the smoothing link 15. This mean value connection 14, 15 is designed for a large smoothing to produce a small ripple of the mean value in ^ (time constant over several grid periods). However, this requires a long setting time when the mean value changes.

Advantageously, a mean value is therefore used

coupling 16' which is shown in fig. 2, and which is based on the fact that the quantities i and u change in time as sinusoidal sturgeon movements, so that each of them can be perceived as one component of a rotating vector with length (value) iA and uA respectively in a right-angled coordinate system. The calculation link 16<1> thus takes the input quantities iA a and u A as a point of departure, in the event that, analogously to fig. 1, the control voltage Ugt <=> ty is available instead of uA, the corresponding input for uA connects an additional function generator that simulates the relationship uA = U, . Q.sin^ in accordance with equation (2) from the iJ;/uA characteristic of the direct converter.

The calculation of the mean values iA and u"A now takes place by vector processing using the vector analyzers 17, 18

which receive each of the input sizes iA resp. uA connected as a component, while the other right-angle shifted components are derived from these input quantities by means of integrators 19, 20. The calculation of the mean value i^^ now takes place as the quantity uA in a function generator 23 corresponding to the ^/uA~ characteristic of the direct converter, is transferred to the steering angle mean value ty, after which the corresponding equation I ^ = IA.cosijI is formed using an angle function encoder

21 and a multiplication term 22.

Similar to the case shown in fig. 2, where there

as the input value for the mean value connection 16' is used

direct converter output voltage uA, this input quantity can also be used for the calculation link 9', where the function generator 11 is connected to the function generator 23 corresponding to the function generator 24 for the transfer of the input voltage uA to the control angle ty.

If the entire average basic vibration blind load, unlike fig. 1 does not remain uncompensated, but must be compensated up to a maximum value and only the reactive load that exceeds this maximum value must remain uncompensated, then according to fig. 3 to the output from the mean value connection 16 resp.

16' which delivers the magnitude Ai^ after connected to an addition link 26 which, in balanced form, is supplied with the value for i^^ which is taken via the line 25 at the output from the smoothing link 15 resp. the multiplication term 22. This trade-off can take place over a threshold value limiter 27 whose limiting threshold value e.g. can be adjusted with a potentiometer 28.

As already explained, the wattage of a single-phase output direct current converter as shown in fig. 1. This pulsation of the watt effect can be compensated if the used compensator is also designed for such compensations. For the modulated active component ^v mains current one can

according to the invention analogously to equation (1) set

when using the appropriate steering link, iA is again set proportional to Ugt and uA proportional to sinijj. To simulate equation (3), one can therefore again, in addition to the inverter output current in A or its desired value iA, proceed from the voltage uA or via a pre-connected angle function encoder,

3 0 (fig. 4) from the control voltage Ugt< In order to imitate the pulsating calculation components of the active current in the mains current corresponding to equation (3), it is then only necessary to supply the quantities iA and uA to a multiplication term 31. As a desired value for watt- the current equipment can there again use the computationally simulating pulsation Aiw = i^ - iw which is formed in a mean value connection 32 which is connected after the multiplication term 31 and analogously to the mean value connection 16 in fig. 1 consists of a smoothing joint 33 and a node 34. However, the mean value generation can also be obtained in a in fig. 5 shows average value coupling 32' as analogous to the average value coupling 16' in fig. 2 forms the mean values i'A and u'A by means of vectorial calculation over integrators 35, 36 and vector analyzers 37 and forms the mean value iw =

iA sin ty over a multiplication term 39.

The calculation links 9 resp. 9' can also be used for direct converters with multi-phase output. In fig. 6 shows such a connection for a three-phase direct current converter. Here, for the individual phase outputs a, b and c, one and one calculation link 9a, 9b and 9c are arranged, as in the case of fig. 1 is supplied with the corresponding phase output current respectively iAa, i^ and iAc (or its respective desired value i<*>a/ i^, and i<*>c) °9 the control angle ty^ resp. ty. respectively ty„ (or in the case of Fig. 2 the phase output voltage UAa or UAc is supplied via a corresponding function converter 24a, 24b and 24c respectively). Each of these calculation links also in this case contains value generators 10a, 10b and 10c respectively, an angle function generator respectively 10a, 10b and 10c and a multiplication term 12a resp. 12b or 12c. The output values from the multiplication sections 12a, 12b, 12c are combined in an addition section 40 to the theoretical value for the basic oscillation reactive current i^-^.

If this value i^ is not to be used as a desired value for compensation of the total basic oscillation reactive load, but only for compensation of the modulation effect, then the output 13 can also here analogously to fig. 1, a mean value connection 16 may be connected afterwards. However, one can also perform a mean value formation here by vectorial calculation of the input quantities,

as shown in fig. 7. To generate the mean value iA, a vector analyzer 17<1> is arranged in this way, which is supplied with the components of the output current vector iA in a right-angled coordinate system. Because the current vector iA is composed of the three phase currents iAa, i^ and iAc, the vector analyzer 17' is connected to a coordinate converter 41. The mean value ty is similar to fig. 2 formed by the phase output voltage u*„r u,u, u,. via a coordinate converter are transformed into de

Aa' Ab Ac

two components of the voltage vector uA with respect to a right-angled coordinate system, and in the vector analyzer 18' there will be

from the two components the vector magnitude is formed, which is then transformed in a function generator 23' into the steering angle mean value ijj. If instead of the phase output voltages uAa«'"- the corresponding control angles tya ..... or control voltages Ugt are to be used, then the control voltages or control angles can be transformed through using function converters 43a, 43b, 43c which are inserted in front corresponding inputs to the coordinate converter 2 2 and gives proportionality between IL

and \ pA, between U^ and ty^ and between UAc and ty . The characteristics of the function generators 43a, 43b and 43c on the one hand and the function generator 23' on the other hand are mutually inverse. These function generators are in fig. 7 only entered dashed

since it turns out that a simultaneous looping of these function generators in practice does not lead to any errors.

Since, furthermore, often for the regulation of the direct beamformer equipment, coordinate converters 41, 4 2 and vector analyzers 17', 18' are already available for the formation of iA and uA, it is thereby possible to compensate the basic oscillation reactive load and its modulation economically without any reactive load regulation with additional hearing measuring devices.

Claims (12)

1. Method for reducing the grid feedback from a grid commutating direct inverter (I, II) which is connected to a three-phase alternating current grid (R, S, T), by equipping a grid-connected controllable reactive power compensator (5, 6, 7), characterized in that the desired value (ini<1>3) for the control (8) of the reactive current to be delivered from the compensator, becomes independent of the measured value of the reactive current formed by the product (/ i^/, costy=±Q^) of the value of the rectifier output current ( iA) and one from the inverter (I, II) control voltage (ust) or the direct inverter output voltage (U ) derived predetermined function (cosi);) of the direct inverter control angle ty (fig. 1) . 2. Method as specified in claim 1, characterized in that the desired value (in) is calculated as the difference (Aig-j^, difference term 14) between the product (/^/.cos^ig-^ and a corresponding product mean value (ig- ^) (fig- 1) • 3. Method as stated in claim 2, characterized in that the product mean value (ig^) is formed by smoothing (15) of the electrical quantity (ig-^) which corresponds to the product (/iA/.cosiJj) (fig. 1) . 4. Method as specified in claim 2, characterized in that the product mean value (ig]_) is calculated as the product of (multiplication term 22) the mean value of the direct inverter output current (i ) over a half-wave of the inverter output frequency, and the predetermined function (cosijj) of the corresponding mean value of the direct inverter steering angle ty (Fig. 2). 5. Method as stated in claim 4, characterized in that another orthogonal component of a current vector is formed respectively at a phase shift of approx. 90° of the measured value (i,) or the desired value (iA) of the direct inverter output current and another orthogonal component of a control vector from the control voltage (U ^.) or the direct inverter output voltage (L>A), and that the mean value of the contribution (in cL) from the direct inverter output current is calculated as the contribution of the current vector and the function (cos[ ty ) of the mean value for the direct inverter control angle ( ty ) from the contribution of the control vector. 6. Method as stated in one of the claims 2-5, characterized in that to the difference (Aig-^) between the product and the product mean value is added the product mean value (Iq-^ (fig- 3) in a weighted form for the desired value formation. 7. Method as specified in claim 1 for the simultaneous reduction of power pulsations in a direct current converter with a single-phase output, characterized in that a controllable compensator is used which is also designed to compensate for modulations of the active current, the product (/iA/.sin^= iw) is calculated from the direct inverter output current value and a predetermined function (sin ty) of the direct inverter control angle, and the difference (Aiw) is formed from the product (iw) and a product mean value (Iw) as the desired value for the control of the active current which must be provided by the compensator (fig. 4 and 5). 8. Control connection for generating the desired value for the equipment of a reactive power compensator to reduce the grid feedback from a direct-current converter connected to a three-phase alternating current grid in the method according to one of claims 1-7, characterized by a calculation link (9) with a value generator (10) which is imposed by the measured value (i,) or the desired value (iA) of the direct-inverter output current, a function generator (11) which is imposed by a quantity corresponding to the direct-inverter control angle ty and has a predetermined function characteristic, and a multiplication term (12) which is imposed by the output quantities from the value generator and the function generator and serve for product formation (Fig. 1). 9. Control coupling as stated in claim 8, characterized in that to form the predetermined function (cos ty) of the direct-inverter control angle ty, the function generator (11) is connected to another function generator (24) which receives the output voltage from the direct-inverter, and whose characteristic gives the relationship between steering angle and output voltage from the direct current converter (fig. 2). 10. Control link as stated in claim 8 or 9, characterized by an average value link (16, 16') connected downstream to the calculation link (9, 9') which forms a product average value: of the output quantity (igj» Fig. 1) or the input quantities (iA, uA ; fig. 2) from the calculation link and includes a difference term (14) incurred by the calculation link and the product mean value. 11. Control connection as stated in one of the claims 8-10 for a direct converter with multi-phase output, characterized in that for each phase (a, b, c) there is arranged a calculation connection (9a, 9b, 9c) with a value generator (10a, 10b, 10c) which is induced by the output current UAa, i^, iA(J from the rectifier phase output, a function generator (lia, 11b, lic) which is induced by a voltage corresponding to the steering angle ( tyat ty-^ i tyc) from the sub-inverter which works on this phase output, and a multiplication term (12a, 12b, 12c) which is incurred from the output quantities from the value generator and the function generator, and that the output quantities from the calculation links (9a, 9b, 9c) are supplied to an addition term (40) (fig. 6). 12. Control coupling' as specified in one of the claims 8-10, for controlling a reactive load compensator which is also intended for compensation of the modulation of the watt current, when using a direct inverter with a single-phase output, characterized in that for the equipment of the watt- current to be delivered, a multiplication term (31) is provided for the direct-converter output current (iA or iA) and the direct-converter output voltage (UA) or the output magnitude (sin ty) from a function generator which is induced by the steering angle (ty or Ugt), and that the output magnitude (i-w) from the multiplication term (31) and a corresponding mean-valued multiplication term output magnitude (Iw) formed in a mean value coupling (32, 32') are supplied to a difference term (34) at whose output the desired value (AiTwT) for the compensating reactive current to be supplied is tapped off (fig. 4 and 5).