Three phase core form transformers

(54) NAME Three-phase nuclear transformer (56) PRESENTING PUBLICATIONS: --- (57) SUMMARY: Three-phase nuclear transformer for a three-phase dual dynamic stabilizer system for compensating and reactive power, which here has a three-phase magnetic core, a number of primary windings 146 (142) corresponding to single phases, and secondary main winding (152, 154, 156) corresponding to the primary winding and two secondary auxiliary winding (A, AA; B, BB; C, CC) per phase, which Ar wound so as to have first and second (A2, AA2, B2 etc.) common bifilar parts. The first common bifilar portion (e.g., 152A1), the secondary main winding (152), the second common bifilar portion (152A2), and the primary winding (142) are arranged coaxially and radially adjacent to one another.

The primary windings are mutually connected. The secondary main windings and secondary auxiliary windings Ar are connected in branched Y-coupling, in which a secondary auxiliary winding in each of the M other two phases is electrically connected to the free end of each and every secondary secondary winding connected in Y .

The primary windings are continuously wound windings, the secondary main windings and the secondary auxiliary windings, preferably helically wound windings.

ALLF13Ö8 122 AA

<img file="SE459702B_D0001.tif" />

The numbers in brackets indicate the international identification code. INID code. Letters in clamps indicate international document code.

459 702 i

This invention relates to reactive power compensation (VAR) and, in particular, to a core coupling transformer used with dynamic stabilizers to provide stabilization of electric alternator or turbine generators for sub-synchronous resonance. in

Maximizing the reliability and efficiency of large electrical systems often requires reactive power (VAR) compensation. For example, industrial consumers typically use switchable shunt capacitors with conventional means to provide the reactive power required for their load. Since the use of capacitors results in better power factor and voltage support on loaded conductors, significant savings and increase in the efficiency of production processes can generally be achieved. Similarly, in power transmission, series capacitors and switchable shunt reactors and capacitors are used at high voltage power lines to increase transmission capacity and translational stability margins, providing voltage support and limiting sustained and transient overvoltages.

An important problem in the use of series condenser compensated transmission lines is the occurrence of sub-synchronous resonance. At sub-synchronous resonance, the transmission line and series capacitor have series resonance at a frequency below the system frequency, which is typically 50 Hz. Disturbances or faults can easily excite low frequency resonant currents. These low frequency resonant currents can affect the turbine generator sets which feed the transmission line. A mechanical resonance between the generator rotor, the turbine rotor and the shaft between the generator and the turbine can be excited by these low frequency resonant currents which, once started, continue to grow indefinitely. If they continue, the sub-synchronous resonance can destroy the machine equipment.

One way to influence sub-synchronous resonance is to periodically connect a reactive load to the generator outputs. In this case, reactors are switched on and off by means of inversely connected parallel triethorps. However, while such a circuit arrangement compensates for the transmission line and the generator, it can also cause undesired odd harmonic harmonic currents through the generator which are proportional to the size of the regulated reactor. Sometimes the conduction angle of the thyristor switches is reduced to reduce the reactive current to the generator, a condition that accentuates the odd harmonic currents in the stabilizer.

By arranging the thyristor-controlled reactors in a delta coupling, the triplet allows odd harmonic currents, for example third,

459 702 ninth, fifteenth overtones, essentially eliminated in balanced three-phase. Operation. However, other harmonic currents are not eliminated. Thus, other devices for eliminating odd harmonic currents other than the triplet are desirable.

In most electrical systems in which dynamic stabilizers are used, the stabilizer is typically connected to the generator over a multiphase coupling transformer to provide the appropriate level of drive voltage for the thyristors and associated reactors. This transformer, which is usually three-phase, can be either of the sheath type or of the core type. When using a core type transformer, primary winding, secondary main winding and secondary winding windings are wound around each leg of the transformer core as shown in Fig. 1. Two secondary winding windings, winding stumps, the secondary main winding and primary winding for each phase are arranged coaxially and radially adjacent to each other. order, from the core of the core.

The transformer used to connect the stabilizers to the generator should result in selective elimination of certain odd harmonic frequencies generated by phase-angle ignition of the thyristors in the stabilizers. Unfortunately, the leakage reactances of the transformer tend to prevent this elimination, especially when the reactance is a significant percentage of the effective reactance of the dynamic stabilizers. Due to the stabilizer operating cycle (dv3. the short periods of time during which the stabilizer is fully engaged) the nominal voltage of the transformer can be significantly less than the maximum voltperamperature of the stabilizer. For example, the required voltampere may be as low as 20% of the stabilizer's highest voltampere number. As a result, a situation arises in which the leakage reactance of the transformer substantially increases the reactor's reactivity when the thyristors are conductive. Therefore, it would be desirable to have a transformer so designed that the effect of its leakage reactance can be advantageously utilized for the selective elimination of the harmonic frequencies of currents occurring in dynamic stabilizers.

Current designs of nuclear transformers cannot economically provide the proper value for the leakage reactances of the windings which require for the selective elimination of subsynchronous resonance of the type described below. Namely, the value of the leakage reactance of the secondary windings must be about twice the value of the common leakage reactance of the secondary auxiliary windings with both values ​​counted positive. In addition, the ratio of the revolutions in the secondary auxiliary windings to the number of turns in the secondary main windings should not be stated as the number of turns between the primary windings and the secondary main windings Sr, since it is used primarily to determine the voltage change required to obtain the operating voltage required. for the stabilizers. Although specific turnover figures can be determined by one of ordinary skill in the art in FIG. 1 shown the arrangement, the value pfl of the leakage reactance of each secondary main winding with this arrangement becomes negative. Analysis can show that the leakage reaction value for the binder arrangement shown in Fig. 1 can be made positive. However, with this arrangement, the distance required between the windings to reach appropriate values ​​of the leakage reactance for each phase would be so large that it would exceed the capability of currently commercially available transformer manufacturing equipment. Therefore, it would be advantageous to have a winding arrangement for a grenade-Y core transformer, which permits manufacturing with existing manufacturing equipment and to achieve the blanket reactance values ​​required for selective elimination of harmonic currents.

According to the present invention, a three-phase core transformer having a three-phase magnetic core, a number of phases associated with the phases, and a number against the number, primary windings corresponding to the number of secondary main windings, comprise two secondary auxiliary windings per phase, so as to obtain first and second common bifilar portions, first common bifilar portion, the secondary head wrap, the second common bifilar portion and the primary winding are coaxial and radially adjacent to each other in said order starting from each leg of the magnetic core, the primary winding being delta-coupled, the secondary main winding and the secondary winding of the winding in branch Y, in which a secondary auxiliary winding in each of the other two phases is electrically connected to a free end of each of the Y-connected secondary main windings; so that all three phases occur in any combination of secondary main windings and secondary auxiliary windings, so that the leakage reactance of the secondary main winding and the common leakage reactance of the secondary, associated auxiliary windings for each phase have a positive value with the leakage reactance of the secondary main winding. greater than the value of the joint leakage reactance of the helical windings.

Preferably, the primary windings are delta-coupled. The secondary main windings and the secondary helical windings are connected in branch Y, wherein a secondary auxiliary winding of each of the other two phases is

459 702 electrically connected to the free end of each of the Y-connected secondary main windings. This means that all three phases occur in each combination of secondary head windings and secondary auxiliary windings. With this winding arrangement, the leakage reactance of the secondary main winding and the common leakage reactance of the secondary to the same winding connected for each phase are positive with the value of the leakage reactance of the secondary main winding selected to about twice the common leakage reactance of the auxiliary windings.

The transformer of the present invention is designed such that the ratio of the displacement number of the secondary main winding to the secondary auxiliary windings is equal to (iT? -1) / 2 + 0.01%. This rotates the phases of resonant currents in the dynamic stabilizers by about 30 ° relative to each other and about 15 ° relative to the phase of the connected secondary main winding. The combination of phase rotation and positive leakage reactance provides for the selective elimination of resonant currents generated in the dynamic stabilizers connected to the secondary auxiliary windings.

The primary windings may be continuously wound windings, the secondary main windings and the secondary winding windings being fully wound windings.

The invention is described in more detail below in connection with the exemplary embodiment shown in the accompanying drawings.

Fig. 1 shows schematically the winding arrangement used in the prior art with the winding shown in section, Fig. 2 is a schematic diagram of an electrical installation in which the present invention is applied; Fig. 3 is a schematic diagram of an embodiment according to the present invention in which a double dynamic Figure 4 illustrates voltage waveforms of the thyristor controlled reactors shown in Figure 3 for a transformer without leakage reactance; 5 Figure 3 is a circuit diagram of a phase of the three-phase transformer of Figure 3; a cross-section along the line VII-VII in Fig. 7.

Fig. 2 shows an electromechanical network 10 consisting of a transmission network 12, a turbine 14 and a generator 16. The network 10 is a

459 702 multi-phase alternating current system, usually three phases. A dynamic stabilizer 20 is connected to the network 10 over a coupling transformer 22. Since the system voltage in the electrical network is typically much higher than that desired for operating the dynamic stabilizer 20, the coupling transformer acts as a network transformer for the system voltage to the appropriate operating level for those of the stabilizer 20. hearing components. Transformer 22 also performs other functions described below.

A control unit 24 is provided for controlling the compensation provided by the dynamic stabilizer 20. The controller 24 senses the rotational speed of the generator 16 via a speed sensor, for example a tachometer 26, to determine the stabilization of the generator 16 required and which, in response to these requirements, generates a control signal which sets the lead angle at the ignition of the thyristors in the stabilizer 20 the reactive currents that must be supplied to the generator 16 as compensation. Although dynamic stabilizer 20 and transformer 22 are shown as connected to an electrical grid of the type illustrated, it is clear that the particular shape of the electrical grid is not critical to operation of the present invention. The dynamic stabilizer and associated switching transformer can provide selective elimination of harmonic currents even in connection with other forms of electrical networks. Since applications for the present invention are most likely to occur with three-phase networks, the embodiment of the invention is shown here in three-phase form. The three electrical phases are denoted by the letters a, b, and c and have substantially balanced alternating voltages and alternating current which are offset relative to each other by about 120 electrical degrees (2 «/ 3 radians) relative to each other. The windings on the coupling transformer belonging to a particular phase are marked with the designation for that phase. The use of the terms phase and degree refers to electrical phase and electrical degree respectively.

Operating voltage, nominal kVA number and isolation level of the transformer provide the physical parameters for the core and windings. The transformer rating data determines the current flow in the windings which in turn gives the size of conductors and windings. In addition, the voltage and kVA for the transformer determine the optimal voltage per revolution, which gives the number of turns and the size of the core. The core is selected so that it has sufficient area to limit the induction to an acceptable value, usually 18,000 gauss or less, so that core saturation is prevented and error-free transformation is obtained. In general, the volt per revolution becomes larger the higher the kVA is and the larger

459 702 will be the core size.

These parameters, as well as the operating frequency, affect the value of the leakage reactance for a given winding. However, when the operating parameters of a particular transformer according to the invention have been specified, the values ​​of the leakage reactances required for the transformer can easily be determined with known computational technique. Therefore, in the following description of the transformer, no special data is given regarding winding dimensions, conductor size, spacing between windings and other dimensional data. Transformers with winding arrangements of this type will generally have a rated kVA over 1000 kVA (1 MVA) and rated voltages above 10000 V alternating voltage.

Fig. 3 shows a general diagram of a dual dynamic stabilizer system 100. This system includes two dynamic stabilizers 110 and 120 and a switching transformer 130. The transformer 130 has a primary side 140 and a secondary side 150, each side having a plurality of multicolored windings. The primary windings 142, 144 and 146 are electrically connected in a delta connection and have the outlet voltages v, v and v respectively with respect to ground 101. The primary side 140 is electrically connected to a generator not shown via the conductors 141, 143 and 145, which are electrically connected to the horns in the delta winding of the primary windings. From the system 100, the electrical system for its stabilization is supplied to the currents in_g, i ^ and

The secondary side 150 of the transformer 130 consists of two groups of multi-winding windings, namely secondary main windings 152, 154 and 156 and corresponding secondary auxiliary windings 152A, 152AA, 1548, 154BB, 156C and 156CC. Two auxiliary windings for each secondary main winding, for example the windings 152A and 152AA for the winding 152, are also known as tertiary or zigzag windings and in the transformer wrap around the same fighter as its associated secondary main winding. Each auxiliary winding pair 152A and 152AA, 154B and 154BB, 156C and 156CC are wound adjacent to each other to maximize the magnetic coupling to each other. These pairs are also known as bifilar winds. For the following detailed reasons, each pair of secondary auxiliary windings is wound so as to obtain first and second common bifilar portions, which are electrically connected in series. Thus, the winding 152A is formed by the winding portion 152A1 and the winding portion 152A2 and the winding 152AA is formed by the winding portions 152AA1 and 152AA2. Similarly, the winding 154B is formed by the winding portions 154B1 and 154B2, the winding 154BB by the winding portions 154BB1 and 154BB2, the winding 156C by the winding portions 156C1 and 156C2, and the winding 156CC by the winding portions.

459 702

156CC1 and 156CC2.

The secondary main winding 152, 154 and 156 are electrically connected to each other in Y connection with the common terminal 158 connected to ground 101. Two secondary auxiliary windings are electrically connected to each secondary main winding, each electrically coupled combination of secondary main winding and two secondary auxiliary windings having equal or matching polarities as shown with the polarity markings in Fig. 3. The two connected secondary auxiliary windings are selected so that all three electrical phases occur in each winding combination. In Fig. 3, the secondary main windings 152, 154 and 156 belong to phases a, b and c, respectively. Accordingly, a secondary auxiliary winding belonging to phase b - 154BB - and one belonging to phase c - 156C - should be electrically connected to the secondary the main winding 152 and in a similar manner to the secondary winding 154 and 156. Each secondary auxiliary winding is in turn electrically connected to an individual phase of one of the two dynamic stabilizers. The connection arrangement of the secondary winding, main winding and auxiliary winding, and the stabilizers are such that the corresponding phases in each stabilizer and secondary main winding are the same. For the stabilizer 110, the three phases are designated a ', b * and c', those for stabilizer 120 a, b and c '. Thus, the secondary main winding 152, which is of phase a, is electrically connected to phases a¹and a "of the stabilizers 110 and 120, respectively, via auxiliary windings 154BB and 156C and conductors 160 and 162, respectively.

For the secondary main winding 154 and phases b * and b of the stabilizers 110 and 120, respectively, the connections are made via the secondary auxiliary winding 156CC and 152A and conductors 166 and 164. Similarly, for the secondary main winding 156 and phases c 'and c of the stabilizers 110 respectively 120 connections made via the secondary auxiliary windings 152AA and 154B and conductors 168 and 170, respectively.

Voltages and currents for the stabilizer-110 are v_A', v &amp;' and ν_βand i · and i for the stabilizer 120 v_A, v ^ and v_c respectively, i, i ^, i, These measured voltages are in relation to ground 101 and the current is as shown by the arrows. The stabilizers 110 and 120 are essentially identical. Each includes a variety of variable reactants that are electrically coupled. In the stabilizer 110 are the variable reactants 111, .113 and 115, while the stabilizer 120 has the reactants 121, 123 and 125. The thyristor pairs, 111A and B, 113A and B, 115A and B, 121A and B, 123A and B, and 125A and B, which are connected in series with their respective reactances, are used to control the reactive currents and create

459 702 thus a variable reactance. The thyristor pairs are coupled in reverse parallel coupling / the angle allows current control during both the negative and positive parts of the system voltages.

When the plant is in operation, a tachometer is used to detect the mechanically resonant oscillating velocity of the generator rotors superimposed on the normal rotational speed of the rotor. In the typical case, the normal speed of the rotam is 3000 rpm while the oscillation frequency can be about 20 Hz. The signal representing the rotational speed 3000 rpm is filtered out with conventional signal processing means and a signal is obtained with a frequency corresponding to the sub-synchronous mechanical resonant frequency of the rotor. This processed rotor speed signal controls the angle of heating of the thyristor pairs of the stabilizers and acts to increase or decrease the value of the reactants. As the rotor speed decreases, the conduction angle increases and the reactive power to the stabilizer increases so that the stabilizers appear to have received less reactance. As the rotor speed increases, the thyristor pairs lead angle decreases, which reduces the reactive power to the stabilizer and makes the stabilizer reactance appear to have increased.

The mechanical sub-synchronous frequency is related to the frequency of the sub-resonant currents in the electrical part of the system in that it is essentially the difference between the normal operating frequency of the system and the frequency of the sub-synchronous resonant currents therein. For example, if a typical system frequency is 50 Hz and a typical value of the frequency of sub-synchronous. resonant currents circulating in the system are 30 Hz, the expected mechanical oscillation frequency would be 50 Hz - 30 Hz or 20 Hz.

Fig. 4 illustrates the relationship between the currents and voltages of the variable reactors in each of the stabilizers 110 and 120. The sinusoidal waves e._A', e ^¹, e_c', e_A, e ^ and e_c are the voltages over their respective reactances. For the stabilizer 110 is^eA ^{= v}b * -^vc ' ^eb ' ^{= v}c^va% ^{= v}a'^vb

The corresponding double-primed voltages are similarly obtained. The arrows represent current pulses P. For each complete period, 6 current pulses are generated in each stabilizer or a total of 12 current pulses for both. For this reason, this stabilizer arrangement is called the 12-pulse system. If the thyristors in each stabilizer are turned on in the usual way so that all the reactors have the same conduction or phase angle, then a current pulse P will occur,

459 702 which corresponds to each zero voltage passage Q with the current pulse P exactly symmetrical about the zero point passage Q.

The current pulses P produce harmonic currents. Due to the symmetry of the current pulses P, they contain no uniform harmonics and the odd harmonics multiples of 3 will circulate within each stabilizer's coupled reactances. The frequencies produced by the odd harmonic currents correspond to series 1, 5, 7, 11, 13,17,19 .... Π where n is an odd number that is not divisible by 3. The harmonic frequencies corresponding to n = 1, 7, 13, 19 ... (add 6 each bunch) produce balanced multi-phase currents with phase rotation a, b, c, while the frequencies corresponding to n - 5, 11, 17 ... (add 6 each thread) produces balanced multi-phase currents with phase rotation a, c, b. According to the principle of symmetrical electrical components, these two stream groups are referred to as positive sequence and negative sequence currents respectively. As is well known to those skilled in the art, the generator does not supply harmonic currents with zero sequence due to the thyristor control.

In determining the efficiency of transformer 130 with respect to the elimination of unwanted harmonic currents generated in the stabilizer, the situation is first analyzed from the point of view of the ideal transformer. In the stabilizer 110, the voltages and currents for the three phases a ', b * and c * are offset 120 ° and in the corresponding manner for the stabilizer 120, the voltages and currents of phases a, b and c. The geometry of the waveform suggests that elimination of selected harmonic currents can be achieved if there is a phase shift of 30 ° between the voltage of phases a ', b * and c¹ of the stabilizer 110 and phases a ", b and c of stabilizer 120. This phase shift is determined by the speed ratios between the secondary auxiliary windings and the secondary main windings and by the polarities of the transformer windings.

The primary delta-coupled windings have a number of turns equal to q times the number of turns in the secondary main windings. For equal voltages on primary winding and output voltage on secondary winding, the theoretical value of q = 3 / V2L is The number of turns of the secondary winding should be k times the number of turns of the secondary main winding. The theoretical value of k is (JJ - 1) / 2.1 In practice, the actual values ​​of q and k can be within + 0.1% from their theoretical values. For the polarities of the transformer windings indicated in Fig. 3, the single voltages designated voltages are 30 ° ahead of the double primers, designated voltages. With respect to the voltages of the secondary main windings, the single prime voltages are about 15 ° after them in phase, whereas

459 702 ίο the double-priming voltages are at risk by about the same amount. The reverse can also be applied provided that the control signals to the thyristors are also shifted accordingly.

Given given values ​​of q and k, the amplitude of the sinusoidal voltages v_q, odi v_Q of the primary windings 142, 144 and 146, respectively, and the single and double prime voltages are assumed to be equal, ie a total voltage transformation of 1: 1. If this is not the case, a scale factor is included in the value of q with the male view of the amplitude differences between the primary voltage and the secondary voltage. For example, if q = n (3/42), the primary voltage would be n times the secondary foreigners. Furthermore, these values ​​of q and k also apply to a transformer with leakage reactance.

The coupling transformer of the present invention, although generally realized as a three-phase transformer, can be regarded as three single-phase transformers for analytical purposes. Fig. 5 shows an equivalent circuit for phase a transformer. Corresponding circuits apply to the transformers in phases b and c. In practice, the two secondary auxiliary windings 360 and 380 are wound bifilar and can therefore be considered 100% coupled. For each phase there are four windings, which generally result in fairly complex equivalent circuits. In this case, since the two secondary auxiliary windings have strong coupling to each other, the simpler equivalent scheme of Fig. 5 can be used. This is a special application of an equivalent three-winding connection.

Transformer 300 has four pairs of terminals, terminal winding 420 for primary winding 320, terminal 440 for secondary main winding 340, and terminals 460 and 480 for secondary auxiliary winding 360 and 380, respectively. The circuit elements within the dashed lines of Fig. 5 constitute a phase equivalent transformer circuit . The equivalent circuit includes four ideal transformers:

the transformer 322 belonging to the primary winding 320 and having a speed ratio q: 1, the transformer 340 belonging to the secondary main winding 340 and meeting the speed ratio 1: 1, the transformers 350 and 370 belonging to the secondary auxiliary winding 360 and 380 respectively.

Of the ideal transformers 350 and 370, transformer 350 has three windings, 352, 354 and 356, each having a 1: 1 turnover

459 702 between any pair of the three windings. Transformer 370 has a turnover number pS k; l.

A study of the equivalent circuit shows that elimination of ampere turns must be found between the four windings 320, 340, 360 and 380 of transformer 300 as well as the required idle voltage conditions. The influence of the leakage reactance of transformer 300 is represented by the inductances L L, L and L ^, where

Lq = the leakage reactance of the primary winding 320,

L = the leakage reactance of the secondary main winding 340 and

L ^ = the common leakage reactance of the secondary auxiliary windings 360 and 380.

Lt can also be considered as the leakage reactance of either of the bifilially wound secondary auxiliary windings. Due to the 100% coupling of bifilar windings, these specifications are extensible. If L = 2l__k or if there is no leakage reaction, it happens that perfect elimination of the harmonic currents of series 5, 7, 17,

19, 21, 31 .... Π, p occurs, where: n = 5 + z, p = 7 + z and z = multiples of 12.

Since it is not possible to manufacture a transformer without leakage reactance, a transformer according to the present invention should be designed in such a way that the leakage reactance L = 2L_k with a variation of about 10% up or down.

In this case, Fig. 6 illustrates the effect of these blanket reactances on the voltages in the stabilizer. This figure shows the voltage waveforms eoe ^ and e_c* for the reactants 111, 113 and 115, respectively, of the stabilizer 110. Similar waveforms are generated for the stabilizer 120. In Fig. 6, the fifth harmonic tone added to the voltage waveforms of the_A', e ^' and e_c'. Study of these waveforms shows the half-wave symmetry, dv3. e J (wt) = e '(-wt), which occurs in the waveforms of Fig. 4. This means that the current pulses generated in the stabilizers will be symmetrical and properly phased to provide currents to the coupling transformer, which will be selectively eliminated. . Addition of other odd harmonic tones will result in a significantly more distorted voltage waveform, but one that still has half-wave symmetry. Thus, sine wave voltages are not required to achieve the desired elimination of harmonic currents.

Figures 7 and 8 show a winding arrangement for the coupling transformer 130 which realizes the positive values ​​of both L and L

459 702 for harmonic elimination. Transformer 130 has a three-legged magnate core 710. Core 710 is laminated and consists of three pS spaced apart legs 712, 714 and 716 which are joined at the ends of an upper yoke 718 and a lower yoke 720. For simplicity, it can be said that legs 712,714 and 716 belong to phases a, b and c, respectively.

In general, the windings belonging to the same phase are wound on insulators 722 and positioned around each leg before the upper yoke is attached. An insulating board 724 is inserted between adjacent windings for the different phases. Since the winding arrangement on each leg of the core for each phase is essentially the same, only the arrangement for phase a is described. The two secondary auxiliary windings per phase are wound so as to obtain first and second common bifilar portions. For phase a, these first and second common bifilar portions are represented by the winding portions 152A1 and 152AA1 and the winding portions 152A2 and 152AA2, respectively. The same arrangement applies to phases b and c.

The first common bifilar portion, the winding portions I52A1 and 152AA1, the secondary main winding portion 152, the second common bifilar portion, the winding portions 152A2 and 152AA2, and the primary winding 142 are located coaxially and radially adjacent to each other in order of the bores of the core 712. the other two phases are arranged in a similar manner. The primary windings 142, 144 and 146 are mutually coupled. The secondary auxiliary winding 152,154 and 156 and the secondary auxiliary winding 152A, 152AA, 154B, 154BB, 156C and 156CC are connected in a branched Y, wherein a secondary auxiliary winding of each of the other two phases is electrically connected to a free end of all the Y-linked secondary main windings, which means that all three phases occur in each combination of secondary main windings and secondary auxiliary windings. By splitting the two secondary auxiliary windings into the first and second common bifilar portions and arranging the latter between the secondary main winding and the primary winding for each phase, the leakage reactances of the secondary main winding and the common leakage reactance of the secondary secondary windings connected to it are obtained. positive value. With this arrangement which gives positive value to the reactances L and the secondary windings can be designed so that the leakage reactance of the secondary main winding becomes about twice the common leakage reactance of the secondary auxiliary windings. This winding arrangement also enables existing manufacturing equipment to be used in the construction of the transformer. The secondary windings

459 702 are thus wound so that they meet the speed ratio (\ TT-1) / 2. When these two conditions are met, the transformer, when connected as described herein, will provide the desired elimination of selected harmonic frequencies.

Preferably, the primary windings are continuously wound windings and the secondary main windings and secondary auxiliary windings are fully wound windings.

459 702