Method and Apparatus for Direct Torque Control of a Three-Phase Machine

1. Field of the Invention

The invention relates to the field of power electronics. It is based on a method for Direct Torque Control (=DTC) according to the descriptive part of the first claim.

2. Discussion of Background

Such a method and a corresponding apparatus are described in the article "Direkte Drehmomentregelung yon Drehstromantrieben" ›Direct torque control of three-phase drives!, ABB Technik 3/1995, pages 19-24. This method is based on so-called "direct self-control", as is described in EP-B1-0 179 356. The DTC method represents a simple, reliable and highly dynamic torque control method for three-phase machines. In principle, it comprises the following steps:

Calculation of an actual flux value from a converter voltage

Calculation of an actual torque value

Comparison of the actual flux value and the actual torque value with corresponding reference values by means of a hysteresis controller

Formation of an optimum control vector for the switches in the converter on the basis of the hysteresis controller.

In order to protect the machine against excessively steep voltage flanks, which may lead to damage to the insulations, it is desirable to connect a harmonic filter between the converter and the motor. A control method is specified, for example, in EP-A2-0 595 319 for a system of pulse-duration-modulated converters with a harmonic filter. The filter capacitance and the filter inductance represent two integrators in control engineering terms. There is thus a 180.degree. phase shift between the motor voltage and the converter voltage at the switching frequency, in the frequency band. This means that the signals required for the DTC method are present with an incorrect mathematical sign. Because of the 180.degree. phase shift, caused by the harmonic filter, between the motor voltage and the converter voltage, the known DTC method can no longer be applied directly to an arrangement with a harmonic filter. In addition, disturbing oscillations occur because of the filter resonance.

Accordingly, one object of the invention is to provide a novel method and apparatus for direct torque control for a three-phase machine fed via a harmonic filter, which method manages as far as possible without major adaptation to already known methods and apparatuses. In addition, the oscillations caused by the filter are intended to be damped effectively.

This object is achieved by the features of the first claim for a method of the type mentioned initially.

The essence of the invention is thus that a current flowing through the filter and/or a filter output voltage are/is measured, and the reference flux value and the reference torque value are corrected on the basis of the measured voltage and/or of the measured current.

The correction comprises a transformation of the filter output voltage into a first component at right angles to a stator flux and into a second component parallel to the stator flux as well as a reduction in the reference torque and flux values in proportion to the oscillating part of the filter output voltage.

The oscillations caused by the harmonic filter can be effectively damped in this way.

The components can be calculated in various ways. Instead of a filter output voltage, a filter current can also be measured, and a filter voltage can be calculated from this. These variants are the subject matter of dependent claims. In addition, an apparatus for carrying out a method according to the invention is also specified.

Other exemplary embodiments result from the corresponding dependent claims.

The advantage of the design according to the invention is that the three-phase machine can be protected by the use of a harmonic filter against excessively steep voltage flanks which can lead to damage to the insulation, and, nevertheless, that essential parts of the DTC method and of a corresponding apparatus can be used without any major change.

In addition, oscillations caused by the filter can be damped effectively.

A more complete appreciation of the invention and many of the attendant advantages thereof will be readily obtained as the same becomes better understood by reference to the following detailed description when considered in connection with the accompanying drawings, wherein:

FIG. 1 shows a circuit arrangement with a three-phase machine fed from a converter via a harmonic filter;

FIG. 2 shows a variant of FIG. 1;

FIG. 3 shows a block diagram of an apparatus according to the invent ion, based on a first exemplary embodiment;

FIG. 4 shows a block diagram of an apparatus according to the invention, based on a second exemplary embodiment;

FIG. 5 shows a block diagram of an apparatus according to the invention, based on a third exemplary embodiment;

FIG. 6 shows a block diagram of an apparatus according to the invention, based on a fourth exemplary embodiment;

FIG. 7 shows a block diagram of an apparatus according to the invention, based on a fifth exemplary embodiment.

The designations used in the drawings and their meanings are listed in summary form in the List of Designations.

Referring now to the drawings, wherein like reference numerals designate identical or corresponding parts through the several views used, FIG. 1 shows a circuit arrangement with a three-phase machine 1 which is fed from a converter 2 via a harmonic filter comprising a three-phase network with filter inductances L.sub.fil and filter capacitances C.sub.fil. The capacitances of the filter are combined to form a star point, which is not connected to the intermediate circuit. The intermediate circuit is represented by the intermediate-circuit capacitor C.sub.zk. A DC voltage U.sub.DC is applied to the intermediate-circuit capacitor. A torque controller 3 is provided for torque control of the motor 1. This is fed, according to FIG. 1, from the intermediate-circuit voltage U.sub.DC, from a current i.sub.inv emitted by the converter, and from a filter output voltage u.sub.fil. In the variant according to FIG. 2, a filter current i.sub.fil may also be measured instead of the filter output voltage u.sub.fil. The converter 2 may be either a two-point invertor or a three-point invertor.

Underlined variables in each case designate vectors in the following text. These may be transformed into vectors from the measured conductor currents or conductor voltages in the phases, using the known 120.degree./90.degree. transformation.

The torque controller 3 uses the measured variables as well as a reference flux value .psi..sub.ref and a reference torque value T.sub.ref to calculate an optimum control vector S (S.sub.A, S.sub.B, S.sub.C) which is used for driving the switches in the converter 2.

FIGS. 3-7 show block diagrams of various variants of torque controllers 3. That part of the regulator which is above the dashed line corresponds to the already known DTC controllers. The lower part comprises the adaptations according to the invention.

In the following text, the fundamental principle of DTC control will be discussed briefly once again, and the problems on which the invention is based will be explained:

In the DTC method, an actual flux value .psi..sub.act is calculated from a converter output voltage u.sub.inv and a correction variable. The correction variable corresponds, to a first approximation, to the voltage drop R.sub.s .multidot.i.sub.s across the copper resistance of the stator. The calculation is much more accurate if a motor model is used as an aid. However, this is not important to the invention, so that further reference at this point will be to the said voltage drop.

The actual flux value is thus given by: .psi..sub.act=u.sub.inv -R.sub.s i.sub.s)dt. The converter output voltage u.sub.inv is in this case formed with the aid of a voltage forming element 8 from the intermediate-circuit voltage U.sub.DC and the actual control vector S. An actual torque value T.sub.act is subsequently formed from the calculated actual flux value .psi..sub.act and a current i.sub.inv emitted by the converter. The current i.sub.inv is obtained by measuring two phase currents and by transformation into a vector. The actual torque value T.sub.act is obtained, finally, by forming the vector product of the current i.sub.inv and of the flux .psi..sub.act. In addition, the magnitude .psi..sub.act of the actual flux value .psi..sub.act is also formed. T.sub.act and .psi..sub.act are subsequently compared, by means of a hysteresis controller 4, with corresponding externally predetermined reference values .psi..sub.ref and T.sub.ref. The result of this comparison and the result of a vector assigner 12, which uses the actual flux value .psi..sub.act to determine an actual flux .vector, are supplied to a logic circuit 13 which determines an optimum control vector S (S.sub.A, S.sub.B, S.sub.C) for the converter.

However, if a harmonic filter is now connected between the converter and the motor as in the case of the invention, then, on the one hand, there is a 180.degree. phase shift in the current and voltage at the switching frequency because of the double integration of the filter, so that the DTC method can no longer be applied directly. On the other hand, the filter causes resonant oscillations, which are intended to be damped. The known method and the known apparatus must therefore be adapted, as is described in the following text.

The adaptation essentially comprises a current flowing through the filter and/or a filter output voltage being measured, and the reference flux value, or its magnitude, and the reference torque value being corrected on the basis of the measured filter variables. In particular, they are reduced in proportion to an oscillating part of the filter output voltage. There are various variants for calculating these oscillating disturbance variables:

A first exemplary embodiment (FIG. 3) comprises measurement of the filter output voltage on each phase and conversion, for example by means of 120.degree./90.degree. transformation, into a vector u.sub.fil. In addition, the product i.sub.inv L.sub.fil is subtracted from the actual flux value .psi..sub.act. A multiplier 5 and a subtractor 7 are provided for this purpose. In this way, a variable .psi..sub.s is obtained, which corresponds to a stator flux. The voltage u.sub.fil is now split into a component u.sub.fild parallel to .psi..sub.s and into a component u.sub.filq at right angles to .psi..sub.s. A vector product forming element 9 and a scalar product forming element 10 are provided for this purpose, which multiply the voltage u.sub.fil by .psi..sub.s. The component u.sub.filq is produced from the vector product, and the component u.sub.fild from the scalar product. u.sub.fild represents just that proportion of the filter output voltage causing the oscillations and is therefore multiplied in a further multiplier 5 by the constant K.sub.1, is supplied to a limiter 6 and is subtracted from the reference flux value .psi..sub.ref in a subtractor 7. The oscillating part of u.sub.filq must also be formed, in contrast. This is done according to FIG. 3 by subtracting a constant value .O slashed..multidot..psi..sub.s from u.sub.filq. .omega. in this case corresponds to the angular frequency of the stator flux. In the steady state, u.sub.filq should correspond precisely to this value, so that the difference u.sub.filq =.omega..multidot..psi..sub.s thus corresponds just to the oscillating part .DELTA.u.sub.filq of u.sub.filq. This value is likewise multiplied in a multiplier 5 by a second constant K.sub.2, and is supplied to a limiter 6. Finally, the reference torque value is reduced by means of a further subtractor 7, using the output of this limiter 6.

The essence of the method and of the apparatus according to the invention is thus based on the fact that the influence of the harmonic filter is included in the control system, in that those elements of the filter output voltage which cause the oscillations are first of all damped and a virtual converter flux and a converter moment, so to speak, are controlled as actual variables instead of directly controlling variables which correspond to the motor flux or motor moment upstream of the filter.

A second exemplary embodiment (FIG. 4) differs from the first in that the oscillating parts of the components u.sub.fild and u.sub.filq are obtained by means of high-pass filtering. Two high-pass filters are provided for this purpose and are connected upstream of the multipliers 5, after the vector and scalar product formation 9, 10.

In the case of a third exemplary embodiment (FIG. 5), a filter current i.sub.fil is measured instead of a filter output voltage u.sub.fil. FIG. 5 is thus based on FIG. 2. i.sub.fil is obtained, for example, by measuring two currents flowing through the filter capacitors C.sub.fil, and by converting these currents into a vector. Two currents are sufficient, since the sum of all the currents must be zero because of the star connection of the capacitors. The voltage u.sub.fil across the capacitor can be calculated from the current i.sub.fil by means of limited integration. A limited integrator 15 is provided for this reason in FIG. 5 at the input of i.sub.fil. Otherwise, the arrangement in FIG. 5 corresponds to that in FIG. 4.

The variant according to FIG. 3 can, of course, also be applied to the third exemplary embodiment. According to a fourth exemplary embodiment (FIG. 6), the integration can also be carried out after the scalar and vector product formation.

Finally, FIG. 7 shows an exemplary embodiment based on FIG. 2, in which the integration for the component formation is carried out, on the one hand, before the scalar product formation 10 and, on the other hand, after the vector product formation 9. Since the current i.sub.fil should be parallel to .psi..sub.s from the start, the remaining part of a vector product represents just the element causing oscillations. The situation with the scalar product is the converse of this, since the voltage multiplied by the .psi..sub.s should result in zero here. For this reason, the voltage is calculated first for the scalar product formation, and not until after this for the vector product formation.

The invention has the advantage that the known DTC method and the known DTC apparatus can also continue to be used essentially for an arrangement having a harmonic filter. The invention makes available an addition which is suitable for use with a harmonic filter and leaves the known part virtually undisturbed. An existing system can therefore be extended in a simple manner without having to dispense with the advantages of a harmonic filter.

Obviously, numerous modifications and variations of the present invention are possible in light of the above teachings. It is therefore to be understood that within the scope of the appended claims, the invention may be practiced otherwise than as specifically described herein.