METHOD AND APPARATUS FOR DETERMINING A SATURATION CHARACTERISTIC OF A SYNCHRONOUS RELUCTANCE MACHINE

Abstract

A method for determining a saturation characteristic of a synchronous reluctance machine includes applying with a pulse inverter a voltage sequence to a stator of the synchronous reluctance machine, wherein the voltage sequence introduces stator fluxes and is applied such that torques acting on a rotor of the synchronous reluctance machine cancel each other out during the application of the voltage sequence. Electrical currents resulting from the stator fluxes are measured and the saturation characteristic is determined based on the stator fluxes and the measured electrical currents.

Claims

1.-10. (canceled)

11. A method for determining a saturation characteristic of a synchronous reluctance machine, comprising: applying with a pulse inverter a voltage sequence to a stator of the synchronous reluctance machine, with the voltage sequence introducing stator fluxes and being applied such that torques acting on a rotor of the synchronous reluctance machine cancel each other out during the application of the voltage sequence; measuring electrical currents resulting from the stator fluxes; and determining the saturation characteristic based on the stator fluxes and the measured electrical currents.

12. The method of claim 11, further comprising: before the application of the voltage sequence, determining a position of the rotor; and initially aligning a stator-fixed coordinate system to the determined position.

13. The method of claim 11, further comprising: when the voltage sequence is applied, outputting a plurality of voltage indicators, and determining for each of the voltage indicators a measured value for the electrical current.

14. The method of claim 13, wherein four voltage indicators are outputted when the voltage sequence is applied.

15. The method of claim 13, wherein eight voltage indicators are outputted when the voltage sequence is applied.

16. The method of claim 13, further comprising determining a pulse duration during which the voltage indicators are outputted as a function of an estimated deflection of the rotor.

17. The method of claim 11, wherein the voltage sequence is specified such that the resulting stator fluxes are equidistant.

18. The method of claim 11, wherein the voltage sequence is specified as a function of known saturation characteristics of the synchronous reluctance machine.

19. The method of claim 11, further comprising: determining the electrical currents for a quadrant of the rotor, and determining the saturation characteristic based on a known symmetry of the rotor.

20. Apparatus for determining a saturation characteristic of a synchronous reluctance machine, comprising: a pulse inverter for applying a voltage sequence to a stator of the synchronous reluctance machine and for impressing stator fluxes, wherein the pulse inverter is designed to apply the voltage sequence so that torques acting on a rotor of the synchronous reluctance machine cancel each other out during the application of the voltage sequence; and a measuring facility configured to measure electrical currents resulting from the stator fluxes, wherein the apparatus is configured to determine the saturation characteristic based on the stator fluxes and the measured currents.

Description

[0026] The invention is explained in more detail hereinafter with reference to preferred exemplary embodiments and with reference to the accompanying diagrams. The diagrams show:

[0027] FIG. 1 A diagrammatic view of a synchronous reluctance machine and an apparatus for determining a saturation characteristic of the synchronous reluctance machine;

[0028] FIG. 2 A diagrammatic view of four flux pointers as a result of voltage pulses which are applied to a stator of the synchronous reluctance machine, and the currents resulting therefrom;

[0029] FIG. 3 Torques acting on the rotor of the synchronous reluctance machine with four voltage pulses and the resulting curve of shaft speed and shaft position as a function of time;

[0030] FIG. 4 Torques acting on the rotor of the synchronous reluctance machine with eight voltage pulses and the resulting curve of shaft speed and shaft position as a function of time;

[0031] FIG. 5 A time profile of the flux and the current in the case of an applied voltage pulse;

[0032] FIG. 6 A diagrammatic view of the specification of flux pulses in the direction of the transverse axis of the rotor on tracks;

[0033] FIG. 7 A diagrammatic flow chart for identifying the rotor position and the adjustment of a coordinate system;

[0034] FIG. 8 An equidistant specification of the proportions of the flux with regard to the axes d and q;

[0035] FIG. 9 The resulting mapping of the current measurement values with the specification according to FIG. 8;

[0036] FIG. 10 A specification of the proportions of the flux according to a priori knowledge of typical flux saturation curves;

[0037] FIG. 11 The resulting mapping of the current measurement values with the specification according to FIG. 10; and

[0038] FIG. 12 A diagrammatic flow chart of a method for determining the saturation characteristic of the synchronous reluctance machine.

[0039] In the figures, elements which are identical or have identical functions are provided with the same reference characters.

[0040] FIG. 1 shows a highly simplified illustration of a synchronous reluctance machine 1 having a stator 2 and a rotor 3. The synchronous reluctance machine 1 can be controlled with a pulse inverter 4.

[0041] Synchronous reluctance machines 1 without a damper cage have an inherently non-linear relationship between stator flux and stator current. This saturation characteristic of the synchronous reluctance machine 1 is to be determined in order to operate the synchronous reluctance machine in a robust and performant manner without a rotor position sensor. For this purpose, voltage sequences are applied to a stator 2 of the synchronous reluctance machine with the pulse inverter 4. This results in stator fluxes in the stator 2. Furthermore, electrical currents i.sub.1 to i.sub.4 resulting from the stator fluxes Ψ.sub.1 to Ψ.sub.4 are measured. The currents i.sub.1 to i.sub.4 are measured by means of a measuring facility 5, The saturation characteristic of the synchronous reluctance machines can then be derived therefrom. The measuring facility 5 and the pulse inverter 4 together form an apparatus 6.

[0042] First, generation of the voltage sequence or sub-sequence is explained. A sub-sequence may comprise a short sequence of applied voltage space vectors. The torques acting on the rotor 3 are to be determined in such a way that they cancel each other out in the second time integral. The connected voltage space vectors are based on the independently specified flux value as the “generator value”. The result of the sub-sequence is the measured value for the current as a function of the specified generator value of the flux. The magnetic symmetries present in the synchronous reluctance machine 1 can be used for the generation of the sub-sequence: the mapping of the stator currents to the stator fluxes and, in reverse, the fluxes to the currents are axisymmetric with respect to the rotor axes.

[0043] FIG. 2 shows the resulting currents i.sub.1, i.sub.2, i.sub.3, and i.sub.4 for four applied flux pointers Ψ.sub.1, Ψ.sub.2, Ψ.sub.3, and Ψ.sub.4 which are constantly equal in magnitude and axisymmetric directions. The stator fluxes to Ψ.sub.1 to Ψ.sub.4 or the flux pointers and the currents i.sub.1 to i.sub.4 are shown with respect to the axes d and q of the synchronous reluctance machine. With given symmetries, the associated resulting torques M.sub.1 to M.sub.4 are likewise constantly equal in magnitude and changing sign: M.sub.1=M.sub.4=−M.sub.2=−M.sub.3.

[0044] In an initially simplified view, the torques M.sub.1 to M.sub.4 acting on the rotor due to the switched pulses are ideal Dirac pulses. For this purpose, FIG. 3 shows the resulting curve of shaft speed n and shaft position P as a function of time t. With the selected pulse sequence, the double integral of the torque disappears, as required, and with it the change in the position of the rotor.

[0045] According to the illustration in FIG. 2, a 4-pulse sub-sequence can be described for measuring the dependence of the current i.sub.k on the flux Ψ.sub.k:

SEQ4(Ψ.sub.k):={Ψ.sub.k, Ψ*.sub.k, −Ψ*.sub.k, −Ψ.sub.k}.

[0046] The measurement point result from the four current values results in:

i.sub.k=½(i.sub.k,1−i.sub.k,4).

[0047] Under actual conditions, the ideal mechanical equation of motion is “disturbed” by the friction torque which occurs. This influences the zero balance of the change in position. The two pulses applied in the middle of the sequence also bring about a different magnitude in the torque compared to the edge pulses due to the position deflection that is effective there. By extending the sub-sequence to eight pulses, this effect can be reduced:

SEQ8(Ψ.sub.k):={Ψ.sub.k, Ψ*.sub.k, −Ψ*.sub.k, −Ψ.sub.k, Ψ*.sub.k, Ψ.sub.k, −Ψ.sub.k, −Ψ*.sub.k}.

[0048] The measurement point result from the eight current measurement values:

i.sub.k=¼(i.sub.k,1−i.sub.k,4+i*.sub.k,5−i*.sub.k,8).

[0049] Ideally, the curve of torque shown in FIG. 3 and FIG. 4 results from torques M.sub.1 to M.sub.4, rotational speed n and position P.

[0050] The actually switched pulse durations are considered hereinafter. The flux pulses are impressed by means of the switched voltage space vector U of the pulse inverter. The current measurement value is recorded over the duration of a switched zero pointer interval N. In the subsequent interval, the current is reduced again via the freewheeling diodes by pulse block z. This can be seen in FIG. 5. Assuming a constantly switched voltage U, with negligible saturation, flux Ψ and current i increase in a linear manner with time t. If the saturation is actually considered, the steepness of the current i increases over the pulse duration. The torque, as a product of flux Ψ and current i, thus increases at least quadratically over time t.

[0051] Based on the nominal torque value reached at the end of the pulse, the minimum pulse duration required to impress the nominal flux value and the mechanical nominal starting time, the resulting position deflection can be estimated via the switched pulse. The following table shows the mechanical nominal running times T.sub.mech and the deflection φ for 50 Hz synchronous reluctance machines of different power P.sub.n.

TABLE-US-00001 P.sub.n/kW T.sub.mech/ms φ/° 0.55 90 0.33 3.0 82 0.37 15 102 0.29 45 284 0.1

[0052] If the final value is fixed, the position deflection increases quadratically with the quotient of the pulse duration that is actually applied to the minimum value. The minimum value can be estimated using the reciprocal value of the nominal angular frequency. For 50 Hz machines, this results in a minimum duration of 3.2 ms.

[0053] The penetration of the stator current into the stator flux decreases depending on the frequency. The cause is eddy currents in the stator and rotor iron, which are induced according to Lenz's law in such a way that they counteract a change in the flux (transformer effect). Compared to quasi-steady-state operation under rotary field conditions, the stator current with pulsed excitation therefore contains additive components that are not coupled with the flux. In order to minimize these effects, the temporal extension of the pulses must not be made arbitrarily mall. Furthermore, the factor between the two types of measurement can optionally be determined by means of a rotary field reference measurement in a controlled current frequency-impressed operation (I/F operation) under no-load conditions. For this purpose, the same flux is specified in both measurement types and the ratio of the required currents is determined. The saturation characteristic measured in pulse form can be scaled with this factor.

[0054] Due to the magnetic symmetries, the specification of the flux Ψ.sub.k (generator value of the sub-sequence) over the first quadrant is sufficient for a complete measurement. The specifications can be followed in the direction of the rotor transverse axis on tracks a, b, c and so on. This is illustrated in FIG. 6.

[0055] During the measurement sequence, the rotor position can be identified repeatedly if necessary and the comparison of the coordinate system can be updated. This is explained below with reference to the flow chart of FIG. 7. In a step S1, a counter is set to 0 and the first measurement sequence is considered. In a step S2, the rotor position is then identified, and the coordinate system is initialized. In a step S3, the counter is increased, and the subsequent measurement sequence is considered. Here, in a step S4, a check is made as to whether a new adjustment is necessary. If this is the case, the method is continued with step S2; otherwise the measurement is ended, and the result is output in step S5. This result can be output in the form of a table in which the measured values of the flux are entered via the current.

[0056] The specification of the measuring point distribution can be equidistant For this purpose, FIG. 8 shows by way of example an equidistant specification of the components of the flux Ψ.sub.k with respect to the axes d and q. FIG. 9 shows the resulting image of the measured current values.

[0057] Alternatively, the measurement point distribution can be specified based on a priori knowledge of the saturation curves which are typical for synchronous reluctance machines. For this purpose, with a small amount of flux, the distance between the specifications can initially be selected to be larger and then compressed as the amount of flux increases. For this purpose, FIG. 10 shows the specification of the flux pulses based on a priori knowledge of typical d-flux saturation curves. FIG. 11 shows the resulting mapping of the current readings. The a priori specification results in a more homogeneous distribution of the measured current values.

[0058] FIG. 12 shows a diagrammatic flow chart for the determination of the saturation characteristic of the synchronous reluctance machine. The result of the measurement sequence is provided in a step S6. The result of the measurement sequence is the allocation of the independently specified flux to the dependently measured current in tabular form for all points approached. The table can be read or evaluated in both directions. For example, the rows and columns of the table can be reversed in a step S7.

[0059] By means of subsequent interpolation of non-equidistantly measured data at any specified support points (according to standard numerical methods), the dependency relationship of the measurement can be reversed, and the respective dependent stator fluxes can be interpolated for the currents (support points) that can now be specified independently (step S8). In a step S9, the tables in which the flux for the support points of the stream is entered can then be output. When using a closed-form flux model, in another step a parameter estimation method can be carried out in a step S10. The interpolation and/or the parameter estimation method can be implemented inside the drive or outside the drive. The algorithms (standard numerical method) can be purchased from third party suppliers or implemented independently.