JPS60256074A - Method and device for determining vector of magnetic flux ofrevolving field machine - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

Detailed Description of the Invention [Industrial Field of Application] The present invention is a method for determining the magnetic flux vector of a rotating magnetic field machine from a stator current and a stator voltage, comprising: a) rotating magnetic flux vector from the current and voltage; A process in which the electromotive force vector of the machine is formed; b) a process in which the electromotive force vector is modified by a feedback signal derived from the magnetic flux vector; and C) a magnetic flux vector is formed by integrating the modified electromotive force vector. and a method comprising: Furthermore, the invention relates to a device for carrying out this method and its application.

[Conventional technology]

Such a method is described in Patent Application No. 30 of the Federal Republic of Germany.
In the device according to No. 26202, it is used for operating a rotating magnetic field machine powered by a frequency conversion device by means of magnetic field orientation. In the magnetic field orientation, the state of the magnetic flux vector is detected, and the converter that supplies power to the rotating machine, in relation to the state of the magnetic flux vector, divides the stator current component parallel to the magnetic flux and the stator current component perpendicular to the magnetic flux. and are controlled so that they can be influenced independently. One given value for the absolute value of the magnetic flux is set through the control of the stator current component parallel to the magnetic flux (magnetizing current), and the other perpendicular to the magnetic flux (active current)
is proportional to the torque and can be used directly for decoupling control of speed or torque.

However, for this magnetic field orientation, it is necessary to know the magnetic flux state. In this case, it is advantageous not to directly measure the magnetic flux using a Hall element, but to calculate it from electrical quantities using a calculation model circuit.

The simplest way to do this is to use a so-called "voltage model" in which the induced emf is determined by subtracting the resistive voltage drop and inductive leakage voltage in the stator from the input voltage of the motor using a single emf detector. It is. In this case, the magnetic flux is obtained by integrating the electromotive force.

To describe the rotating machine current, rotating machine voltage, emf, and magnetic flux, each has two determinants, e.g., a stationary (i.e., stator-referenced or "space-fixed") coordinate system or a rotor-referenced coordinate system that rotates with the rotor axis. ”) coordinate system or a coordinate system rotating with the magnetic field axis (“field reference”), a plane vector given by a rectangular coordinate component or a polar coordinate component in a coordinate system rotating with the magnetic field axis (“field reference”) may be used. It is easiest to consider using a standard orthogonal coordinate system.

This is because, for example, in a three-phase rotating machine, the
° From three out-of-phase voltages and currents to one “
The corresponding stator current vector ±
, and the space-fixed orthogonal component of the stator voltage vector and S (
This is because such "stator-based" vector components need only form the subscripts S and s2). At that time, the electromotive force vector 1s takes into account the stator resistance r5 and leakage inductance 11! , -and, -rS -Sat, -1' ·dt.

/dt by component-wise addition. The orthogonal stator-based components of the magnetic flux vector S are obtained by integrating the corresponding components of the electromotive force vector. In a coordinate system that has a coordinate axis φ1 parallel to the magnetic field and a coordinate axis φ2 perpendicular to the magnetic field and rotates together with the magnetic flux vector, the electromotive force vector is the component e of the "magnetic field reference"? + and e 92, and from the physical vector relation ヱ=S−・dt,
According to Pn= Therefore, the voltage model always operates with respect to the stator.

The open integrator required for integrating the electromotive force is prone to drift and must be stabilized, for example via a zero adjuster located in the return conductor of the integrator.

However, the zero point drift of the integral also suppresses correspondingly slow changes in the flux vector at low frequencies. Furthermore, during steady-state operation, if the setpoint value for the stator current is predetermined in terms of the magnetic field, angular errors also occur which appear particularly at low frequencies and lead to an incorrect magnetic field orientation. However, despite this drawback, the voltage model has good dynamic properties.

However, one model value for the rotor flux is defined as the rotor current (i.e. the stator current vector ±S and the exciter current 1 in the case of a synchronous machine) and the measured rotor position λ
Alternatively, it is also possible to start with a rotor speed of 1, which is often advantageous in terms of measurement technology. This "current model" electronically simulates the processes occurring in the rotating machine, insofar as they lead to the formation of magnetic flux. For this current model it is advantageous to use a magnetic field-based coordinate system, in which the rotor time constant is taken into account as the time constant of one smoothing element, and the current model is A model flux frequency is formed from which a flux angle can be formed by integration.

The transformation from one coordinate system to another coordinate system rotated by a given angle involves inputting the corresponding components of the vector to be transformed and one corresponding angle signal of the rotation angle at the angle input, such as the sine and cosine. This is done by applying one so-called "one-vector rotator".

In the current model, model parameters for rotating machine parameters must be set as accurately as possible.

For example, changes in rotor resistance due to temperature lead to fluctuations in the model magnetic flux in both static and dynamic processes. For high operating frequencies, therefore, the voltage model is advantageous, whereas at low frequencies the current model, despite its static inaccuracies, leads to better model values for the magnetic flux.

Therefore, the said Federal Republic of Germany Patent Application No. 30262
The 02 specification describes a method of using both models in combination. Correspondingly to the voltage model, two components of a model electromotive force vector +Its(u) are formed from the rotating machine current and the rotating machine voltage, which are assigned to the voltage model, and from these components the components are then assigned to this voltage model. The corresponding component of the assigned magnetic flux vector is formed as a command vector for the zero adjuster. In this case, the circuit operates with reference to the stator and includes one integrator for each quadrature emf component for the formation of the magnetic flux. For the stabilization of this integral, each component of this flux vector is applied in one return conductor to one regulator, the output signal of which is used to correct the corresponding component of the model emf vector. It is given to the integrator input terminal as a correction amount. The corresponding component of the model flux vector, which is formed from the stator current and the rotor position λ in the current model, is then supplied as command variable 0 to the target value input of the regulator.

Thus the regulator has at its input the difference vector 'I'
The difference vector is adjusted to zero on average by receiving the space-fixed orthogonal component of s (u) 'P''s and providing the space-fixed orthogonal component of one correction vector and feeding it to the voltage model. Thereby the voltage model is made to follow the current model, at least with respect to its steady-state behavior,
It is thus possible to maintain good dynamic properties of the voltage model, while at low frequencies a better steady-state flux determination of the current model is utilized.

The outputs of the two known correction adjusters are essentially orthogonal stator reference components of one correction vector rotating at the frequency of the vector ヾ. The regulator thus has to process alternating current at all times, which can not only be disadvantageous at high operating frequencies, but also creates difficulties, especially if the method is to be implemented by a microprocessor.

[Problem that the invention seeks to solve]

SUMMARY OF THE INVENTION Accordingly, it is an object of the present invention to provide an improved method and apparatus for determining the magnetic flux vector of a rotating field machine.

[Means for solving problems]

According to the invention, this objective is achieved according to claim 1:
This is achieved by the method described in item 11 or the apparatus described in item 21. Preferred embodiments of the invention are set out in claims 2 to 10, 12 to 20, and 22 to 31.

Similarly, in the present invention, first, an electromotive force vector of a rotating magnetic field machine is formed from voltage and current by one electromotive force detector,
The starting point is that this electromotive force vector is then modified by a feedback signal derived from the requested magnetic flux vector, and the flux vector is formed by the integration of this modified electromotive force vector.

According to two basic embodiments, to modify the electromotive force vector one starts from a rotating Cartesian coordinate system, in which the electromotive force vector is subsequently processed. One of the coordinate axes then ultimately indicates the direction of the magnetic flux vector, as will be explained later, so that the transformed emf vector has an orthogonal component e? l and e with 92. A magnetic flux vector is now formed from this modified emf by a special integration method.

In order to explain this method, FIG.
Two regulating loops are shown, the output quantity 2 of which is passed through one negative feedback loop with an amplification factor y to the regulator input quantity X
has been added to.

Since the amplification factor of the open integrating amplifier is also represented by , the mathematical structure of this circuit is as shown in the center of FIG. 1, and the following equation holds true.

(x-y-z) ・V=z or z=x/ (y+1/V) In that case, if the amplification factor is large enough, 1im z=x/y Yyu^, and finally Fig. 1 1 as shown on the right side of
is equivalent to one division.

In FIG. 2, the magnetic flux components A95, A! in the reference frame rotating at an (first arbitrary) frequency -3 are shown. A circuit is shown for forming a spatially fixed coordinate system (axis 1, s
2) and the rotating coordinate system (11, jL2) is generated by an integrator 2 from a corresponding frequency signal -s. With this angle signal φS, one vector rotator 3 forms a component e〒e,2 converted from the electromotive force component e 1+, e 52 based on the stator, and at this time, one angle is changed from the angle φ. The function generator 4 generates angle functions cosφ3, sinφ
3, i.e. the stator-based orthogonal component of one corresponding unit vector is formed and fed to the vector rotator 3 as a corresponding rotation angle signal pair, so that the vector 3 only needs to perform a small number of simple algebraic operations. It is particularly advantageous to do so.

In order to calculate the magnetic flux from the electromotive force, the converted component e〒1
．． It is not enough to integrate e〒2 (integrators 5.6), but also to integrate the rotor components eも! in these integrators corresponding to the vector relationships mentioned above. 1 = Nasal S - Arch S2 and eRvr =
'14 s.A, 1 must be supplied. Second
A corresponding feedback loop with multipliers 7 and 8 is shown in the figure.

The rotating coordinate system is now magnetically oriented when the coordinate axis L1 actually points in the direction of the magnetic flux vector. After that, e v+ and ey2 indicate the electromotive force component based on the magnetic field, and! 〒1=11de1,! ! ? −〇 holds true.

This means that one zero point regulator 9, which has to operate very quickly, is component A, ? This is achieved by pre-giving the frequency -3 by adjusting to zero. Finally, FIG. 2 shows how the electromotive force vector is to be integrated in a coordinate system based on the magnetic field, taking into account the rotor component.

However, the quick regulator 9 always makes sure that P,=
Since Q is forced, multiplier 7 in FIG. 2 can be omitted. This leads to the circuit shown in FIG. Third
In the figure, each vector provided by orthogonal components is indicated by a double arrow, corresponding to the signal line for the orthogonal component.

Components that are the same as those in FIG. 2 are given the same reference numerals. As you can see from this figure, the integrator 6
and a configuration group 10 consisting of a very fast regulator 9
If the amplification factor of is sufficiently high, the structure already described in FIG. 1 results, and elements 6.8 and 9 in FIG. Yes (Figure 4).

In this way, a signal (φS) is formed that defines the rotating orthogonal coordinate system (φ1, φ2), and thereby one electromotive force detection ``a←'' causes the electromotive force vector 1. is transformed into a rotating coordinate system (vector rotation 53), thus the magnetic field-based component e of the electromotive force vector
□ and a n represent the modified electromotive force vector.

Now the absolute value of the magnetic flux vector 1, I is formed by integrating the first modified electromotive force component e91 (integrator 5
). The frequency of the magnetic flux vector -8 is formed from the ratio 8yt/'F (divider 11). Integration of this frequency (integrator 2) produces a feedback signal φS that simultaneously determines the rotation angle of the rotating coordinate system and the direction of the magnetic flux vector. FIG. 5, in which the stator-based polar coordinate component of the magnetic flux vector has already been determined, shows how this method can be used to determine the magnetic field orientation of a rotating field machine 21, here an asynchronous machine, which is powered by a frequency converter. It is shown how it can be incorporated into the control section by encroachment. A 372 coordinate transformer converts the measured values for current and voltage into the corresponding stator reference vector i3 and.

is formed, and from these, the electromotive force detector 24 detects the stator-based electromotive force vector! , form. This is illustrated in FIG. 5, where the multipliers 25, 26 and one differentiator 27 form the resistive voltage drop vector d, and the leakage voltage drop vector l-diS/dt, forming the summing point 28.
and 29 by subtraction from the voltage vector.

The vector rotator 3, the circuit group 30 consisting of elements 2.5 and 11 and the angle function generator 4 then generate the magnetic flux absolute value 1ヱ, 1 and the stator-referenced quadrature component of the unit vector US pointing in the direction of the magnetic flux vector. form. 1ヱ in the magnetic field reference coordinate system! 1 = instep 7. Katsu? ? = 0, and in the stator-based coordinate system, the polar coordinate magnetic flux component lヱ, 1 =! The multiplication of φS or the unit vector component by the magnetic flux absolute type produces the stator-based orthogonal component of the vector 〈S=A・is, ``sl=A・cosφS and A, 2=A・sinφ$.

Thereby, information regarding the magnetic flux of the rotating magnetic field machine required for operation by magnetic field orientation of the rotating magnetic field machine can be obtained. In operation with such a magnetic field orientation system, for example, a rotation speed adjustment (rotation speed adjustment 531 ) may be performed. The output signal of the rotational speed regulator 31 supplies a setpoint value for the torque that the rotating machine must produce in order to maintain the rotational speed, or a setpoint effective current 17 proportional thereto. here"
By "active current" is meant the component which is perpendicular to the magnetic flux and which produces a torque. The command quantity for this active current can of course also be predetermined by a torque control or adjustment or in other ways.

Furthermore, FIG. 5 shows that the magnetic flux of the rotating magnetic field machine can be adjusted to a predetermined target value 1A and 1'' by means of a magnetic flux angle adjuster 33.This magnetic flux adjuster The output signal of 33 provides a command quantity i ?+ for the "magnetizing current", i.e. the component of the stator current parallel to the magnetic flux. However, often the flux adjustment is omitted and the magnetizing current remains at a constant value during rated operation. The magnetic flux and the reduced magnetic flux in the field weakening range are predetermined accordingly, so that the command quantities for the active current and the magnetizing current are represented by a vector representing the orthogonal component of the target stator current in the coordinate system of the magnetic field reference.
Z) Also, from this magnetic field-based command amount, information about the magnetic flux angle is used to form an appropriate stator-based command amount for the stator current given to the rotating machine via the frequency converter 20 and its control device 34. All you need to do is

In the embodiment according to FIG. 5, the control device 34 has two separate inputs for the absolute value and direction (phase) of the stator current vector. The manipulated variable for the absolute current value is calculated from the target vector ±〒 based on the magnetic field using one vector analyzer 35.
can be formed by This vector analyzer has a target absolute value 11 on its absolute value output terminal, and an angular function pair for the target angle α''! between the stator current vector and the axis 11=15 (magnetic flux axis) on its angle signal output terminal. cosα4〒
, Sinα0! will occur.

A setpoint current vector predetermined in this ``field reference'' coordinate system can be used as a setpoint value for a current regulator, which is provided with a corresponding magnetic field reference component of the actual current vector. For this purpose, a vector rotator 36 transforms the actual current vector into a coordinate system referenced to the magnetic field by means of the flux angle φ, and a subsequent vector analyzer 37 converts the actual current vector by means of the magnetic flux angle φ, while a subsequent vector analyzer 37 converts the actual current vector with the actual absolute value lit and the current angle α , (angle between magnetic flux and current).The adjustment of the absolute value of the current determined in this way (absolute value regulator 38) forms an absolute value manipulated variable for the control device 34, while correspondingly An angle adjuster 39 provides the frequency control variable in advance in such a way that the current angle α with reference to the magnetic field is equal to the corresponding target value α.For this angle adjustment with reference to the magnetic field, the angle α9, the tangent of this angle Or other functions of this angle can be used.In FIG. 5, the orthogonal component sin α! of the magnetic field reference appears at the vector outputs of vector analyzers 35 and 37 as actual and setpoint values.
is used.

FIG. 5 shows an example of a magnetic field reference adjustment unit, in which the adjustment of the current to a predetermined target value with the magnetic field reference takes place within the coordinate system of the magnetic field reference. That is, the actual current value is transformed by the vector rotator 36 into a magnetic field-based coordinate system so that the regulators 38 and 39 can be preloaded with a DC flow. At this time, by giving the magnetic field frequency to the summing point 40 connected later in the sense of preliminary control, the burden on the adjuster IB unit 39 can be significantly reduced.

The absolute value of the magnetic flux is finally determined by adjusting the rotating machine's magnetic field using orienting silane! The values determined according to the invention of the magnetic flux angle φ, and the magnetic flux angle φ determine the control of the rotating machine. However, at low frequencies, the level of voltage measurements required for flux determination is low enough to cause inaccuracies in flux determination. These inaccuracies are particularly noticeable during steady-state operation with low frequencies, while dynamic behavior can still be detected relatively accurately by the device. Thus, in FIG. 6, the circuit is commanded by the absolute value of the magnetic flux and the angle so that the steady-state error is adjusted to approximately zero, but the detection of dynamic behavior is maintained.

That is, at low frequencies, the magnetic flux angle is a unit vector J of one command quantity, in this case one magnetic flux command vector s.
), it is sufficient to follow the angle φS or its unit vector 1s accordingly with one tracking adjuster 40. In FIG. 6, in order to calculate the command amount φ, a current model 41 given the actual current value and the rotor position angle is used, and −φ−8 is set to the coordinates of the magnetic field reference by one vector rotator 42. It is converted into a system. This transformation corresponds to one angular difference formation. Instead of using a current model, the command variable φ″s can also be determined in other ways, for example from a setpoint value that is formed in a magnetically oriented control.

The angle tracking adjuster 40 thus supplies a frequency such that the direction of the detected magnetic field vector on average corresponds to the direction pregiven by φ'. The magnetic field frequency can thus be extracted from the output of regulator 40. On the other hand, the magnetic field direction (i.e. one coordinate axis of the magnetic field reference coordinate system) is determined by the integral (integrator 43) and the subsequent angle function cosφ
5SsinφS (function generator 4
4). FIG. 6 shows that if the directional information obtained in this way is required for controlling the rotating machine by magnetic field orientation (vector rotator 36), for example the stator voltage vector or other variables can be fixed. It has been shown that it can be used to transform from a child reference coordinate system to a magnetic field reference coordinate system.

Since the angle tracking adjuster 40 already provides the field reference direction of the magnetic flux vector, the elements 2 and 11 necessary for determining this direction in FIG. 5 can be omitted in the circuit of FIG. 6. However, this leads to the fact that the feedback signal derived from the output of the integrator 5 no longer has any effect and is therefore also no longer needed. However, the integrator 5 now supplies only an electromotive force component parallel to the magnetic flux, which component is not mathematically equal to the absolute value of the magnetic flux. However, this error is constant, and the integrator output is the absolute value command amount! ' can be adjusted to zero by being fed to a subtraction point 45 which supplies the control deviation for an absolute value follow-up regulator 46 which is given as setpoint value. The output signal of the absolute value follow-up adjuster 46 is applied to the input end of the integrator 5 together with the electromotive force component e v+ based on the magnetic field.

In this way, according to FIG. 6, the stator-based electromotive force vector 1. However, due to the signal φ, which defines the rotating orthogonal coordinate system, the electromotive force is Hector! , is transformed into a rotating coordinate system (vector rotator 3), and the feedback signal Δ″
F' (the output signal of the absolute value tracking adjuster 46) is corrected by being added to the component e1 of the converted electromotive force vector parallel to the magnetic field. The absolute value of the magnetic flux vector is then formed by the integration of this modified electromotive force component, and the absolute value feedback signal itself is determined from the control deviation between the absolute value and the absolute value command quantity 'P''. The frequency φS of the magnetic flux vector is determined from the control deviation between the signal that determines the rotation angle and the command signal φS for the direction of the magnetic flux vector.
is formed, and the integration of this frequency forms a signal is which simultaneously determines the angle of rotation and the direction of the magnetic flux vector.

The formation of the vector I −d±,/dt in an electromotive force detector can lead to difficulties. This is because mathematically exact differentiation of rapidly changing quantities is not feasible. However, in FIGS. 12 and 13 of German Patent Application No. 3,034,275, the integral of one quantity a,
The subsequent subtraction of one quantity with integral constant t results in c
= (a+db/dt)/(1+st) where,
A circuit is described for calculating the quantity τ given by 1/(1+st), which indicates the temporal behavior of a situational element with a time constant t.

By applying this circuit to the vector and ±S components, as shown in FIG. 53 and 54, a vector es= (u3 r ・13 1 ・dls/d t−
) / (1+s t) is obtained.

This smoothed electromotive force vector I] formed by the constituent group 55 now has a temporal behavior that would mislead the formation of the flux vector as an integral of the electromotive force vector. However, this integration is performed by an integrator 56 for each component, and the output is obtained at a summing point 57 as a smoothed electromotive force vector multiplied by a smoothing time constant t (or time constant ratio t /T). can be added, one vector (1/Tles・d
t='Ps/T, that is, the magnetic flux vector of the rotating magnetic field machine is obtained. The integration time constant T, which normalizes during all integrations, is simply a proportionality factor in all circuits under consideration and does not need to be explained in detail.

Integrations within the configuration group 58 have already been performed in Figures 1 to 4.
It can also be formed by converting the smoothed emf vector into a rotating reference frame (vector rotator 3) and subsequent integration within this rotating reference frame in the manner described in connection with the figure. This is shown in FIG.

In this case, the rotation angle required for the conversion at the angle signal input terminal of the vector rotator 3 is the second component of the corrected electromotive force vector taken out from the terminal 60 and the second component integrated by the integral a5. divider 1 as the ratio of the component of 1
1 is formed from the integral of the feedback signal formed by . Since elements 2 to 5 and 11 form the integral of the vector , the output signal of the integrator 5 is the absolute value of the smoothed magnetic flux vector s, which belongs to the smoothed electromotive force. Similarly, the integrator 2 or the function generator 4 detects the direction of the smoothed flux vector, which in the dynamic process deviates from the direction of the actual (unsmoothed) flux vector as a result of the smoothing action of the component group 55. Supply the angle φS or the attached unit hector j. Therefore, the first rotating electromotive force component e v+ occurring at the terminal 60 corresponds to the component of the smoothed electromotive force perpendicular to the smoothed magnetic flux, while the second rotating electromotive force component e v+ It corresponds to the component perpendicular to the smoothed magnetic flux vector ヱ of the electromotive force.

The smoothing action is now carried out by the multiplier 61 and the summing element 62 to calculate the value supplied from the integrator 5 which is equal to the component of this vector parallel to the smoothed flux vector as the absolute value of the smoothed flux vector. is the amount (t/T)
can be compensated for by adding il, a signal proportional to the smoothing time constant t and the first component (7I) of the modified emf vector. Thereby, the first component 'P? of the magnetic flux vector in the coordinate system jgo, φ2? + is formed. Since the second component of the smoothed flux vector in this coordinate system is zero, the second component of the unsmoothed flux vector is directly at the output of the multiplier 63,
It can be taken out as one signal proportional to the smoothing time constant and the second component 7M of the modified electromotive force vector.

As a result of the directional difference between the smoothed flux vector (rotation angle Ss) and the actual flux angle (φS), at the output of the configuration group 58 there is a coordinate system referenced to the direction of the smoothed flux vector. The orthogonal components of the actual flux vector within are obtained. One vector analyzer 64 then calculates the actual magnetic flux absolute value and the orthogonal component CO8φ with respect to the angular difference or angle φ between the actual magnetic flux vector and the smoothed magnetic flux vector or the smoothed magnetic flux angle.
T and sin$〒 can be obtained. The actual magnetic flux direction φS with reference to the stator then occurs as the angular sum φv+17 (vector rotator 65).

If the actual frequency is then required, it must be determined from the frequency position of the smoothed flux vector and the derivative of the angle φT. However, the frequency corresponding to this derivative rarely exceeds 0.1% of the smoothed frequency. If this frequency is required only, for example, for the precontrol of the angle adjustment according to FIG. As the frequency, the frequency of the smoothed magnetic flux vector can be used.

Integrators are generally prone to zero point drift and other integration errors that can be detrimental, especially at low frequencies (steady or quasi-steady conditions for rotating field machines). Fig. 9 shows the vector of the magnetic flux vector s in the stator-based coordinate system S1, S2, in the case where the rotating magnetic field machine is operating steadily, but the integrator used has a zero point deviation. The vector trajectory of the claimed magnetic flux vector whose trajectory is shown is eccentric. That is, the center point O of the vector locus
is shifted by one vector A, the "DC component" of the magnetic flux vector, with respect to the coordinate origin Qo. In this case, the magnetic flux vector components calculated by the circuit are mixture quantities in which the sine-like motion of the stator-based orthogonal coordinates of the actual rotating machine magnetic flux is superimposed on each orthogonal component of the DC component vector A. . In order to suppress this DC component in the stationary case, it is advantageous that a correction hector is added to the electromotive force vector. This correction vector changes from the magnetic flux vector to zero when the magnetic flux vector is rotated uniformly, and during steady operation of the rotating magnetic field machine, it suppresses the DC component in the space-fixed vector locus of the magnetic flux vector. It is being led out. That is, the absolute value of the correction vector must be proportional to one "volatile quantity" of the magnetic flux vector, where the volatile quantity is defined as one in which the vector locus of the magnetic flux vector is not eccentric (uniform rotation). It means the amount that disappears when

Such a volatile quantity is, for example, a magnetic flux that is positive during positive rotation above the straight line 00-0 drawn with a broken line in FIG. 9, and negative below the straight line 0o-0. The angular acceleration of the vector is φS.

Another preferred volatile quantity is the time derivative φ of the magnetic flux absolute value, the sign of which is inverted.

If we give the correction vector in advance a direction perpendicular to the direction of the magnetic flux itself, then, for example, in the coordinate system based on the magnetic field, δ square 9. =0, δK, 2=-U holds true. In this case, the first
As shown in FIG. 0, the correction vector is formed by rotating the magnetic flux vector 〉 by εma=+π/2, corresponding to the negative sign of φ, that is, the positive sign of δvn. When passing through the straight line 00-0, dK/dt=Q holds true, so the correction vector W becomes zero when passing through the straight line 00-0, and after that, it corresponds to the sign reversal of φ within the half-year plane at the bottom right. is rotated by angles -ε, -π/2 with respect to the vector s.

As shown in Fig. 10, the DC component of the correction vector perpendicular to the vector A is 1 along the vector locus.
The components δ≈Δ that are averaged to zero during rotation, while the components parallel to the DC component always have the same direction, opposite to the DC component vector.

By adding this correction vector to the transformed electromotive force vector at the input of the circuit provided for integration in the coordinate system of the magnetic field reference, the vector locus of the magnetic flux vector obtained by integration is The larger the DC component vector is, the more strongly it is shifted in the opposite direction to the DC component vector. However, in a steady state, the correction vector completely disappears, so the magnetic flux is not constantly erroneously distorted.

Such a circuit is shown in FIG.

As new elements, one absolute value follow-up regulator 70 and one angle follow-up regulator 71 and one switch 72 are introduced in this circuit. Both articulators 70, 71 are inactivated by closing their switches, and the switching device 7
2 is placed in the position shown, the configuration shown in FIG. 8 is obtained. Before the electromotive force detector 55 which forms an electromotive force vector or a smoothed vector i from measured values of current and voltage, the DC component in the components of the voltage vector is collected using a relatively low amplification factor and the voltage is calculated for each component. A DC component adjustment unit that is subtracted from the vector is connected. This DC component adjustment section 73 is
In order to avoid actually causing phase distortion of the voltage vector,
It is designed to have a weak effect.

A calculation circuit 74 is connected after the electromotive force detector, which supplies the modified electromotive force vector. This calculation circuit is
It includes a vector rotator 3 that forms an orthogonal electromotive force component in a rotating coordinate system that is rotated by a rotation angle with respect to a stator reference coordinate system. The calculating circuit 74 is followed by an integrating circuit 58, whose integrator 5 calculates the absolute value of the magnetic flux vector (
In this case, first the absolute value of the smoothed magnetic flux vector)
supply. The rotation angle i is supplied by an angle signal former which includes a second integrator 2 and, if angle function pairs are always used as the angle signal, a function former 4. The input signal of this integrator is the signal φS fed back from the integrating circuit 58 as a signal indicating the frequency of the smoothed magnetic flux vector.

Said correction vector d'P is provided by one correction vector former. μ is orthogonal to the flux vector or the smoothed flux vector, and therefore there is only one component δ that is also perpendicular to the flux! , t or δ''?t, the correction vector former simply has one differentiator 76' and a vector analyzer 64.
It is only necessary to include one signal conductor branching at the absolute value output terminal of . Moreover, this differentiation is not necessary in the first place. This is because the derivative of the absolute value of the magnetic flux essentially coincides with the input signal of the integrator 5, and therefore also the absolute value of the correction vector can be calculated with sufficient accuracy by the converted emf vector or the corrected emf This is because it can be extracted from the corresponding components of the vector.

The correction vector ultimately achieves not only a direct current adjustment but also a damping of the entire flux determining device. The dynamic properties of the magnetic flux determination are thereby adversely affected.

However, this can be avoided by determining the absolute value of the correction vector by the difference φ-ψ゛ (here φ' is the command amount for the change in magnetic flux) rather than by the value alone. In particular, φ' can be taken from the current model described at the top or from the target value of the rotating field machine control.

Furthermore, it has proven advantageous in some cases that the correction vector is not exactly perpendicular to the magnetic flux vector or the smoothed magnetic flux vector. Particularly at low frequencies, it is advantageous for the correction vector to also have a component parallel to the magnetic flux vector. 1st
21! I is a predetermined advantageous control vector for various values of the magnetic flux frequency as or 1τ! The vector locus of the magnetic field reference is shown.

As you can see from this figure, 0. At frequencies above 1, the component perpendicular to the magnetic field significantly exceeds the component parallel to the magnetic field. Although at low frequencies the angle ε ma does deviate from about 90° to about 180°, values belonging to the frequency spread, at which the component perpendicular to the magnetic field would completely disappear, are not reached. This is because there is a singular state in which the magnetic flux vector no longer rotates but remains stationary. If the device is also to determine the magnetic flux at rest, the superposition of correction vectors is disabled.

Providing in advance the angle ε between the corrected vector trajectory and the magnetic flux vector by the vector trajectory programmed according to FIG. This means that it is the basis of the correction vector within the system. Thus, the correction vector thereby formed not only has an angle related to ↓ or j, but also has its absolute value equal to −
is proportional to the volatile quantity via a proportionality coefficient (i.e. the absolute value of a pre-given control vector in relation to the function).

If the device according to FIG. 11 is used to act on the control of a rotating field machine with a claimed magnetic flux vector, it may be advantageous to modify the control vector 6 in relation to this vector trajectory operation. could be. In particular, it may be advantageous to vary the control vector Δ as a function of the load state, for example the angle between current and voltage of the rotating field machine or other state variables W of the rotating machine.

The action via the state quantity W is such that in some cases the correction vector is no longer antiparallel to the DC component vector after averaging over one revolution of the magnetic flux vector on the vector locus, so that this DC component vector also It may not be properly zeroed, but the interaction with the rotary machine and its control may nevertheless act in such a way that a stable steady-state rotary machine operation is achieved.

Negative branch (εr〈0) in the vector trajectory according to Fig. 12
is calculated when the magnetic flux rotates with a negative sign mathematically. In this case, in the eccentric vector locus according to the first theta diagram, the value is positive in the upper left half-plane and negative in the other half-plane, so that in this case also the direction of the control vector is in the first-half plane. are the same.

Control hector soil corresponding to Figure 12! Alternatively, in the device according to FIG. 11, the presetting of the earth mass can be accessed by the input signal formula of the second integrator 2 and possibly also by the load angle of the rotating field machine or other operating variables W of the rotating field machine. This can be done by one function memory 75 (FROM).

This control vector can be multiplied component by component by -φ or e〒 or by the difference formed by the command variable φ′ and then fed to a summing point 77 in calculation circuit 74 . In this case, depending on the motor or generator operation, angles <90° are also possible.

The output of the integrating circuit 58 is, as described above, the orthogonal component of the unsmoothed magnetic flux vector with respect to the direction of the smoothed magnetic flux vector. There is no difficulty in converting the orthogonal components to polar coordinates or to other coordinate systems. For example, the vector analyzer 64 calculates absolute value coordinates! and the angular coordinate φT (treated as a unit vector). The vector rotator 65 then sets the stator-based angular coordinate φS=
Tτ+φT is formed, and a multiplier 80 supplies the stator-based orthogonal coordinates of the vector 〈$ from the angular coordinates and absolute value commands expressed as unit vectors. moreover,
FIG. 11 shows that one actual vector or one setpoint vector (for example, the stator current vector ±S in the rotating field machine control according to FIG. 6) is determined for further processing in the rotating field machine control. It is shown that the vector rotator 81 provided with seven beads can transform the φ coordinate system, and the vector rotator 82 can transform the coordinate system into a stator-based coordinate system.

If the switch 72 is now switched from the position shown to another position, the divider 11 will no longer be activated.
The configuration shown in the figure is obtained. Although the integrator 2 always acts as an angle generator for forming the rotation angle φS, the angle following joint 71 is currently connected in front of this integrator. This angle tracking regulator, when activated by its switch, adjusts the rotation angle φS by shifting the frequency 17 relative to the rotation angle 7 formed by the integrator 2 (pregiven, for example, from the current model). Command angle φ4. Adjust to In this arrangement the angles are used as angle signal pairs, so the angular difference is determined by one vector rotator 8
Formed by 3. However, for the control of the angle tracking regulator 71 one component of the angular difference, e.g.
Only φ′, −[) is required.

As mentioned above, in this case the absolute value tracking regulator 70 is configured to be activated by opening its shorting switch. This regulator has a difference between A2 and A of one absolute value command amount and the absolute value of the magnetic flux taken out at the output end of the integrator 5! The component of the modified electromotive force vector e
? is adjusted to zero by superimposing the regulator output signal on +.

In this switch position, the smoothing of the electromotive force carried out according to FIG. 11 and the subsequent correction in the integrating circuit 58 also take effect. Similarly, the attenuation performed by the correction vector can be continued if desired. The switch is placed in this position, especially when the frequency of the rotating field machine is low. In this case as well, the command quantity essentially determines the magnetic flux calculation. The errors due to low levels of voltage measurements become significant in the stationary case, while on the other hand the good dynamic properties of the voltage model are practically preserved.

The device according to the invention can also be used for monitoring the magnetic flux of a rotating field machine for adjustment and testing purposes independently of the control of the rotating field machine. The device according to the invention also acts on the rotary machine control in the upper frequency range and simultaneously runs in a preparatory run in the lower speed range in which the rotary machine is controlled in another way (e.g. by a current model). It can be applied to By appropriately pre-giving the command quantities φ□, normal°, and φ′, the voltage model can be stably maintained within the lower rotational speed range, where it is difficult to use the voltage model by itself due to the low level of the voltage measurement value. It can be determined by a command variable (for example a current model), with the voltage model dynamically correctly detecting dynamic deviations from the steady state. In this case, the transition from the state commanded by the command amount to the state not commanded can be effected discontinuously, simply by switching the switch 72 to the state shown in FIG. However, continuous transitions are also possible by switching the switches alternately with frequency-related duty ratios.

[Brief explanation of drawings]

FIG. 1 is a circuit diagram and its equivalent circuit diagram for explaining the method according to the present invention, FIGS. 2 to 8 are connection diagrams of embodiments of each part of the apparatus according to the present invention, and FIGS. 9 and 10 are FIG. 11 is a vector ilLm diagram for explaining the present invention, FIG. 11 is a connection diagram of another embodiment of the apparatus according to the present invention, and FIG. 12 is a vector locus diagram for explaining the present invention. DESCRIPTION OF SYMBOLS 1... Operational amplifier, 2... Integrator, 3... Vector rotator, 4... Angle function generator, 5.6... Integrator, 7, 8... Multiplier, 9...Adjuster, 11...
- Divider, 20... Frequency converter, 21... Rotating magnetic field machine, 24... Electromotive force detector, 25.26... Multiplier, 27... Differentiator, 28.29...・Additional points, 3
1... Rotation speed regulator, 33... Magnetic flux regulator, 34.
...Control device for frequency conversion device, 35... Vector analyzer, 36... Vector rotator, 37... Vector analyzer, 38... Absolute value adjuster, 39...
Angle adjuster, 40... Angle follow-up adjuster, 43... Integrator, 46... Absolute value follow-up adjuster, 50... Integrator, 51.52... Multiplier, 56... Integrator, 58・
...Integrator circuit, 64...Vector analyzer, 65.
... Vector rotator, 71 ... Angle following joint device, 72
. . . Switching device, 74 . . . Calculation circuit, 80 . . . Multiplier, 81 . . . Vector rotator.

Claims (1)

[Claims] 1) From the stator current (±S) and stator voltage (and), the magnetic flux vector of the rotating magnetic field machine (21) ('Ps-A・Sat,
), the method includes: a) a process in which an electromotive force vector (±S) of a rotating machine is formed from measured values of current (±S) and voltage (and S); and b) an electromotive force vector. C) A process in which the magnetic flux vector (↓S) is modified by the feedback signal (↓S) derived from the requested magnetic flux vector; d) a modified emf vector is defined by orthogonal emf components in a rotating coordinate system; and e) the absolute value ('V) of the magnetic flux vector is determined by a first rotation. r) The feedback signal is formed by the integration of the electromotive force component (eve), and the magnetic flux vector (e , ) as the frequency (small $), and the rotation angle (φS) of the coordinate system rotated by the integral of the frequency (nose s)
) and an angular signal that simultaneously determines the direction of the magnetic flux vector (
-ψ-S) is formed. 2) Transformation of the vector formed from the current and voltage measurements and corresponding to one smoothed emf to form orthogonal emf components in the rotating coordinate system and the integration of the modified emf vector. The first component (8i), the circumstantial time constant (t) and the first component of the modified electromotive force vector
addition with one signal (multiplier 61) proportional to the component of (
The addition element 62) forms the first orthogonal component (A) of the magnetic flux vector (layer 2) in the rotating coordinate system, and also the smoothing time constant (1) and the second orthogonal component of the modified electromotive force vector. 2. Method according to claim 1, characterized in that the second orthogonal component (Ai) of the magnetic flux vector is formed from a signal (multiplier 63) proportional to the component (i). 3) A correction vector (aV) is added to the converted electromotive force vector, which is derived from the magnetic flux vector and suppresses the DC component in the space-fixed vector locus of the magnetic flux vector during steady operation of the rotating magnetic field machine. A method according to claim 1 or 2, characterized in that: 4) In the correction vector (aV), 1 for the magnetic flux vector
It is characterized in that two predetermined directional differences and one absolute value that is proportional to the amount (-φ or -8??) that disappears when the vector locus of the magnetic flux vector is not eccentric are given in advance. The method according to claim 3. 5) The correction vector is given in advance so as to be averaged over one rotation on the vector locus and to be in the opposite direction to the direct current component in the spatially fixed vector locus. Method. 6) the absolute value of the correction vector with a given proportionality coefficient, in particular a proportionality coefficient determined as a function of the frequency of the magnetic flux vector (φS) and/or of the state quantity (W) of the rotating field machine; proportional to the derivative of the magnetic flux absolute value (φ) or proportional to the first orthogonal emf component (T1) of the transformed emf vector; 1 component and an unsteady command amount (φ“
5. The method according to claim 4, characterized in that the method is proportional to the difference between (+-φ1 or v〒-φ'). 7) A component perpendicular to the magnetic flux vector in the direction of the correction vector, which is related to the frequency of the magnetic flux vector, in particular to the frequency of the magnetic flux vector and to the state quantity of the rotating magnetic field machine, and which is not equal to zero. 5. The method according to claim 4, wherein a component that does not disappear is provided in advance. 8) The component perpendicular to the magnetic flux is positive or negative in relation to the sign of the product of the magnetic flux frequency and the absolute value of the correction vector. Method. 9) The command quantity is formed from a setpoint value of the rotating machine or from the actual values of the current of the rotating machine and the rotor position by simulation of the magnetic flux.
The method described in section. 10) Applied to the operation by magnetic field orientation of a rotating magnetic field machine powered by a frequency converter, the position of the magnetic flux is determined, and the frequency converter is applied to the magnetic flux of the stator current in relation to the determined magnetic flux position. 2. A method as claimed in claim 1, characterized in that the components parallel to and perpendicular thereto are controlled in such a way that they can be influenced independently of each other. 11) A method for determining the magnetic flux vector ('Ps) of a rotating magnetic field machine from the stator current (±S) and the stator voltage (base), comprising: a) the current (±S) and the voltage (and S); b) The process in which the electromotive force vector (mut, ) of the rotating machine is formed from the measured values of C) a magnetic flux vector ('Ps) is formed by integration of the modified electromotive force vector (to)), d) the modified electromotive force vector is in a rotating coordinate system. The first orthogonal electromotive force component (ev+) and the first feedback signal (Δ'
P'') and the second in the rotating coordinate system (e!, +ΔA'')
e) the absolute value of the magnetic flux vector ('f) is formed by the integration of the first modified electromotive force component (+3y++ΔA'') and the first feedback A process in which the signal is formed from the absolute value control deviation (A"-v) from the absolute value command quantity ('P");
f) The second feedback signal (φ$) determines the magnetic flux from the control deviation of the signal (f s ) defining the rotation angle (φS) of the rotating coordinate system from the angle command signal (7,) with respect to the direction of the magnetic flux vector. One angular signal (j-s
) is formed. 12) Transformation of the vector formed from current and voltage measurements and corresponding to one smoothed emf to form orthogonal emf components in the rotating coordinate system, and integration of the modified emf vector. The rotating coordinate is determined by addition (addition element 62) of the first component (eve), the circumstantial time constant (t), and one signal (multiplier 61) proportional to the first component of the modified electromotive force vector. A first orthogonal component (Ai) of the magnetic flux vector (E7) in the system is formed and is also proportional to the situational time constant (1) and the second orthogonal component (11) of the modified electromotive force vector 1 signal (multiplier 6
12. A method according to claim 11, characterized in that the second orthogonal component (Ai) of the magnetic flux vector is formed from 3). 13) A correction vector (earthwork) is added to the converted electromotive force vector, which is derived from the magnetic flux vector and suppresses the DC component in the space-fixed vector locus of the magnetic flux vector during steady operation of the rotating magnetic field machine. 13. The method according to claim 11 or 12, characterized in that: 14) Add to the correction vector (ay) one predetermined directional difference to the flux vector and 1 proportional to the amount (-! or -1) that disappears when the vector trajectory of the flux vector is not eccentric. 14. The method according to claim 13, wherein the two absolute values are given in advance. 15) Claim 14, characterized in that the correction vector is averaged over one rotation on the vector trajectory and is given in advance so as to be in the opposite direction to the direct current component in the spatially fixed vector trajectory. the method of. 16) Given the proportionality coefficient, especially the frequency of the magnetic flux vector ($
``s) and (or) a proportionality coefficient determined as a function of the state quantity (W) of the rotating magnetic field machine, the absolute value of the correction vector is proportional to or transformed to the derivative of the i flux absolute value (φ). The first orthogonal electromotive force component (
“”? +) or put an unsteady command with the derivative or the first component of the transformed emf vector
15. The method according to claim 14, characterized in that the method is proportional to the difference (φ - φ' or eR - ψ゛). 17) A component perpendicular to the magnetic flux vector in the direction of the correction vector, which is related to the frequency of the magnetic flux vector, in particular to the frequency of the magnetic flux vector and to the state quantity of the rotating magnetic field machine, and which is not equal to zero. Claim 1 characterized in that a component that does not disappear is provided in advance for the
The method described in Section 4. 18) Claim 15 or 16, characterized in that the component perpendicular to the magnetic flux is positive or negative in relation to the sign of the product of the magnetic flux frequency and the absolute value of the correction vector.
The method described in section. 19) The command according to claim 11 or 12, characterized in that the command quantity is formed from the setpoint value of the rotating machine or from the actual values of the current and rotor position of the rotating machine by magnetic flux simulation. Method. 20) Applied to the operation by magnetic field orientation of a rotating magnetic field machine powered by a frequency converter, the position of the magnetic flux is determined, and the frequency converter is applied to the magnetic flux of the stator current in relation to the determined magnetic flux position. 12. A method as claimed in claim 11, characterized in that the components parallel to and perpendicular thereto are controlled in such a way that they can be influenced independently of each other. 21) A device for determining the magnetic flux vector of a rotating magnetic field machine from voltage and current, comprising: a) one electromotive force vector (1s) forming an electromotive force vector (1s) from the measured values of current (±S) and voltage (us); Power detector (5
5) and b) one feedback signal derived from the magnetic flux vector (m
l's) by one modified emf vector (±7
) and C) one integrating circuit (58) forming a magnetic flux vector (!7) as an integral of the modified electromotive force vector, d) the calculation circuit comprises means (3, 79) for forming orthogonal components of the electromotive force vector in a rotating coordinate system; e) the integrating circuit forms a first integrator signal ('P j+ ); In order to 1 component or, if the coordinate axes are parallel to the magnetic flux, the absolute value coordinates of the vector; The second integrator (2) is provided with two integrators (2).
) is derived from one quantity (as or , ) of the claimed magnetic flux vector and is given a feedback signal (T〒) which is associated with the frequency of the magnetic flux vector,
A magnetic flux vector determining device for a rotating magnetic field machine, characterized in that the second integrator signal (is) is supplied to a calculation circuit (74) as a rotation angle of a rotating coordinate system. 22) a divider (11) for the integrator circuit (58) to form a ratio between the second orthogonal electromotive force component and the first integral signal;
, and the output signal (depression) of this divider is the second
and a second integrator signal is supplied to an output device (65, 85) for the angular or orthogonal component of the magnetic flux vector. Apparatus according to clause 21. 23) One correction vector former (75, 76) for pre-giving one correction vector (ay) that rotates with respect to the magnetic flux vector, is not particularly parallel to the magnetic flux vector, and disappears in a steady state. ) and means (77) for adding the electromotive force vector and the correction vector. 24) means for forming a quantity approximately proportional to the derivative (A) of the magnetic flux from the first component of the corrected electromotive force vector or the absolute value of the magnetic flux vector; 24. Device according to claim 23, characterized in that it comprises absolute value forming means (76) for determining the absolute value. 25) A control vector former (75) is provided which is fed with a quantity (W) indicative of the frequency of the magnetic flux vector or rotational reference system and preferably the load condition of the rotating machine; characterized in that a multiplication circuit (76) is provided which determines the components of the control vector in the rotating reference frame in a functional relationship and which supplies a correction vector by multiplying the control vector by a quantity determining the absolute value. An apparatus according to claim 23 or 24. 26) The angle signal former (2, 4) is connected to the second integrator (2
) and is provided with the difference between the rotation angle i]) and one commanded rotation angle (φ%S) for the formation of the feedback signal (φS). includes a first integrator signal and one command absolute value (V
′) forms another feedback signal (Δ%F) and a calculation circuit (74)
22. The apparatus of claim 21, wherein includes one summing element for the feedback signal and the first orthogonal emf component. 27) A patent claim characterized in that a switching means (72) is provided which can switch the input of the second integrator between the output of the divider (11) and the output of the angle tracking adjuster (71). The apparatus according to any one of the ranges 26 and 22 to 25. 28) the electromotive force detector (55) comprises means for forming a smoothed electromotive force vector (jT) which is fed to a calculation circuit (74) for transformation into a rotating reference frame; , from the first integrator signal ('Pi) and the signal (multiplier 61) taken out at the calculation circuit output for the first orthogonal electromotive force component and proportional to the smoothing time constant (1). The first orthogonal component of the magnetic flux vector in the reference frame (
'Pv+) and a signal (multiplier 63) taken out at the calculation circuit output end for the second orthogonal electromotive force component and proportional to the smoothing time constant 28. Device according to any of claims 21 to 27, characterized in that it is formed as component (Ai). 29) Device according to claim 27, characterized in that means (65, 80) are provided for converting the magnetic flux vector into a stator-based coordinate system. 30) Any one of claims 2 to 28, characterized in that the device (23) for suppressing the DC voltage is connected in front of the voltage input terminal of the electromotive force detector. The device described. 31) The rotation angle i of the rotating coordinate system is the means (8) for coordinate transformation of the actual vector and/or the target vector.
1,82) and the transformed actual vector and/or target vector are supplied to a frequency conversion device feeding the stator current of the rotating field machine. Apparatus according to any of the ranges 21 to 29.