Method and Apparatus for Direct Regulation of Output Currents of an Inverter Feeding a Predetermined Number of Induction Machines

The invention relates generally to a method and a apparatus for the direct regulation of output currents of an inverter, and more particularly to such an inverter feeding one or more induction machines without rotational speed sensors. The inverter uses a secondary current control to regulate the output currents. The secondary current control receives a current command variable system.

"Drehzahlveranderbare Antriebe in der Praxis" pages 34 to 38, Siemens, Bereich Energie- und Automatisierungstechnik, Order No. A 19100-E319-A365, discloses a pulse frequency converter which outputs a three-phase current system of variable voltage and frequency for a predetermined number of induction machines. The pulse frequency converter comprises an uncontrolled power rectifier, a dc voltage intermediate circuit and a machine-side inverter. The power-side current converter generates from the supply voltage a constant dc voltage which is smoothed by capacitors in the intermediate circuit. The machine-side current converter forms from the constant dc voltage, a three-phase voltage system with variable frequency and voltage. Transistors or gate turn-off thyristors (GTO thyristors) are used as current switches for the inverter. The pulse patterns for driving the inverter are generated in the microprocessor drive unit. For this purpose pulse patterns are selected which ensure optimal operation with nearly sinusoidal motor currents and minimum total losses, over the entire setting range of the frequency converter. The drive and regulation of the pulse frequency converter is fully digital. All functions including drive unit, ease of operation, and extensive protective functions are processed via a 16 bit microprocessor. A vector regulator is provided as the regulator, which achieves high regulating dynamics. The magnetization current and the active current forming the torque are determined from the available actual motor current and voltage. These current components of the current vector are compared with the setpoints. The difference is eliminated by regulation. In this way, given torque setpoints can be adhered to exactly. Rotational speed actual value sensors are not required up to a rotational speed setting range of 1:10.

Due to unavoidable switching and delay times of the switches of the inverter and the drive circuit whose durations are variable and not exactly known in the individual case (depending upon for example the parameter spread of the electronic components instantaneous current, and temperature) the frequency converter output voltage differs from the settings of the voltage drive unit. These erroneous voltage/time areas cause lower order harmonics and dc components in the output voltage, particularly with low frequency converter output frequencies. These harmonics lead to a non-sinusoidal curve of the output currents, to additional losses in the motor, and to concentricity properties which are not too good at low frequency converter output frequencies.

Marked improvement in the behavior of a drive is possible if the frequency converter outputs a regulated three-phase current system of variable frequency and amplitude, instead of a controlled voltage system. The non-ideal switching response of the power semiconductors as well as delay times occurring in the signal flow are detected by the regulator and eliminated by regulation.

DE 30 46 392 C2 discloses a method for controlling three-phase pulse inverters in which a three-phase sinusoidal current command variable system is given. The phase currents of an induction machine with a virtual neutral point following the pulse inverter are measured and compared to the command variable input. As a function of the deviation, the phases of the inverter are switched by means of hysteretic flip-flop controllers.

In groups drives wherein several motors, whose number can also be varied during operation, are connected in parallel, a field-oriented control cannot be employed. This results from the following facts:

(1) the different motors of the group drive can have different loads placed on the motors which cause the voltage drops across the winding resistances and across the leakage inductances to be different in each motor, which in each instance have a different phase position of the flux relative to the common inverter output voltage. If these motors are not equipped with a position detector the position of the flux cannot be determined.

(2) the number of motors of the group drive is not known since the number changes during operation, also the nominal value for the flux-forming current components depending on the number or the type of the motors present is also hot known.

"Drehzahlveranderbare Antriebe in der Praxis" discloses that in group drives, frequency control is provided as the variable regulation, with the voltage given as a function of frequency and with current limiters. With single motor drives, a vector regulator as a rotational speed regulator without a rotational speed sensor but with current and moment limiters is provided as the variable regulator.

The present invention is directed towards the problem of developing a method and an apparatus for the direct regulations of invertor output currents by means of a secondary current control wherein this invertor feeds one or more induction machines without rotational speed sensors.

The present invention solves this problem by calculating the actual values of an active current (i.sub.Wact) and a reactive current (i.sub.Bact) from the stator currents (i.sub.Rx, i.sub.Sx, i.sub.Tx) and the state signals (S.sub.R, S.sub.S, S.sub.T) of the current switches of the invertor. Then the method converts the actual value of the active current (i.sub.Wact) <into an actual value of the torque-forming current (i.sub.qact)> and the actual value of the reactive current (i.sub.Bact) and, into an actual value of the flux-forming current (i.sub.dect). Next, the method simulates the actual value of the rotational speed (n.sub.act) by using the actual value of the torque-forming current (i.sub.qact). The method then compares a given rotational speed setpoint (n.sub.set) with the simulated actual rotational speed value (n.sub.act), and generates a torque-forming current (i.sub.qw) from this comparison. The method then compares the flux-forming setpoint of the current component (i.sub.set) given as a function of the stator frequency (f.sub.1) of the induction machine with the actual value of the flux-forming current component (i.sub.dact) and generates a flux-forming command current (i.sub.dw) from this comparison. The method transforms the flux-forming and torque-forming currents (i.sub.qw, i.sub.dw), which are field-oriented currents, into a stator-oriented current vector (i.sub.S) and converts the stator-oriented current vector (i.sub.S) into a command current variable system (i.sub.Rw, i.sub.Sw, i.sub.Tw). Finally, the method supplies the command current variable system (i.sub.Rw, i.sub.Sw, i.sub.Tw), to the secondary current control.

The basis of the regulation is supplying the actual values of the active and reactive currents to the motor which are calculated using the invertor output currents and the states of the inverter switches. The flux-forming and the torque-forming current components of the motor current are determined from these actual active and reactive currents, via an R.sub.S /X.sub.o transformation which accounts for the resistance of the winding and the leakage inductance. The nominal value for the flux-forming components is obtained as a function of the frequency converter output frequency. The setpoint for the torque-forming current component is available as the output signal of a rotational speed controller wherein the actual rotational speed is simulated. The amplitude and the angle of the command current vector are determined from the calculated command variables for the two current components by means of a coordinate transformation.

A fast secondary current control ensures that the frequency converter output currents or the phase currents of the frequency converter-fed induction machine correspond at all times to the command variables given by the superior field-oriented regulation. This applies even for very low output frequencies so that even at low rotational speeds very good concentricity behavior can be achieved.

The present invention solves the problem for an inverter feeding several induction machines without speed sensors by calculating actual values of the active power (P/U.sub.d) and reactive power (P.sub.q /U.sub.d) supplied to the induction machines from the output currents (i.sub.Rx, i.sub.Sx, i.sub.Tx) and the state signals (S.sub.R, S.sub.S, S.sub.T) of the current switches of the inverter. Then, the method calculates the actual value of a functional-r.m.s. of the output voltage (U.sub.x) the actual values of the active current (i.sub.Wact), and a reactive current (i.sub.qact), from the actual values of the active and reactive power by using an absolute value of stator-oriented current vector. Next, the method converts the actual active current (i.sub.wact) <into the actual torque-forming current (i.sub.qact)>, and the actual reactive current (i.sub.Bact) and into the actual flux-forming current (i.sub.dact). The next step simulates the actual rotational speed (n.sub.act) by using the actual torque-forming current (i.sub.qact). The next step compares the given rotational speed setpoint (n.sub.set) with the simulated actual rotational speed (n.sub.act) and generates a torque-forming current (i.sub.qw) from this comparison. The next step compares the output voltage setpoint (U.sub.w) of the inverter given as a function of the stator frequency (f.sub.1) of the induction machines, with the actual output voltage of the inverter at a calculated fundamental frequency and generates the flux-forming current (i.sub.dw) from this comparison. Next, the method transforms the flux-forming and torque-forming current components (i.sub.qw, i.sub.dw), which are field-oriented currents, into a stator-oriented current vector (i.sub.s), and converts the stator-oriented current vector (i.sub.S) into a current command variable system (i.sub.Rw, i.sub.Sw i.sub.Tw), and supplies the current command variable system (i.sub.Rw, i.sub.SW, I.sub.Tm) to the secondary current control.

This method calculates the actual value of a fundamental-r.m.s. of the inverter output voltage from the actual values of the active and reactive powers supplied to the induction machines. These powers are calculated from the inverter output currents and the states of the current switches of the inverter by using the absolute value of a stator-oriented stator current. The actual value of a fundamental-r.m.s. of the inverter output voltage is compared with a setpoint of the inverter output voltage and forms a command variable of the flux-forming current component as a function of the resulting difference. The command variable of the torque-forming current component is generated in the same way as in the regulation of individual drives. In contrast to the regulation of individual drives, the regulation of group drives for the generation of the command variable of the field-oriented current components a setpoint and actual value of the flux-oriented current component are not compared with each other. Rather, a setpoint of the inverter output voltage, formed as a function of the stator frequency, is compared to the actual value of a fundamental-r.m.s. of the inverter output voltage formed from the active and reactive powers supplied to the induction machines. This method orients the flux of the individual motors of the group drive. The quality of this orientation is high, especially if the loads of the individual motors are approximately equal. In addition, through the calculation of the actual value of a fundamental-r.m.s. of the inverter output voltage it is possible to recognize where or not individual motors, during operation, are switched to or away from the group drives, and the command variable of the flux-oriented current component will be changed accordingly.

In an advantageous method for the determining the command variable of the flux-forming current component, the nominal value of the flux-forming current component is set equal to the command variable of the torque-forming current component. In application of this method, care should be taken that a minimum magnetization or a maximum magnetization (for example nominal magnetization) are never exceeded. For particularly small or particularly large loads corrections via a characteristic function are possible. In this way, the motor is magnetized according to a load whereby the active and reactive current supplied to the motor are of same magnitude. This means that the motor is operated at partial load with a phase angle of approximately 45.degree. whereby the partial load efficiency of inverter and motor is raised to its maximum. In this way, the inverter and the motor are operated with the lowest possible losses. This load-adaptive magnetization offers the advantage that the possibility of operating with improved efficiency under partial load is not tied to a particular rotational speed-torque characteristic of the operating machine, but rather this is possible in any given drive.

The present invention also solves the problem of developing an apparatus for the direct regulation of the output currents of an inverter feeding an induction machine without a rotational speed actual value sensor. This apparatus supplies the current command variable system to a secondary current control. A command variable former calculates the current command variable system (i.sub.Rw, i.sub.Sw, i.sub.Tw) that is manipulated by the inverter, and supplying the current command variable system to the secondary current control. The command variable former has an actual value computer, which receives the output currents (i.sub.Rx, i.sub.Sx i.sub.Tx) and the state signals (S.sub.R, S.sub.S, S.sub.T) of the current switches of the inverter as inputs. The command variable former has a comparator with a given rotational speed setpoint (n.sub.set) as a positive input, and a simulated speed (n.sub.act) as a negative input. The command variable former includes a rotational speed simulator which simulates the rotational speed (n.sub.act). The rotational speed simulator has an input coupled with the output of the actual value computer, and outputs the simulated rotational speed (n.sub.act ). The command variable former has a transformation device with three inputs, and a rotational speed controller which follows the comparator. The rotational speed controller has an output coupled with the rotational speed simulator, and with one of the inputs of the transformation device. The command variable former includes another comparator and a flux-forming current component controller, which follows this comparator. The flux-forming current component controller is coupled with one of the inputs of the transformation device. Another transformation device with an output coupled to the negative input of the second comparator follows the actual value computer. A flux-forming current component setpoint former has an output coupled with the positive input of the second comparator. Also included is a slip compensator with an input coupled with the output of the rotational speed controller, and another input coupled with the output of the flux-forming current component setpoint former. An adder has an input coupled with the output of the slip compensator, another input coupled with the output of the rotational speed simulator, and an output coupled with one of the inputs of the transformation device.

According to the invention, the current command variable former generates a three-phase sinusoidal current command variable system from the actual values of the inverter output currents, the states of the current switches of the inverter, and a given rotational speed setpoint. The three-phase sinusoidal current command variable system is supplied to a secondary current control. The actual value computer calculates the actual values of the active and reactive currents supplied to the induction machine, initially from the inverter output currents (phase currents of the induction machine), the states of the current switches of the inverter, and the calculated amplitude of the current command variable. The active and reactive currents are subsequently transformed into a torque-forming and a flux-forming current component, which are field-oriented values. The torque-forming and the flux-forming current command variables are generated separately from each other by these field-oriented values. Moreover, the actual rotational speed is determined by a formed acceleration signal. The actual rotational speed is required for determining the command variable of the torque-forming current component. Consequently, this current command variable former generates a current command variable system from the field-oriented regulation using only the actual values of the inverter output currents and the states of the current switches of this inverter. The current command variable system is supplied to a secondary current control that easily controls an induction machine without a rotational speed sensor, achieving very good concentricity even with low rotational speeds.

An apparatus according to the invention for the direct regulation of the output currents of an inverter feeding several induction machines each without a rotational speed sensor comprises a secondary current control and a command variable former. The command variable former calculates a current command variable system (i.sub.Rw, i.sub.Sw, i.sub.Tx) that is manipulated by the inverter, and supplies this current command variable system to the secondary current control. The command variable former comprises an actual value computer with the output currents (i.sub.Rx, i.sub.Sx, i.sub.Tx) and the state signals (S.sub.R, S.sub.S, S.sub.T) of the current switches of the inverter as inputs. A comparator has a given rotational speed setpoint (n.sub.nom) as the positive input, and a simulated speed (n.sub.act) as the negative input. A rotational speed simulator simulating a rotational speed (n.sub.act) has an input coupled with the output of the actual value computer, and outputs the simulated rotational speed (n.sub.act). The command variable former includes a transformation device with three inputs, and a rotational speed controller following the comparator. The rotational speed controller has an output coupled with the rotational speed simulator and coupled with an input of the transformation device. The command variable former has a second and third comparator, and a voltage controller following the third comparator. The voltage controller has an output coupled to an input of the transformation device. Another transformation device follows the actual value computer, and has an output coupled with the negative input of the third comparator. A setpoint former calculates an output voltage (U.sub.W) of the inverter, and has an output coupled with the positive input of the third comparator. A slip compensator has an input coupled with the output of the rotational speed controller, and has another input coupled with the output of the voltage controller. An adder has an input coupled with the output of the slip compensator, has a second input coupled with the output of the rotational speed simulator, and has an output coupled with an input of the transformation device, and coupled with the input of the setpoint former.

According to the present invention, the current command variable former generates a three-phase sinusoidal current command variable system from the actual values of the inverter output currents, the states of the current switches of the inverter, and the rotational speed setpoint. This current command variable system is supplied to a secondary current control. The actual value computer of the current command variable former calculates the actual values of the active and reactive powers supplied to the induction machines, initially from the inverter output currents and the states of the current switches of the inverter. The value of a fundamental-r.m.s. of the inverter output voltage is subsequently calculated from the active and reactive powers by means of a calculated current amplitude of the current command variable. This value of the inverter output voltage at the fundamental frequency is compared with a setpoint of the inverter output voltage, and the command variable of the flux-forming current component is formed as a function of the resulting difference. The torque-forming current component is generated as in the case of regulation of an individual drive. Consequently a group drive that does not use any rotational speed sensors can be approximately regulated with field-oriented variables.

In order to improve the dynamic behavior of the drive in the simulating the actual rotational speed by means of a calculated acceleration signal, this acceleration signal is switched onto the simulated rotational speed. Moreover for determining the phase of the current command variable vector, the load angle is differentiated and this angle change added to the stator frequency and subsequently integrated. This changes the phase of the current command variable corresponding to a load change of the drive.

In order to achieve especially low loss operation of the motor and the inverter in the partial load range, the setpoint of the flux-forming current component is set equal to the command variable of the torque-forming current component. This ensures that the motor under a partial load is operated with a load angle of 45.degree.. In this way, the partial load efficiency of the inverter and the motor is raised to its maximum. This load-adaptive magnetization offers the advantage of the possibility of operating with improved partial load efficiency that is not dependent upon the particular rotational speed/torque relationship of the operating machine but rather can be used in any drive.

A particularly advantageous embodiment of the current command variable former occurs when this current command variable former is a microprocessor. Relative to the analog technique, the microprocessor technique offers the advantage of increased precision, precise selectability, and reproducibility of setpoints and regulating parameters.

In the method and apparatus of the present invention for generating the current command variable system for a secondary current control, and inverter output currents can be regulated directly without needing to detect the rotational speed actual value of the drive (both for individual and group drives). Moreover a voltage intermediate circuit or a current intermediate circuit inverter can be used as the inverter.

FIG. 1 illustrates an embodiment of the apparatus of the present invention for single motor drive.

FIG. 2 illustrates an embodiment of the apparatus of the present invention for a group drive.

FIG. 3 illustrates an advantageous embodiment of the apparatus of the present invention of FIG. 1.

FIG. 4 illustrates another advantageous embodiment of the apparatus of the present invention of FIG. 1.

FIG. 5 illustrates a segment of an advantageous embodiment of the apparatus of the present invention of FIG. 2.

FIG. 6 illustrates a segment of an especially advantageous embodiment of the apparatus of the present invention of FIG. 2.

FIG. 7 illustrates a block circuit diagram of one of the transformation devices of the circuit configuration according to FIG. 2.

The figures use the same reference symbols for identical objects.

FIG. 1 shows an apparatus for directly regulating the inverter output currents i.sub.Rx, I.sub.Sx, and i.sub.Tx of a frequency converter that feeds an induction machine 2. For the sake of clarity only the load-side inverter 4 of the frequency converter is shown. A voltage intermediate circuit or a current intermediate circuit frequency converter can be used as the frequency converter. The load-side inverter 4 is built with current switches. Transistors or gate turn-off thyristors (GTO thyristors) can be used as the current switches. The states of these current switches are indicated by the state signals S.sub.R, S.sub.S, and S.sub.T.

The regulation comprises a current command variable former 6 and a secondary current control 8. As an example, the current command variable former 6 generates three-phases of a sinusoidal current variable system i.sub.Rw, i.sub.Sw, and i.sub.Tw for the secondary current control 8 from the inverter output currents i.sub.Rx, i.sub.Sx, and i.sub.Tx, a selectable rotational speed setpoint, n.sub.set and the state signals S.sub.R, S.sub.S, and S.sub.T of the current switches in the inverter. In addition, the actual values of the inverter output currents i.sub.Rx, i.sub.Sx, and i.sub.Tx are also supplied to the secondary current control which controls the inverter switches by comparing the command currents i.sub.Rw, i.sub.Sw, and i.sub.Tw and the inverter output currents i.sub.Rx, i.sub.Sx, and i.sub.Tx.

The current command variable former 6 comprises an actual value computer 10 and a first comparator 12 on the input side, and a transformation device 14 on the output side. In addition, the command variable former 6 comprises a rotational-speed simulator 16, a setpoint generator for the flux forming current 18, a slip compensator 20, a rotational speed controller 22, and a flux generating current controller 24. The setpoint generator 18 forms the flux forming current setpoint from the stator frequency f.sub.1.

A given rotational speed setpoint n.sub.set is connected to the positive input of the first comparator 12 and a simulated rotational speed n.sub.act is connected to the negative input of the first comparator 12. The simulated rotational speed n.sub.act is formed by the rotational speed simulator 16. The rotational speed controller 22 generates a torque forming current i.sub.qw from the rotational speed difference at the output of the comparator 12. This torque forming current i.sub.qw is supplied to a positive input of the rotational speed simulator 16, the slip compensator 20, and to the first input of transformation device 14. An actual value of the torque forming current i.sub.qact is supplied to the negative input of the rotational speed simulator 16.

The rotational speed simulator 16 has a comparator 26 with positive and negative inputs. The comparator calculates the difference i.sub.qa of the command value i.sub.qw and of the actual value i.sub.qact of the torque forming current. These currents i.sub.qw and i.sub.qact are equivalent to the motor torque and to the load torque, respectively. Consequently, the current difference i.sub.qa is equivalent to the torque difference. This torque difference is equal to the acceleration moment. The equivalent current difference i.sub.qa is supplied to an integrator 28 and a proportional controller 30. The outputs of the integrator 28 and the proportional controller 30 are added by an adder 32 which outputs the simulated rotational speed n.sub.act. The proportional controller 30 significantly improves the dynamic behavior of the drive. This simulated rotational speed n.sub.act is also supplied to another adder 34 whose second input is the slip frequency f.sub.2. The adder 34 outputs the stator frequency f.sub.1 which is then supplied to the transformation device 14.

The actual value of the torque-forming current i.sub.qact is one of the two field-oriented variables which is generated by the input-side actual value computer 10 and a succeeding transformation device 36 from the inverter output currents i.sub.Rx, i.sub.Sx, and i.sub.Tx and the state signals S.sub.R, S.sub.S, and S.sub.T of the inverter switches.

The actual value computer 10 comprises a power detector 38 and an actual value computing device 40. The structure of the power detector 38 is described in detail in DE 35 04 623 Al. The power detector 38 outputs the active power P/U.sub.d and the reactive power P.sub.q /U.sub.d, which relate to the intermediate circuit voltage U.sub.d. From these values, the actual value computing device 40 calculates the actual values of the active current i.sub.Wact and the reactive current i.sub.Bact that are supplied to the motor and the phases. These actual values are calculated by the following equations. ##EQU1##

The actual values of the torque forming current i.sub.qact and the flux-forming current i.sub.dact components of the motor current are formed from the actual values of the active i.sub.Wact and reactive i.sub.Bact currents via a transformation that is a function of the winding resistance R.sub.S of the stator winding and the leakage inductance X.sub.6.

The actual value of the flux-forming current component i.sub.dact is supplied to the negative input of a another comparator 42 which has the setpoint of the flux-forming current component i.sub.dnom as a positive input. This setpoint i.sub.set is derived from a characteristic of the output frequency (stator frequency) f.sub.1 by a load-independent setpoint calculator 44 of the current setpoint, calculator 18 which also outputs the setpoint i.sub.set. This setpoint i.sub.set is also supplied to the slip compensator 20. The command value of the flux-forming current component i.sub.dw is generated from the difference of setpoint i.sub.set and actual value i.sub.dact by the current controller 24.

The value .sub.w and the load angle .epsilon..sub.w are formed from the field-oriented command variables i.sub.qw and i.sub.dw and the stator frequency f.sub.1 initially by the transformation device 14. The amplitude .sub.w and the load angle .epsilon..sub.w are obtained from the field-oriented current components i.sub.qw and i.sub.dw by a Cartesian to Polar (K/P) converter 46. The current phase angle .phi..sub.i of the current vector i.sub.S is obtained from this load angle .epsilon..sub.w and the integral of the stator frequency f.sub.1. A three-phase sinusoidal command variable system i.sub.Rw, i.sub.Sw, and i.sub.Tw is formed from this current phase angle .phi.hd i and the amplitude .sub.w by means of a command variable system former 48.

FIG. 2 represents a circuit for the direct regulation of inverter output currents i.sub.Rx, i.sub.Sx, i.sub.Tx of an inverter 4 feeding several induction machines 2. Compared to the circuit configuration of FIG. 1, a switch 58 follows the current controller 24. The output of the switch 58 is connected to an input of the transformation device 14 and also to an output of a voltage controller 60 of a voltage control unit 62. The adder 34 is also followed by a switch 64 whose output is connected with the input of the setpoint current calculator 18. The output of adder 34 is electrically connected with the transformation device 14 and with a setpoint former 66 for the inverter output voltage U.sub.w. The output of the voltage controller 60 is connected with an input of the slip compensator 20. The switches 58 and 64 are synchronized with each other, so that they open and close simultaneously. Electrically actuated switches, such as transistors, can be used as switches 58 and 64 which are switched off by means of a pulse generated by the actuation of a key "group control" in an operator control panel of the frequency converter. The setpoint former 66 is followed by a third comparator 68 whose output is connected to the input of the following voltage controller 60 and whose negative input is connected to the output of another transformation device 70. The block diagram of the second transformation device 70 is shown in more detail in FIG. 7. The inputs of the transformation device 70 are the calculated relative active and reactive powers P/U and P/U from the actual value computer 10, and an absolute value .sub.w of a calculated stator-oriented current vector i.sub.S from the output of the transformation device 14. In addition the value of the intermediate circuit voltage U.sub.d of the frequency converter is supplied since the actual value computer 10 calculates the active and reactive powers relative to the intermediate circuit voltage U.sub.d. The transformation device 70 outputs the value of the fundamental -r.m.s. of the inverter output voltage U.sub.x which is compared with a setpoint of the inverter output voltage U.sub.w. The setpoint U.sub.w is determined as a function of the stator frequency f.sub.1 by means of the setpoint former 66. This setpoint former has stored in it the characteristics for different drives (constant torque drive, turbo machine drive). The resulting regulating difference value U.sub.e is supplied to the voltage controller 60 which outputs a command value of the flux-forming current component i.sub.dw which is supplied to the input of the transformation device 14. This command value of the flux-forming current component i.sub.dw is changed by means of the voltage controller 60 until the value of a fundamental-r.m.s. of the inverter output voltage U.sub.x agrees with the setpoint of the inverter output voltage U.sub.w.

FIG. 3 represents an advantageous circuit for the direction regulation of inverter output currents i.sub.Rx, i.sub.Sx, and i.sub.Tx. This circuit differs from the circuit of FIG. 1 in that the current setpoint, calculator 18 has a load-dependent setpoint former 50 and a switch 52. The torque-forming current component i.sub.qw generated by the rotational-speed controller 22 is supplied to the load-dependent setpoint former 50. In the case of partial loads the setpoint of the flux-forming current component i.sub.set is set equal to the command value of the torque-forming current component i.sub.qw. This ensures that the motor is magnetized according to load. Frequency converter and motor are then operated with the lowest possible loss, i.e. the active and the reactive current are then of equal magnitude. Since the magnetization never should exceed a minimum or maximum with particularly low or particularly large loads of the drive, the magnetization is corrected via a characteristic. This means under partial loads, the load-dependent setpoint former 50 is connected by the switch 52 with the comparator 42, whereas outside of this partial load range the load-independent nominal value former 44 is connected with the comparator 42 by the switch 52. Consequently, not only drives with known and fixed rotational speed/torque characteristics, but also any given drive can be operated with decreased power loss.

FIG. 4 represents another advantageous circuit for the direct regulation of the inverter output currents i.sub.Rx, i.sub.Sx, and i.sub.Tx. This circuit differs from the circuit configuration according to FIG. 3 in that a pull-out protection device 54 is present. This commutation failure protection device 54 is connected with the actual value computer 10 and with the slip compensator 20 and outputs to a setpoint setting element 56. The pull-out protection device 54 compares the slip frequency f.sub.2 with a predetermined stored pull-out slip frequency f.sub.sk or a frequency f.sub..phi. corresponding to the phase angle .phi. of the predetermined pull-out slip frequency f.sub.sk. As soon as the slip frequency f.sub.2 or the frequency f.sub.100 reaches a predetermined value below the pull-out frequency f.sub.sk, the setpoint setting element 56 decreases the rotational speed setpoint n.sub.set with motor operation. This means in order to prevent the pull-out of a drive under load torque (the value of this torque can correspond for example to 90% of the value of the pull-out torque it is necessary to decrease the rotational speed setpoint n.sub.set as quickly as possible, perhaps suddenly, thus decreasing the magnitude of the slip. The pull-out protection device 54 and the setpoint setting element 56 achieve sudden reduction capability. With high dynamic demands, the evaluation of the phase angle .phi. must be given priority in order to recognize the danger of pull-out of a drive under load.

FIG. 5 shows an advantageous implementation of the voltage control unit 62 of the circuit according to FIG. 2. In this circuit, the setpoint former 66 is followed by an adder 72 whose output is connected to the positive input of the third comparator 68. The second input of the adder 72 is connected to the output of a constant element 74. The command value of the flux-forming current component i.sub.dw is present at the input of this constant element 74. The constant element 74 is set to a value R corresponding to the sum value of the stator resistances of the several induction machines 2, when the number of motors 2 is predetermined. At the output of the constant element 74 a correction value U.sub.KR =i.sub.dw R is generated, which is combined with the setpoint of the inverter output voltage U.sub.w. This compensates for the voltage drops from the motor feed lines and the winding resistances of the stator windings of the induction machines 2.

In FIG. 6 a particularly advantageous implementation of the voltage control unit 62 of the circuit of FIG. 2 is shown in more detail. In this implementation the assumption is made that the compensation is a function of the inverter output current or a function of the active component of this current. If a motor 2 from the motor group is switched off, the nominal value for the inverter output voltage U.sub.w would be reduced since the correction value U.sub.KR becomes smaller. For this reason it is useful to introduce a new correction value U.sub.KR, instead of the correction value U.sub.KR. This new correction value U.sub.KR, carries out a weighting of the original compensation with the ratio of the flux-forming current component I.sub.dRef (assumed on starting initial operation of the group drive, also called the flux-forming reference current component I.sub.dRef), to the current command value of the flux-forming current component i.sub.dw given by the voltage controller 60. Instead of the flux-forming current component i.sub.dw, the corresponding reactive currents can also be used. This weighting of the original compensation occurs by following the constant element 74 with a multiplier 76. The output of this multiplier 76 is connected with the second input of the adder 72. The second input of the multiplier 76 is preceded by a divider 78 whose first input is the flux-forming reference current component I.sub.dRef and whose second input is the output of the voltage controller 60 i.sub.dw.

FIG. 7 illustrates a block diagram of the second transformation device 70. This transformation device 70 has side two squares 80 and 82 at the input side and a divider 84 at the output side. The outputs of the two squarers 80 and 82 are connected with each other by an adder 86 whose output is connected with a square root calculator 88. The square root calculator 88 is followed by a multiplier 90 whose output is connected with the first input of the divider 84. The value of the intermediate circuit voltage U.sub.d of the frequency converter is present at the second input of the multiplier 90. The second input of the divider 84 is preceded by a constant element 92 which has as an input an absolute value .sub.w of the stator-oriented current vector i.sub.S. This circuit determines the value of the inverter output voltage at the fundamental frequency from the actual values of the active and reactive powers P/U.sub.d and Pq/U.sub.d, in each instance relative to the value of the intermediate circuit voltage U.sub.d. By means of the value of a fundamental-r.m.s. of the inverter output voltage U.sub.x, the regulator recognizes whether or not the motors of the group drive are connected to or away from the group drive, or whether the operating points of the individual motors are moving away from each other. If for example motors are switched away from the group drive, then the value of a fundamental-r.m.s. of the inverter output voltage U.sub.x increases. Thereby the regulating difference U.sub.e and the command value of the flux-forming current component i.sub.dw are lowered until the inverter output voltage value of a fundamental-r.m.s. of the U.sub.x becomes equal to the nominal value U.sub.w of the inverter output voltage, which is predetermined as a function of the inverter frequency f.sub.1.