DE102018206286A1 - Method and device for current measurement of an inverter - Google Patents

Abstract

A method of measuring the current of an inverter having three half bridges, the method comprising determining a respective PWM value associated with each of the half bridges of the inverter for a first switching period, and determining whether predetermined minimum distances between the PWM values are met be, has. If the minimum distances between the PWM values are not met, the method further comprises identifying a critical PWM value from the PWM values, determining a correction value by which the critical PWM value must be corrected so that the minimum distances are maintained correcting the critical PWM value by the correction value, correcting a corresponding PWM value for a second switching period of the inverter, the corresponding PWM value being associated with the same half-bridge as the critical PWM value, the sum of all corrections being 0 ,

Description

The invention relates to a method and a device for current measurement of an inverter. The invention further relates to a computer program product and a system.

Methods and apparatus for measuring the current of an inverter are known from the prior art. However, these are complicated and expensive because at least two current sensors must be used.

It is therefore an object of the present invention to provide an improved method and apparatus for measuring the current of an inverter.

This object is achieved by a method for measuring the current of an inverter having three half-bridges, the method comprising determining a respective PWM value associated with each of the half-bridges of the inverter for a first switching period, and determining whether predetermined minimum distances between the PWM values. If the minimum distances between the PWM values are not met, the method further comprises identifying a critical PWM value from the PWM values, determining a correction value by which the critical PWM value must be corrected so that the minimum distances are maintained correcting the critical PWM value by the correction value, correcting a corresponding PWM value for a second switching period of the inverter, the corresponding PWM value being associated with the same half-bridge as the critical PWM value, the sum of all corrections being 0 , The method additionally includes detecting a first current at a first time within the first switching period in which a first half-bridge in a second switching state and a second and third half-bridges in a first switching state, and detecting a second current to a second Time within the first switching period in which the first half-bridge and the second half-bridge are in a second switching state and the third half-bridge in a first switching state, or in which the first half-bridge and the third half-bridge are in a second switching state and the second half-bridge are in a first switching state.

In this case, the inverter is a device for generating a three-phase alternating current, for example a three-phase rotary current, from a constant direct current. The DC current is provided by an accumulator in most applications, for example in an automobile.

The inverter controls the three phases by means of three half-bridges. A half-bridge consists of a terminal for the phase, which is connected via a first switch with a positive potential, and is connected by a second switch with a negative potential. The positive potential is made available for all half-bridges together at one pole of a DC voltage source, the negative potential at the other pole of the DC voltage source. The voltage between the positive potential and the negative potential is also called the intermediate circuit voltage.

The switches, also called current valves, can be implemented, for example, as MOSFET, IGBT, thyristors, or GTO. Therefore, the two switches of a half bridge must never be closed at the same time at the same time, as this would result in a short circuit of the DC voltage source, which would destroy the semiconductor components. In order to prevent a short circuit, the second switch is usually operated with a signal inverse relative to the first switch. In this case, both switches are closed in a switching operation for the duration of a blocking period, also called dead time. Thus, each half-bridge can be controlled with one signal.

In order to impose a desired voltage curve on a phase, a pulse width modulation method can be used to control the corresponding switching bridge. In this case, the switching bridge is switched in a first switching state and switched to a certain portion of a switching cycle in a second switching state. The proportion of a switching cycle after which the switch is switched is given a PWM value between 0 and 1, and determines how long the phase is connected to the positive potential and how long the phase is connected to the negative potential. This results in what potential is present for the duration of the switching cycle, or the switching period, on the corresponding phase. If the switching periods are sufficiently short, a voltage curve can thus be imposed on the phase. High-frequency components of the voltage profile which result from the alternating application of the phase with a positive potential and a negative potential, thereby approximately fall away due to the low-pass characteristic of electrical machines.

To control the frequency judge, in particular for the operation of an electrical machine, information about the controlled variables for each control period is necessary. These include the currents DC link currents, which flow over the individual phases. For this purpose, current sensors can be used over the phases. Since all half-bridges are connected together at the intermediate circuit potential, two current sensors are sufficient for determining the three phase currents. Two phase currents are measured with corresponding current sensors. The third current results with the node rule from the two measured currents, since the sum of the three currents is zero. In particular, with small and cheap engines, such as servomotors or power steering or as a servomotor, but only a current sensor can be used. In particular, the fact is exploited that with each of the poles on the half-bridges at certain times a phase, at other times but two phases are connected. Therefore, instead of making one measurement on each two current sensors, two measurements can be made on a current sensor at different times. In order that the actual phase currents from the measured currents can be reliably calculated, the measurements must take place close to one another in terms of time, namely in a common switching period.

For the sampling of the DC link current in two different times within a switching period of the inverter requires a minimum time interval between two adjacent switching edges of the inverter to perform a clean current measurement. The inverter contains stray inductors and small capacitors. These cause with each switching edge, as a result of switching a half-bridge, a vibrating behavior, which only fades after a certain time. In order not to erroneously sample the current during this transient response, a time interval between two switching operations must therefore be maintained.

The PWM values for a switching period in which the current measurement is carried out are changed so that the minimum interval between the switching edges is maintained in order to achieve a clean current measurement. In this case, preferably only minimal changes are carried out, so that a clean current measurement is barely guaranteed. In a switching period, all three switches are usually in a first switching state at the beginning. Upon expiration of a portion of the switching period indicated by a smallest of the three PWM values for that switching period, the switch associated with the smallest PWM value is toggled. The switching over of the switch belonging to the smallest PWM value must not take place before a dead time has elapsed in order to avoid a short circuit of the intermediate circuit voltage. After expiration of a portion of the switching period indicated by an average of the three PWM values for that switching period, the switch associated with the average PWM value is also switched. The switching of the switch associated with the average PWM value must occur at a minimum time interval from the switching of the switch belonging to the smallest PWM value, so that an intermediate current measurement is precise. After expiration of a portion of the switching period indicated by a largest of the three PWM values for that switching period, the switch associated with the largest PWM value is also switched. The switching of the switch associated with the largest PWM value must be done with a minimum time interval from switching the switch associated with the average PWM value so that an intermediate current measurement is accurate. Finally, the switching of the switch belonging to the largest PWM value must not be too close to the end of the switching period and thus the beginning of a subsequent switching period. Rather, at least one dead time, or blocking time, must be in between in order to avoid a short circuit of the intermediate circuit voltage.

If one of the minimum time intervals described above is not met, there is a first possibility to reduce the smaller of the affected PWM values and a second possibility to increase the larger of the affected PWM values. If one of the two possibilities results in or touches a further violation of a minimum time interval, the other option is preferably realized. The PWM value whose correction is preferred is called the critical PWM value. If neither of the two options results in or touches a further violation of a minimum time interval, it is preferable not to correct the average of the PWM values. Accordingly, the other affected PWM value is the critical PWM value. In the case when both the switching of the switch belonging to the largest PWM value falls below a minimum distance to the switching of the switch belonging to the mean PWM value, and a switching over of the switch belonging to the average PWM value, a minimum distance to the latter Switching the switch below the smallest PWM value, the smallest PWM value and the largest PWM value are corrected so that the minimum distances are just kept, while the average PWM value remains unaffected. That is, in this case, both the smallest PWM value and the largest PWM value are critical PWM values.

Correcting a PWM value may, as mentioned above, mean decreasing the PWM value or increasing the PWM value. Correcting for a positive correction value means that the PWM value is increased, while correcting for a negative correction value means that the PWM value is reduced. The correction value is determined so that the minimum time interval is just fulfilled. A minimum difference between two PWM values can be calculated from a minimum time interval between switching the switches associated with the PWM values by dividing by a switching period duration. A correction value is calculated as a minimum difference less the actual difference of the PWM values concerned.

Due to the so changed, i. corrected, PWM values, the phase voltage waveforms distortions and harmonics, which lead to acoustic problems and power loss. However, in many applications, for cost reasons, large control periods (or control cycle) are used to avoid costly and expensive fast microprocessors. Thus fall in a control period several switching periods. The distortions of the voltages due to the new PWM values can be partially compensated. The new distorted voltages of the machine can be recalculated using the corrected PWM values.

The voltage distortion is compensated by correcting a PWM value corresponding to the critical PWM value. A corresponding PWM value is a PWM value that belongs to the same switch as the critical PWM value, but for a switching period in which no current measurement is performed. With the corresponding PWM value, the correction does not serve to ensure minimum time intervals between the switching of the switches belonging to the PWM values. Rather, by correcting corresponding PWM values, the actual averaged voltage over all switching periods of a control period is again brought into agreement with that voltage value which would have resulted without the measurements of the current measurements of the critical PWM value as averaged voltage over all switching periods of a control period , Preferably, a PWM value corresponding to the critical PWM value in the switching period, which follows immediately after the switching period of the current measurements, can be corrected for this purpose. The corresponding PWM value is corrected by the same correction value as well as the critical PWM value, but in the opposite direction. Thus, if the critical PWM value has been increased by the correction value, the corresponding PWM value is reduced by the correction value. If, on the other hand, the critical PWM value has been reduced by the correction value, the corresponding PWM value is increased by the correction value. However, the corresponding PWM value does not necessarily have to be corrected by the correction value, as will be seen later. Rather, a corresponding PWM value can be corrected by an arbitrary value as long as the sum of all corrections, including the direction as a sign, yields 0. For example, when the critical PWM value is corrected by a correction value kor and the corresponding PWM value is corrected by a value -kor, the sum of the corrections is given as (kor) + (-kor) = 0.

The method according to the invention has the particular advantage that precise current measurement can be carried out simply and inexpensively, since only one shunt is required. Because the method requires only one shunt, the method is particularly suitable for realizing a robust and fail-safe control of a current control. The method has the advantage that, despite a current measurement with only one shunt distortions are avoided. This avoids acoustic problems, control problems and power losses. Next, the method has the advantage that despite a current measurement with only one shunt harmonics are avoided. As a result, an electric machine can be operated with higher efficiency, which additionally reduces the cooling effort. The method also has the advantage that the control effort is low, so no costly fast microcontroller must be used. In particular, the method can be operated with a large control period, whereby cost-effective microcontroller can be used with low processing power.

The method can be developed by following the first switching period on the second switching period. This means that the second switching period expires first and then only the first. A sequence here means in particular directly, but it can also be one or more switching periods between the second and the first switching period.

This development has the particular advantage that the position of the PWM correction and the correction compensation can be varied. This then, for example, a frequency spectrum of the corrections are optimized to minimize an impact on a machine operation.

The method may be further developed in that the method further comprises correcting a corresponding PWM value for a third switching period, wherein the corresponding PWM value is assigned to the same half-bridge as the critical PWM value.

This development has the particular advantage that the compensation of the necessary for the current measurements correction of the critical PWM value can be distributed over several switching periods and thereby mitigated. Thus, high-frequency interference can be avoided.

The method can be further developed in that the second and / or the third switching period adjoin the first switching period. This means, in particular, that the first, second and third switching periods directly adjoin one another. For example, the second switching period immediately follows the first switching period and the third switching period immediately follows the second switching period.

This refinement has the advantage that voltage values are compensated averaged over the shortest possible period of time so that the distortions in the voltage profile remain within a narrow range.

The method may be further developed by having the minimum distance between 0 and a minimum of the PWM values, and the minimum distance between a largest of the PWM values and 1, a quotient of an operating condition-locked period and a switching period duration, and wherein the minimum distance between the minimum PWM value and a next larger one of the PWM values, and the minimum distance between the largest PWM value and a next smaller one of the PWM values, a quotient of a component-related minimum time interval and the switching period duration.

This refinement has the advantage that both a design-related dead time, or blocking time, and a minimum time interval required for a current measurement are adequately taken into account in the correction of the PWM values.

In a switching period, all three switches are usually in a first switching state at the beginning. Upon expiration of a portion of the switching period indicated by a smallest of the three PWM values for that switching period, the switch associated with the smallest PWM value is toggled. The switching over of the switch belonging to the smallest PWM value must not take place before a dead time has elapsed in order to avoid a short circuit of the intermediate circuit voltage. After expiration of a portion of the switching period indicated by an average of the three PWM values for that switching period, the switch associated with the average PWM value is also switched. The switching of the switch associated with the average PWM value must occur at a minimum time interval from the switching of the switch belonging to the smallest PWM value, so that an intermediate current measurement is precise. After expiration of a portion of the switching period indicated by a largest of the three PWM values for that switching period, the switch associated with the largest PWM value is also switched. The switching of the switch associated with the largest PWM value must be done with a minimum time interval from switching the switch associated with the average PWM value so that an intermediate current measurement is accurate. Finally, the switching of the switch belonging to the largest PWM value must not be too close to the end of the switching period and thus the beginning of a subsequent switching period. Rather, at least one dead time, or blocking time, must be in between in order to avoid a short circuit of the intermediate circuit voltage.

The object mentioned in the introduction is moreover achieved by a computer program product with program code means for carrying out a method according to one of the abovementioned embodiments.

For the advantages of the computer program product reference is made to the corresponding method features.

The object mentioned at the outset is moreover achieved by a device for measuring the current of an inverter which has three half-bridges, comprising a correction device which is designed for a first switching period each having a PWM value associated with each of the half-bridges of the inverter determine and determine whether predetermined minimum clearances are maintained between the PWM values, wherein the correction device is further configured, if the minimum distances between the PWM values are not met, to identify a critical PWM value from the PWM values, a correction value to determine which of the critical PWM value needs to be corrected so that the minimum clearances are met to correct the critical PWM value by the correction value; and correct a corresponding PWM value for a second switching period of the inverter, the corresponding PWM value being associated with the same half-bridge as the critical PWM value, and the sum of all corrections being 0; and a current sensing device configured to detect a first current at a first time within the first switching period in which a first half-bridge is in a second switching state and a second and third half-bridges are in a first state; and detect a second current at a second time within the first switching period, in which the first half-bridge and the second half-bridge are in a second switching state and the third half-bridge is in a first switching state, or in which the first half-bridge and the third half-bridge in a second switching state and the second half-bridge are in a first switching state

The inventive device can be further developed in that the correction device is designed to correct a corresponding PWM value for a third switching period, the corresponding PWM value being assigned to the same half-bridge as the critical PWM value.

For the advantages of the devices, reference is made to the corresponding method features.

The object according to the invention can furthermore be achieved by a system comprising a rotating field machine, an inverter and a device for measuring current according to one of the above-mentioned exemplary embodiments.

For the advantages of the system, reference is made to the corresponding device features.

• 1: eine schematische Darstellung einer Drehfeldmaschine mit unterschiedlichen Koordinatensystemen;
• 2: ein Schaltbild einer feldorientierten Regelung gemäß dem Stand der Technik;
• 3: ein Schaltbild eines Wechselrichters mit einem Shunt zur Strommessung;
• 4: ein Zeitverlauf dreier PWM Schaltsignale;
• 5: Schaltzustände der Halbleiter des Wechselrichters gemäß 3 zum Zeitpunkt t1
• 6: Schaltzustände der Halbleiter des Wechselrichters gemäß 3 zum Zeitpunkt t2
• 7: Abtastzeiten in einem Regelzyklus
• 8: ein Zeitverlauf dreier PWM Schaltsignale mit unterschiedlichen Zeitpunkten der Abtastung
• 9: ein Schaltbild einer feldorientierten Regelung mit einer Vorrichtung gemäß einer Ausführungsform der Erfindung
• 10: ein Ablaufdiagramm einer Ausführungsform eines erfindungsgemäßen Verfahrens
• 11: ein Ablaufdiagramm einer weiteren Ausführungsform eines erfindungsgemäßen Verfahrens;
• 12: ein Ablaufdiagramm einer weiteren Ausführungsform eines erfindungsgemäßen Verfahrens;
• 13: ein Ablaufdiagramm einer weiteren Ausführungsform eines erfindungsgemäßen Verfahrens;
• 14: ein Ablaufdiagramm einer weiteren Ausführungsform eines erfindungsgemäßen Verfahrens;
• 15: Stromverlauf einer elektrischen Maschine mit idealer Strommessung;
• 16: Stromverlauf einer elektrischen Maschine, wobei der Strom mit einer Ausführungsform eines erfindungsgemäßen Verfahrens gemessen wurde;

darstellen.The invention will now be described in more detail with reference to the attached figures, in which

• 1 : a schematic representation of a rotating field machine with different coordinate systems;
• 2 a circuit diagram of a field-oriented control according to the prior art;
• 3 a circuit diagram of an inverter with a shunt for current measurement;
• 4 a time course of three PWM switching signals;
• 5 : Switching states of the semiconductors of the inverter according to 3 at time t1
• 6 : Switching states of the semiconductors of the inverter according to 3 at time t2
• 7 : Sampling times in a control cycle
• 8th : a time course of three PWM switching signals with different sampling times
• 9 FIG. 4 is a circuit diagram of a field-oriented control with a device according to an embodiment of the invention
• 10 a flowchart of an embodiment of a method according to the invention
• 11 a flowchart of a further embodiment of a method according to the invention;
• 12 a flowchart of a further embodiment of a method according to the invention;
• 13 a flowchart of a further embodiment of a method according to the invention;
• 14 a flowchart of a further embodiment of a method according to the invention;
• 15 : Current profile of an electric machine with ideal current measurement;
• 16 : Current profile of an electrical machine, wherein the current was measured with an embodiment of a method according to the invention;

represent.

In the figures, the same reference numerals designate the same or similar features.

1 shows a schematic representation of a rotating field machine 100 , in particular a permanent-magnet synchronous machine (PSM), in different coordinate systems. The induction machine 100 is preferably intended for use in a motor vehicle, for example in a hybridized or electric drive train, in a power steering or as a servomotor. The induction machine 100 includes one Stator (not shown) and a rotor 101 related to a rotation axis 102 are rotatably mounted against each other. At the stator at least three coils are uniformly offset on a circumference about the axis of rotation 102 appropriate. There are three phases U . V and W provided, with each phase U . V . W Usually connected with the same number of coils and the coils are distributed equidistantly on the circumference. The number of pole pairs Zp and the associated number of coils are selectable depending on the application. On the rotor 101 is preferably at least one permanent magnet 103 arranged, wherein a torque when driving the phases U . V . W is generated with phase-shifted AC currents, which is the rotor 101 around the axis of rotation 102 rotates with respect to the stator.

A phase-shifted control of the phases U . V . W can be displayed in different coordinate systems. In the statorfesten U . V . W Coordinate system, the coordinate axes are rotated by 120 ° from each other. Because the currents of the phases U . V . W result in zero, a current vector or current vector s can also be represented in a stator-fixed, two-dimensional α / β coordinate system. Furthermore, in 1 a rotor fixed d, q coordinate system provided, whose d-component rectified with the magnetic flux ΨPM of the permanent magnet 103 runs. A q-component is perpendicular to it. An angle between the d-axis and the α or. U-axis corresponds to an electrical rotation angle Od or gel of the rotary field machine 100 between the rotor 101 and the stator. An electrical rotation angle Od or gel corresponds to the mechanical rotation angle Omech multiplied by the number of pole pairs ZP , With the transformation of the state variables into the d . q Coordinate system simplify the differential equations of the PSM machine and the PSM machine can be controlled like a DC machine. This is called field-oriented regulation or FOR. In a field-oriented control, a total target current to flow through the induction machine, with respect to a rotor (flux) fixed d . q Coordinate system, so that some control operations are easier to perform and some calculations are simplified.

2 shows an example of a field-oriented control (FOR) for the induction machine 100 out 1 , A control component 205 generated on the basis of predefined d and q components of a total nominal current isd . Isq the induction machine 100 d and q components of a voltage usd . Usq , The d - and q components isd . Isq clamp a current vector which corresponds to the total nominal current. The in the engine 205 generated and by the d - and q components usd . Usq expressed voltage is by means of a converter 210 from d . q Coordinate system in a three-dimensional coordinate system, in particular the U . V . W Coordinate system, implemented. This results in three voltages Usu . Usv . Etc , which by means of a vector modulator 215 on the basis of a DC link voltage Udc into three corresponding pulse width modulation values PWM1 . PWM 2 , and PWM3 be implemented. The DC link voltage Udc can correspond to an on-board voltage or even a battery voltage when used in a motor vehicle. A pulse inverter 220 is set to each of the phases U . V . W alternating with a high and a low potential of the intermediate circuit voltage Udc to connect, so that is at the stage U . V . W sets a desired voltage. The applied voltages cause actual phase currents through the phases U . V and W , At least one actual phase current is determined by means of at least one sampling device 225 sampled, which also includes current sensors. Via a position sensor 230 is the electric angle gel of the PSM based on the measured mechanical rotor position Omech and the number of pole pairs ZP determined. To the mutual effect of the two streams isd and Isq can compensate for decoupling 240 be introduced.

3 shows an example and schematically an inverter, in particular designed as a pulse inverter 220 with a DC link 305 , to which a DC link capacitor 310 is connected in parallel. The inverter 220 is formed in B6 bridge circuit with three half-bridges, each of which has two current valves S1 and S2 . S3 and S4 , respectively. S5 and S6 includes. The flow valves S1 to S6 are preferably each formed in the form of a semiconductor such as a MOSFET, IGBT, thyristor or GTO. The indicated diodes are optional and may be included in each of the respective semiconductors. Three circuit breakers S1 . S3 . S5 of the inverter 100 are as high-side circuit breakers, three switches S2 . S4 . S6 arranged as a low-side circuit breaker.

The switches S1 to S6 are usually realized as semiconductors. Turning off a semiconductor is not sudden, but it requires a so-called dead time t0 until all charges in the stopband of the semiconductor have been removed and the semiconductor is completely switched off. This dead time depends on the type of semiconductor used and, for example, is less than 1 μs for a MOSFET and between 1 μs and 5 μs for the IGBT. In the inverter 220 allow the two complementary semiconductors in one phase, z. B. S1 and S2 , never be turned on at the same time to make a short circuit in the DC link 305 to avoid, leading to a destruction of the DC link capacitor 310 or the semiconductor can lead. For this reason Typically, a lock time (also called dead time) between turning off a semiconductor (until all charges have been removed) and turning on the complementary semiconductor at the same phase U . V . W inserted.

One between the DC link 305 and the induction machine 100 total flowing DC IsSum may be based on a shunt at 315 declining voltage, to be determined. The shunt preferably has a very small resistance, for example in the range of less than about 1 mΩ.

For the implementation of the FOR according to 2 are next to the rotor position and the DC link voltage Udc Information about the phase currents required to feedback the variable to be controlled (d, q currents, see 2 ) of the control loop. This means that current measuring sensors 225 needed to measure the currents. Usually, two phase currents (for example Is 1 and Is 2 ) and the third phase current is calculated by the equation Is3 = -Is1-Is2. In some applications, only one current measurement is used to determine the three phase currents of the machine. This is done using a DC IsSum based on a shunt 315 declining voltage can be determined. The shunt preferably has a very small resistance, for example in the range of less than about 1 mΩ. To obtain phase currents, the current IsSum becomes two different times t1 and t2 sampled. For both times, the phase with the smallest and the second smallest (average) PWM value is determined in each case.

4 shows an example of the switching signals in two consecutive switching periods Ts of the inverter 220 , At the time t1 are the flow valves S1 to S6 in switching states that are in 5 are shown, and at the time t2 in switching states that are in 6 are shown. At the time t1 the current is right ISSUM with the flow Is 1 in phase 1 match, as the lower flow valve S2 alone the stream ISSUM through the induction machine 100 flow. $$I_{sSum\left(t1\right)}=I_{s1\left(t1\right)}$$

At the time t2 is the current IsSum divided into phases 1 and 2, as the flow control valves S2 and S4 are closed. $$I_{sSum\left(t2\right)}=I_{s1\left(t2\right)}+I_{s2\left(t2\right)}=-I_{s3\left(t2\right)}$$

$$I_{s3}=-I_{sSum\left(t2\right)}$$

$$I_{s2}=-I_{s1}-I_{s3}$$

With the two sampled values of the DC link current ISSUM (at the times t1 and t2 ) gives the three phase currents Is 1 . Is 2 and s3 as follows: $$I_{s1}=I_{sSum\left(t1\right)}$$

$$I_{s3}=-I_{sSum\left(t2\right)}$$

$$I_{s2}=-I_{s1}-{\mathrm{<figure-callout\; id="3"\; label="I\; s"\; filenames="DE102018206286A1\_0021.png,DE102018206286A1\_0022.png"\; state="\{\{state\}\}">I\; </figure-callout>}}_s$$

In 7 an example of a quadruple PWM output in a control period is shown. In every switching period of the inverter ts a PWM value is set. The control cycle TA contains four switching periods ts of the inverter. The machine current is sampled within a switching period of the inverter 700 (in this example, the third WR Switching period), while in other switching periods within the control cycle no current measurement is performed. In the WR switching period with the current measurement, a minimum time interval between two edges is required.

Because of the minimum time interval between two edges in a switching period of the inverter, the PWM values can be changed to maintain the minimum distance. The resulting PWM change leads to a falsification of the desired machine voltage.

The original tension of the FOR usd and Usq (respectively. Usα and Usβ) are set via the calculated PWM values in the inverter. The new falsified voltages of the machine can be calculated back on the basis of the falsified PWM values. In doing so, they are talking about the PWM values the individual phases ( PWM_u . PWM_v . PWM_w ) and the DC link voltage Udc of the inverter. From those can be the two components of the voltage vector Usα and Usβ be calculated. Using a Clarke inverse transformation, the voltages of the machine in the rotor coordinate system usd and Usq be won.

The influence of the PWM values on the machine voltage can be found in Equation 6. Each PWM change is directly in the change of the two components Usα and Usβ the voltage vector of the machine (and thus also Usu . Usv . Etc . usd . Usq ) Act: $$\left(\begin{array}{c}Us\alpha \\ Us\beta \end{array}\right)=\left(\begin{array}{ccc}\frac{2}{3}& -\frac{1}{3}& -\frac{1}{3}\\ 0& \frac{1}{\sqrt{3}}& -\frac{1}{\sqrt{3}}\end{array}\right)\left(\begin{array}{c}PWM\_u\\ PWM\_v\\ PWM\_w\end{array}\right)\cdot Udc$$

Within the switching period of the inverter ( k ) (See image 7 ), at which the current idc (equivalent to ISSUM in Equation 1 and Equation 2), the PWM values may be changed, causing distortion of the voltage. The modified PWM values PWM_un . PWM_vn and PWM_wn generate a modified voltage with the two components Usαn and Usβn in the machine Equation 7: $$\left(\begin{array}{c}Us\alpha n\\ Us\beta n\end{array}\right)=\left(\begin{array}{ccc}\frac{2}{3}& -\frac{1}{3}& -\frac{1}{3}\\ 0& \frac{1}{\sqrt{3}}& -\frac{1}{\sqrt{3}}\end{array}\right)\left(\begin{array}{c}PWM\_un\\ PWM\_vn\\ PWM\_wn\end{array}\right)\cdot Udc$$

The relationship between the original machine voltage with the two components Usα and Usβ and the modified voltage with the components Usαn and Usβn can be represented by equation 8: $$\left(\begin{array}{c}Us\alpha n\left(k\right)\\ Us\beta n\left(k\right)\end{array}\right)=\left(\begin{array}{c}Us\alpha \left(k\right)\\ Us\beta \left(k\right)\end{array}\right)+\left(\begin{array}{ccc}\frac{2}{3}& -\frac{1}{3}& -\frac{1}{3}\\ 0& \frac{1}{\sqrt{3}}& -\frac{1}{\sqrt{3}}\end{array}\right)\left(\begin{array}{c}PWM\_un\left(k\right)-PWM\_u\left(k\right)\\ PWM\_vn\left(k\right)-PWM\_v\left(k\right)\\ PWM\_wn\left(k\right)-PWM\_w\left(k\right)\end{array}\right)\cdot Udc$$

If you change in the next switching period (k + 1) of the inverter (see 7 ) the original PWM value around the PWM difference between the original and the changed PWM value for each phase (opposite sign), one obtains a modified new voltage of the machine in switching period (k + 1): $$\left(\begin{array}{c}Us\alpha n\left(k+1\right)\\ Us\beta n\left(k+1\right)\end{array}\right)=\left(\begin{array}{c}Us\alpha \left(k\right)\\ Us\beta \left(k\right)\end{array}\right)-\left(\begin{array}{ccc}\frac{2}{3}& -\frac{1}{3}& -\frac{1}{3}\\ 0& \frac{1}{\sqrt{3}}& -\frac{1}{\sqrt{3}}\end{array}\right)\left(\begin{array}{c}PWM\_un\left(k\right)-PWM\_u\left(k\right)\\ PWM\_vn\left(k\right)-PWM\_v\left(k\right)\\ PWM\_wn\left(k\right)-PWM\_w\left(k\right)\end{array}\right)\cdot Udc$$

Calculating the mean value of the machine voltage over the two adjacent switching periods (k) and (k + 1), one obtains the original desired voltage vector of the machine: $$\frac{1}{2}\cdot \left(\begin{array}{c}Us\alpha n\left(k+1\right)+Us\alpha n\left(k\right)\\ Us\beta n\left(k+1\right)+Us\beta n\left(k\right)\end{array}\right)=\left(\begin{array}{c}Us\alpha \left(k\right)\\ Us\beta \left(k\right)\end{array}\right)$$

This can be done by changing the PWM Values of one phase in one inverter switching period (due to the current measurement) and compensation for this change in the next one WR Switching period on average set the same voltage in the inverter. Alternatively, the PWM Values of the previous one WR Period (k-1) can be used for the correction.

To explain the procedure, the following steps are performed: First, the respective PWM Values are determined by a direct comparison You get the smallest PWM -Value PWM1 , the mean PWM value PWM 2 and the biggest PWM -Value PWM3 , The PWM1 . 2 and 3 for example PWM_u or PWM_v or PWM_w to be the three phases of the machine. In each case, a current in one phase Isx1 first sampled and then sampled and calculated a second current in a second phase Isx2 (x = u, v, w) within the switching period Ts. Table 1: Representation of all possible combinations of the PWM value comparison and the corresponding phase current measurements case measured Corresponds to 1-shunt measurement 6 Isu1, Isv2 Case 6: PWM_w>PWM_v> PWM_u 5 Isv1, Isu2 Case 5: PWM_w>PWM_u> PWM_v 4 Isu1, Isw2 Case 4: PWM_v>PWM_w> PWM_u 3 Isw1, Isu2 Case 3: PWM_v>PWM_u> PWM_w 2 Isv1, Isw2 Case 2: PWM_u>PWM_w> PWM_v 1 Isw1, Isv2 Case 1: PWM_u>PWM_v> PWM_w

For each inverter, due to the blocking time and operating situation, a minimum and a maximum switch-on time are defined tmin and tmax (see 8th ). From this one determines a minimum PWM -Value PWM min and maximum PWM -Value PWMmax , which must not be exceeded or fallen short of, so that these switching time limits are met. These two limits must be adhered to in order to guarantee the functionality of the inverter and to avoid damage (for example due to a short circuit). For some applications, tmin = 0 [sec] and tmax = Ts.

The times of the individual switching edges within the switching period of the inverter are in 8th shown. The time difference of the adjacent switching edges can be formed. The relationship between these time differences and the associated PWM Values of the phases can be taken from the following equations. The PWM Differences are used in the developed algorithm.

The PWM Difference ( ΔPWMmin1 ) between the minimum PWM -Value ( PWM min ) and the smallest PWM value ( PWM1 ) is calculated as follows: $$\Delta PWM\text{min1}=PWM1-PWM\text{min}=PWM1-\frac{t_{\text{min}}}{T_s}$$

The PWM -Difference ΔPWM12 between the smallest PWM -Value PWM1 and the middle one PWM -Value PWM 2 is calculated as follows: $$\Delta \mathrm{<figure-callout\; id="12"\; label="P\; W\; M"\; filenames="DE102018206286A1\_0027.png,DE102018206286A1\_0031.png"\; state="\{\{state\}\}">P\; </figure-callout>}WM12=PWM2-\mathrm{<figure-callout\; id="1"\; label="P\; W\; M"\; filenames="DE102018206286A1\_0022.png,DE102018206286A1\_0023.png"\; state="\{\{state\}\}">P\; </figure-callout>}WM1=\frac{\Delta t_{12}}{T_s}$$

The PWM -Difference ΔPWM23 between the mean PWM value PWM 2 and the biggest PWM -Value PWM3 is calculated as follows: $$\Delta \mathrm{<figure-callout\; id="23"\; label="P\; W\; M"\; filenames="DE102018206286A1\_0027.png,DE102018206286A1\_0032.png"\; state="\{\{state\}\}">P\; </figure-callout>}WM23=PWM3-\mathrm{<figure-callout\; id="2"\; label="P\; W\; M"\; filenames="DE102018206286A1\_0022.png,DE102018206286A1\_0023.png"\; state="\{\{state\}\}">P\; </figure-callout>}WM2=\frac{\Delta t_{23}}{T_s}$$

The PWM -Difference ΔPWM3max between the largest PWM -Value PWM3 and the maximum PWM -Value PWMmax is calculated as follows: $$\Delta \mathrm{<figure-callout\; id="3"\; label="P\; W\; M"\; filenames="DE102018206286A1\_0021.png,DE102018206286A1\_0022.png"\; state="\{\{state\}\}">P\; </figure-callout>}WM3\text{Max}=\text{PWM max}-\text{PWM3}=\frac{t_{\text{Max}}}{T_s}-\mathrm{<figure-callout\; id="3"\; label="P\; W\; M"\; filenames="DE102018206286A1\_0021.png,DE102018206286A1\_0022.png"\; state="\{\{state\}\}">P\; </figure-callout>}WM3$$

From the minimum time interval between two adjacent flanks, the minimum PWM -Difference ΔPWMDelay between two adjacent ones PWM Values of the three machine phases as follows: $$\Delta PWMDelay=\frac{TDelay}{Ts}$$

It puts ts the switching period of the inverter (corresponds to 1 / fs, where fs represents the switching frequency of the inverter) and TDelay the minimum time interval between two adjacent edges in the inverter.

For compliance with the minimum time interval between two adjacent flanks in the inverter TDelay can the determined PWM Values are changed. It is tried, the PWM Values as minimal as possible. If between PWM1 and PWM 2 a distance less than ΔPWMDelay is, then you can get the PWM value PWM1 adjust as follows: $$\begin{array}{l}\text{n) PWM1}n=PWM1-\left(\Delta PWMDelay-\Delta PWM12\right)\\ \text{p) PWM1}n=\mathrm{<figure-callout\; id="1"\; label="P\; W\; M"\; filenames="DE102018206286A1\_0022.png,DE102018206286A1\_0023.png"\; state="\{\{state\}\}">P\; </figure-callout>}WM1+\left(\Delta PWMDelay-\Delta PWM12\right)\end{array}$$

Alternatively, also the PWM -Value PWM 2 be adjusted as follows: $$\begin{array}{l}\text{p) PWM2}n=\mathrm{<figure-callout\; id="2"\; label="P\; W\; M"\; filenames="DE102018206286A1\_0022.png,DE102018206286A1\_0023.png"\; state="\{\{state\}\}">P\; </figure-callout>}WM2+\left(\Delta PWMDelay-\Delta PWM12\right)\\ \text{n) PWM2}n=PWM2-\left(\Delta PWMDelay-\Delta PWM12\right)\end{array}$$

Is the distance between PWM 2 and PWM3 less than ΔPWMDelay , then he can PWM -Value PWM 2 be adjusted as follows: $$\begin{array}{l}\text{n) PWM2}n=PWM2-\left(\Delta PWMDelay-\Delta PWM23\right)\\ \text{p) PWM2}n=\mathrm{<figure-callout\; id="2"\; label="P\; W\; M"\; filenames="DE102018206286A1\_0022.png,DE102018206286A1\_0023.png"\; state="\{\{state\}\}">P\; </figure-callout>}WM2+\left(\Delta PWMDelay-\Delta PWM23\right)\end{array}$$

Alternatively, the PWM -Value PWM3 be adjusted as follows: $$\begin{array}{l}\text{p) PWM3}n=\mathrm{<figure-callout\; id="3"\; label="P\; W\; M"\; filenames="DE102018206286A1\_0021.png,DE102018206286A1\_0022.png"\; state="\{\{state\}\}">P\; </figure-callout>}WM3+\left(\Delta PWMDelay-\Delta PWM23\right)\\ \text{n) PWM3}n=PWM3-\left(\Delta PWMDelay-\Delta PWM23\right)\end{array}$$

9 shows a block diagram of an embodiment of the device according to the invention within the FOR. It is between the designated as a vector modulator PWM generator 215 and the inverter 220 a correction device 245 arranged from the individual PWM values PWM_u . PWM_v and PWM_w by means of the method according to the invention, the corrected PWM values PWM_uN . PWM_vN and PWM_wN generated and sent to the inverter 220 provides. The current detection device 235 detects a first and a second current at different times.

10 shows a flowchart of a method 1000 according to an embodiment of the invention. The procedure begins in step 1010 ,

In step 1020 will be out of the the PWM Values as shown in Table 1 above. Six cases will be different. If a corresponding case occurs, then the smallest PWM -Value PWM1 , the middle one PWM -Value PWM 2 and the biggest PWM -Value PWM3 certainly. Then the two times for measuring the DC link current t1 and t2 within the sampling period ts established.

The measured DC link current ISSUM in the time interval Δt12 corresponds to the phase current of the phase with the smallest PWM -Value PWM1 , The measured DC link current ISSUM in the time interval Δt23 corresponds to the negative phase current of the phase with the largest PWM -Value PWM3 ,

ΔTDelay is the minimum time between two adjacent switching edges within a switching period of the inverter, in order to avoid the influence of the switching effects on the 1-shunt measurement. It follows from Equation 15 that the minimum PWM -Difference ΔPWMDelay that between the PWM1 and PWM 2 as well as between PWM 2 and PWM3 to be complied with.

From the individual determined PWM values PWM1 . PWM 2 and PWM3 and the minimum ( pwmmin ) and the maximum PWM -Value ( PWMmax ) you first determine the individual PWM -Differenzen ΔPWMmin1 , from equation 11 to 14.

After that, in step 1030 from the individual determined PWM values PWM1 . PWM 2 and PWM3 as well as the minimum PWM -Value pwmmin and the maximum PWM who PWMMax the values for ΔPWMmin1 . ΔPWM12 . ΔPWM23 and ΔPWM3max calculated using the equations 11, 12, 13 and 14 listed in detail above.

Based on in step 1030 calculated results will be in the steps 1040 the values for ΔPWMmin1 and ΔPWM3max with the value ΔPWMDelay compared. This results in four different case constellations: In the case 1041 is both ΔPWMmin1 as well as ΔPWM3max greater or equal ΔPWMDelay , This value pair for to the method as in 11 explained in more detail. In the case 1042 is ΔPWMmin1 smaller ΔPWMDelay and ΔPWM3max greater or equal ΔPWMDelay ,

This value pair leads to the procedure as in 12 explained in more detail. In the case 1043 is ΔPWMmin1 greater or equal ΔPWMDelay and ΔPWM3max less than ΔPWMDelay , This value pair leads to the procedure as in 13 explained in more detail. In the case 1044 is both ΔPWMmin1 as well as ΔPWM3max less than ΔPWMDelay. This value pair for to the method as in 14 explained in more detail.

The respective run of the in the 11 to 14 closer explained process steps then provides the new values for PWM1n . PWM2n and PWM3n back. These are in the step 1050 the respective phases U . V . W assigned. This results in the new values for PWM_uN . PWM_vN and PWM_wN which in the step 1060 be provided to the inverter. In step 1070 the procedure ends 1000 ,

11 shows a further embodiment of the method according to the invention 1100 , The procedure begins in step 1110 , After that, by the method 1000 as in 10 shown values obtained for ΔPWMmin1 . ΔPWM12 . ΔPWM23 and ΔPWM3max such as ΔPWMDelay pass and step 1120 checked that both ΔPWMmin1 as well as ΔPWM3max greater or equal ΔPWMDelay is like in step 1040 determined. This corresponds to the case constellation 1041 , Based on this case constellation are now in step 1130 distinguished four sub-constellations: In the case 1131 is both ΔPWM12 as well as ΔPWM23 greater or equal ΔPWMDelay , This value pair is followed by the steps described below 1141 and 1151 at. In the case 1132 is ΔPWM12 smaller ΔPWMDelay and ΔPWM23 greater or equal ΔPWMDelay , This value pair is followed by the steps described below 1142 and 1152 at. In the case 1133 is ΔPWM12 greater or equal ΔPWMDelay and ΔPWM23 less than ΔPWMDelay , This value pair is followed by the steps described below 1143 and 1153 at. In the case 1134 is both ΔPWM12 as well as ΔPWM23 less than ΔPWMDelay , This value pair is followed by the steps described below 1144 and 1154 at.

In step 1140 then, based on the in step 1130 If necessary, correct different subgroup constellations. Likewise, in the step 1150 based on the in step 1130 if necessary, corrected again for the different subgroup constellations. At the same time the step takes place 1140 for a first switching period and the step 1150 for a second switching period.
For the one in step 1130 different case 1131 be both in step 1141 as well as in the step 1151 no changes to the values PWM1n (k) . PWM2n (k) and PWM3n (k) respectively. PWM1n (k + 1) . PWM2n (k + 1) and PWM3n (k + 1) performed. These are each PWM1 . PWM 2 and PWM3 ,

For the one in step 1130 different case 1132 remain the values for PWM2n (k) and PWM3n (k) respectively. PWM2n (k + 1) and PWM3n (k + 1) each unchanged. For the value PWM1n (k) However, this is according to equation 16n in step 1142 and for the value PWM1n (k + 1) according to equation 16p in step 1152 corrected.

For the one in step 1130 different case 1133 remain the values for PWM1n (k ) and PWM2n (k) such as PWM1n (k + 1) and PWM2n (k + 1) each unchanged. For the value PWM3n (k) However, this is according to equation 19p in step 1143 and for the value PWM3n (k + 1) according to equation 19n in step 1153 corrected.

For the one in step 1130 different case 1134 remain the values for PWM2n (k) and PWM2n (k + 1) each unchanged. For the value PWM3n (k) However, this is according to equation 19p in step 1144 and for the value PWM3n (k + 1) according to equation 19n in step 1154 corrected. Likewise, for the value PWM1n (k) this according to equation 16n in step 1144 and for the value PWM1n (k + 1) according to equation 16p in step 1154 corrected.

In step 1160 will be the values for PWM1n (k) . PWM1n (k + 1) . PWM2n (k) . PWM2n (k + 1) . PWM3n (k) and PWM3n (k + 1) to the procedure 1000 returned and then continued as described above. In step 1170 is the procedure 1100 completed.

12 shows a further embodiment of the method according to the invention 1200 , The procedure begins in step 1210 , After that, by the method 1000 as in 10 shown values obtained for ΔPWMmin1 . ΔPWM12 . ΔPWM23 and ΔPWM3max such as ΔPWMDelay pass and step 1220 checked that ΔPWMmin1 less than ΔPWMDelay and ΔPWM3max greater or equal ΔPWMDelay is like in step 1040 determined. This corresponds to the case constellation 1042 , Based on this case constellation are now in step 1230 distinguished three sub-constellations: In the case 1231 is both ΔPWM12 as well as ΔPWM23 greater or equal ΔPWMDelay , This value pair is followed by the steps described below 1241 and 1251 at. In the case 1232 is ΔPWM12 smaller ΔPWMDelay and ΔPWM23 greater or equal ΔPWMDelay , For this value pair close, after another case distinction 1235 which is described below, the steps described below 1242 and 1252 respectively. 1243 and 1253 at. In the case 1233 is ΔPWM12 greater or equal ΔPWMDelay and ΔPWM23 less than ΔPWMDelay , This value pair is followed by the steps described below 1244 and 1254 at. The case that both ΔPWM12 as well as ΔPWM23 less than ΔPWMDelay are for this, in step 1220 described case constellation not possible.

In step 1240 then, based on the in step 1230 If necessary, correct different subgroup constellations. Likewise, in the step 1250 based on the in step 1230 if necessary, corrected again for the different subgroup constellations. At the same time the step takes place 1240 for a first switching period and the step 1250 for a second switching period.

For the one in step 1230 different case 1231 be both in step 1241 as well as in the step 1251 no changes to the values PWM1n (k) . PWM2n (k) and PWM3n (k) respectively. PWM1n (k + 1) . PWM2n (k + 1) and PWM3n (k + 1) performed. These are each PWM1 . PWM 2 and PWM3 , analogous to the steps 1141 and 1151 ,

For the one in step 1230 different case 1232 will be in step 1235 made a further case distinction. It is discriminated whether ΔPWM12 + ΔPWMmin1 is greater than or equal to ΔPWMDelay is or not. If so, the values for PWM2n (k) and PWM3n (k) such as PWM2n (k + 1) and PWM3n (k + 1) each unchanged. For the value PWM1n (k) However, this is according to equation 16n in step 1242 and for the value PWM1n (k + 1) according to equation 16p in step 1252 corrected, analogous to in 11 process shown in the steps 1142 and 1152 ,

If no, the values for PWM1n (k) and PWM3n (k) such as PWM1n (k + 1) and PWM3n (k + 1) each unchanged. For the value PWM2n (k) this is according to equation 17p in step 1243 and for the value PWM2n (k + 1) according to equation 17n in step 1253 calculated.

For the one in step 1230 different case 1233 remain the values for PWM1n (k) and PWM2n (k) such as PWM1n (k + 1) and PWM2n (k + 1) each unchanged. For the value PWM3n (k) this is according to equation 19p in step 1244 and for the value PWM3n (k + 1) according to equation 19n in step 1254 corrected, analogous to in 11 process shown in the steps 1143 and 1153 ,

In step 1260 will be the values for PWM1n (k) . PWM1n (k + 1) . PWM2n (k) . PWM2n (k + 1) . PWM3n (k) and PWM3n (k + 1) to the procedure 1000 returned and then continued as described above. In step 1270 is the procedure 1200 completed.

13 shows a further embodiment of the method according to the invention 1300 , The procedure begins in step 1310 , After that, by the method 1000 as in 10 shown values obtained for ΔPWMmin1 . ΔPWM12 . ΔPWM23 and ΔPWM3max such as ΔPWMDelay pass and step 1320 checked that ΔPWMmin1 greater or equal ΔPWMDelay and ΔPWM3max less than ΔPWMDelay is like in step 1040 determined. This corresponds to the case constellation 1043 , Based on this case constellation are now in step 1330 distinguished three sub-constellations: In the case 1331 is both ΔPWM12 as well as ΔPWM23 greater or equal ΔPWMDelay , This value pair is followed by the steps described below 1341 and 1351 at. In the case 1332 is ΔPWM12 smaller ΔPWMDelay and ΔPWM23 greater or equal ΔPWMDelay , This value pair is followed by the steps described below 1342 and 1352 at. In the case 1233 is ΔPWM12 greater or equal ΔPWMDelay and ΔPWM23 less than ΔPWMDelay , For this value pair close, after another case distinction 1335 which is described below, the steps described below 1343 and 1353 respectively. 1344 and 1354 at. The case that both ΔPWM12 as well as ΔPWM23 less than ΔPWMDelay are for this, in step 1320 described case constellation not possible.

In step 1340 then, based on the in step 1330 If necessary, correct different subgroup constellations. Likewise, in the step 1350 based on the in step 1330 if necessary, corrected again for the different subgroup constellations. At the same time the step takes place 1340 for a first switching period and the step 1350 for a second switching period.

For the one in step 1330 different case 1331 be both in step 1341 as well as in the step 1351 no changes to the values PWM1n (k) . PWM2n (k) and PWM3n (k) respectively. PWM1n (k + 1) . PWM2n (k + 1) and PWM3n (k + 1) performed. These are each PWM1 . PWM 2 and PWM3 , analogous to the steps 1141 and 1151 respectively. 1241 and 1251 ,

For the one in step 1330 different case 1332 remain the values for PWM2n (k) and PWM3n (k) respectively. PWM2n (k + 1) and PWM3n (k + 1) each unchanged. For the value PWM1n (k) However, this is according to equation 16n in step 1342 and for the value PWM1n (k + 1) according to equation 16p in step 1252 corrected, analogous to the steps 1142 and 1152 respectively. 1242 and 1252 ,

For the one in step 1330 different case 1333 will be in step 1235 made a further case distinction. It is discriminated whether ΔPWM23 + ΔPWM3max is greater than or equal to ΔPWMDelay is or not. If so, the values for PWM1n (k) and PWM2n (k) respectively. PWM1n (k + 1) and PWM2n (k + 1) each unchanged. For the value PWM3n (k) this is according to equation 19p in step 1343 and for the value PWM3n (k + 1) according to equation 19n in step 1353 corrected, analogous to the steps 1143 and 1153 respectively. 1244 and 1254 , If no, the values for PWM1n (k) and PWM3n (k) respectively. PWM1n (k + 1) and PWM3n (k + 1) each unchanged. For the value PWM2n (k) this is according to equation 18n in step 1344 and for the value PWM2n (k + 1) according to equation 18p in step 1354 corrected.

In step 1360 will be the values for PWM1n (k) . PWM1n (k + 1) . PWM2n (k) . PWM2n (k + 1) . PWM3n (k) and PWM3n (k + 1) to the procedure 1000 returned and then continued as described above. In step 1370 is the procedure 1300 completed.

14 shows a further embodiment of the method according to the invention 1400 , The procedure begins in step 1410 , After that, by the method 1000 as in 10 shown values obtained for ΔPWMmin1 . ΔPWM12 . ΔPWM23 and ΔPWM3max such as ΔPWMDelay pass and step 1420 checked that both ΔPWMmin1 as well as ΔPWM3max less than ΔPWMDelay are as in the step 1040 determined. This corresponds to the case constellation 1044 , Based on this case constellation are now in step 1430 distinguished three sub-constellations: In the case 1431 is both ΔPWM12 as well as ΔPWM23 greater or equal ΔPWMDelay , This value pair is followed by the steps described below 1441 and 1451 at. In the case 1432 is ΔPWM12 smaller ΔPWMDelay and ΔPWM23 greater or equal ΔPWMDelay , For this value pair close, after another case distinction 1435 which is described below, the steps described below 1442 and 1452 respectively. 1443 and 1453 at. In the case 1433 is ΔPWM12 greater or equal ΔPWMDelay and ΔPWM23 less than ΔPWMDelay , For this value pair close, after another case distinction 1436 which is described below, the steps described below 1444 and 1454 respectively. 1445 and 1455 at. The case that both ΔPWM12 as well as ΔPWM23 less than ΔPWMDelay are for this, in step 1420 described case constellation not possible.

In step 1440 then, based on the in step 1430 If necessary, correct different subgroup constellations. Likewise, in the step 1450 based on the in step 1430 if necessary, corrected again for the different subgroup constellations. At the same time the step takes place 1440 for a first switching period and the step 1450 for a second switching period.

For the one in step 1430 different case 1431 be both in step 1441 as well as in the step 1451 no changes to the values PWM1n (k) . PWM2n (k) and PWM3n (k) respectively. PWM1n (k + 1) . PWM2n (k + 1) and PWM3n (k + 1) performed. These are each PWM1 . PWM 2 and PWM3 , analogous to the steps 1141 and 1151 . 1241 and 1251 respectively. 1341 and 1351 ,

For the one in step 1430 different case 1432 will be in step 1435 made a further case distinction. It is discriminated whether ΔPWM12 + ΔPWMmin1 is greater than or equal to ΔPWMDelay is or not. If so, the values for PWM2n (k) and PWM3n (k) such as PWM2n (k + 1) and PWM3n (k + 1) each unchanged. For the value PWM1n (k) However, this is according to equation 16n in step 1442 and for the value PWM1n (k + 1) according to equation 16p in step 1452 corrected, analogous to the steps 1142 and 1152 . 1242 and 1252 respectively. 1342 and 1352 , If no, the values for PWM1n (k) and PWM3n (k) such as PWM1n (k + 1) and PWM3n (k + 1) each unchanged. For the value PWM2n (k) this is according to equation 17p in step 1443 and for the value PWM2n (k + 1) according to Equation 17n in step 1453 corrected, analogous to the steps 1243 and 1253 ,

For the one in step 1430 different case 1433 will be in step 1436 made a further case distinction. It is discriminated whether ΔPWM23 + ΔPWM3max is greater than or equal to ΔPWMDelay is or not. If so, the values for PWM1n (k) and PWM2n (k) respectively. PWM1n (k + 1) and PWM2n (k + 1) each unchanged. For the value PWM3n (k) this is according to equation 19p in step 1343 and for the value PWM3n (k + 1) according to equation 19n in step 1353 corrected, analogous to the steps 1143 and 1153 . 1244 and 1254 respectively. 1343 and 1353 , If no, the values for PWM1n (k) and PWM3n (k) respectively. PWM1n (k + 1) and PWM3n (k + 1) each unchanged. For the value PWM2n (k) this is according to equation 18n in step 1444 and for the value PWM2n (k + 1) according to equation 18p in step 1454 corrected, analogous to the steps 1344 and 1354 ,

In step 1460 will be the values for PWM1n (k) . PWM1n (k + 1) . PWM2n (k) . PWM2n (k + 1) . PWM3n (k) and PWM3n (k + 1) to the procedure 1000 returned and then continued as described above. In step 1470 is the procedure 1400 completed.

Are the new PWM values PWM1n . PWM2n and PWM3n from the respective procedures as in the 11 . 12 . 13 or 14 such as 10 determined, so they are on the current detection device 235 in 9 forwarded to perform the DC link current sampling of the two measurements without parasitic effects of the switching edge or oscillations by corresponding triggers of the ADC converter. Based on the investigations from the flowchart as in 10 shown (with reference to Table 1), the individual PWM values are assigned to the associated phases ( u . v . w ). Finally, these are sent to the inverter 220 forwarded (see 9 ), where you should set the nominal voltage.

• 1501, 1601 PWM_u/PWM_uN: PWM-Wert in Phase U original und modifiziert
• 1502, 1602 PWM_v/ PWM_vN: PWM-Wert in Phase V original und modifiziert
• 1503, 1603 PWM_w/PWM_wN: PWM-Wert in Phase W original und modifiziert
• 1504, 1604 Amplitude des Spannungsvektors original und modifiziert in Volt
• 1505, 1605 Phase des Spannungsvektors original und modifiziert in Grad
• 1506, 1606 Phasenströme der Maschine Isu, Isv, Isw in Ampere

In 15 and 16 are each simulations for a run-up 1500 . 1600 a machine with different current measurements shown. It is shown from top to bottom:

• 1501, 1601 PWM_u / PWM_uN: PWM value in phase U original and modified
• 1502, 1602 PWM_v / PWM_vN: PWM value in phase V original and modified
• 1503, 1603 PWM_w / PWM_wN: PWM value in phase W original and modified
• 1504, 1604 Amplitude of voltage vector original and modified in volts
• 1505, 1605 Phase of the voltage vector original and modified in degrees
• 1506, 1606 phase currents of the machine Isu . isv . sw in amps

In 15 is the course 1500 The machine is shown with an ideal current measurement (eg: a LEM converter as a current sensor) (no 1-shunt measurement). There are no PWM changes needed. The phase currents are sinusoidal and show no distortions. The desired amount of tension of the machine increases continuously.

In 16 became the same course 1600 the machine as in 15 with the same regulator block diagram as in 9 shown performed. It is in the courses according to 16 the control with 1-shunt measurement for the current measurement carried out according to an embodiment of the method according to the invention. An exemplary switching frequency of the inverter of 16 kHz gives a WR Switching period of Ts = 62.5μs. The control cycle (TA = 250μs) is four times the WR switching period Ts. To maintain a minimum time interval of TDelay = 6.25μs between two adjacent switching edges (corresponding to a minimum PWM difference of 10%), the PWM values of the individual phases based on the flowcharts as in the 10 to 14 shown changed.

It can be seen that the individual PWM values depend on the case found according to the flowcharts as in 10 to 14 have been modified. This results in a change in the voltage amplitude and phase in order to perform the minimum time interval between the corresponding adjacent switching edges for the 1-shunt current measurement.

LIST OF REFERENCE NUMBERS

100

Induction machine

101

rotor

102

axis of rotation

103

permanent magnet

205

control component

210

converter

215

vector modulator

220

inverter

225

Current measuring sensors

230

position sensor

235

Current detection device

245

corrector

240

decoupler

305

DC

310

Link capacitor

315

shunt

700

sampling

1000-1470

steps

1500-1606

Simluationsergebnisse

k

k -th switching period

(K + 1)

k + 1 te switching period

PWM1

Pulse width modulation value

PWM 2

Pulse width modulation value

PWM3

Pulse width modulation value

S1

flow control valve

S2

flow control valve

S3

flow control valve

S4

flow control valve

S5

flow control valve

S6

flow control valve

t1

time

t2

time

t3

time

TA

control period

TS

switching period

U

phase

V

phase

W

phase

Zp

number of pole pairs

Θmech

mechanical angle of rotation

Θel

electrical angle of rotation

Claims (9)

A method of measuring the current of an inverter (220) having three half-bridges, the method comprising the steps of: Determining a respective PWM value associated with each of the half-bridges of the inverter for a first switching period; Determining if predetermined minimum clearances are maintained between the PWM values; if the minimum distances between the PWM values are not met: Identifying a critical PWM value from the PWM values; Determining a correction value by which the critical PWM value must be corrected so that the minimum distances are maintained; Correcting the critical PWM value by the correction value; and Correcting a corresponding PWM value for a second switching period of the inverter, wherein the corresponding PWM value is associated with the same half-bridge as the critical PWM value; the sum of all corrections being 0; Detecting a first current at a first time (t1) within the first switching period in which a first half-bridge in a second switching state and a second and a third half-bridge are in a first switching state; and Detecting a second current at a second time (t2) within the first switching period in which the first half bridge and the second half bridge are in a second switching state and the third half bridge is in a first switching state, or in which the first half bridge and the third Half-bridge are in a second switching state and the second half-bridge are in a first switching state. Method according to one of the preceding claims, characterized in that the first switching period follows the second switching period. The method of any one of the preceding claims, wherein the method further comprises: Correcting a corresponding PWM value for a third switching period, wherein the corresponding PWM value is assigned to the same half-bridge as the critical PWM value. Method according to one of the preceding claims, wherein the second and / or the third switching period adjacent to the first switching period. Method according to one of the preceding claims, wherein the minimum distance between 0 and a minimum of the PWM values, and the minimum distance between a maximum of the PWM values and 1, a quotient of a operating condition-related blocking time and a switching period duration, and wherein the minimum distance between the is the minimum PWM value and a next larger one of the PWM values, and the minimum distance between the largest PWM value and a next smaller one of the PWM values, a quotient of a component-related minimum time interval and the switching period duration. Computer program product with program code means for carrying out a method according to one of the preceding Claims 1 - 5 , An apparatus for measuring the current of an inverter (220) having three half-bridges, comprising: a correction device (245) configured to detect and supply one PWM value each associated with one of the half-bridges of the inverter for a first switching period determining whether predetermined minimum clearances are maintained between the PWM values, wherein the correction device is further configured, if the minimum distances between the PWM values are not met, to identify a critical PWM value from the PWM values, to determine a correction value, by which the critical PWM value must be corrected so that the minimum distances are kept, to correct the critical PWM value by the correction value; and correct a corresponding PWM value for a second switching period of the inverter, the corresponding PWM value being associated with the same half-bridge as the critical PWM value, and the sum of all corrections being 0; and a current sensing device (235) configured to detect a first current at a first time within the first switching period in which a first half-bridge is in a second switching state and a second and third half-bridges are in a first state; and detect a second current at a second time within the first switching period, in which the first half-bridge and the second half-bridge are in a second switching state and the third half-bridge is in a first switching state, or in which the first half-bridge and the third half-bridge in a second switching state and the second half-bridge are in a first switching state. Device after Claim 7 wherein the correction device (245) is further configured to correct a corresponding PWM value for a third switching period, wherein the corresponding PWM value is associated with the same half-bridge as the critical PWM value. A system comprising an induction machine (100), an inverter (220) and a current measuring device Claim 7 or 8th ,

Images