DE102012210671A1 - Method for controlling inverter by space pointer modulation for controlling electric machine in motor vehicle drive strand, involves determining sets of controlling periods in dependence of predetermined load reference values of inverter - Google Patents

Abstract

The method involves providing an inverter (10) with a set of controllable power semiconductor switches, and providing an electric machine (14) with multiphase electric currents (IU, IV, IW). The inverter is controlled such that switching states of one of the switches are set up. The inverter is switched to voltage-free switchable switching states. The inverter is controlled for two sets of controlling periods using a set of sequences, where the sets of controlling periods are determined in dependence of predetermined load reference values of the inverter. Independent claims are also included for the following: (1) a device for controlling an inverter to control an electric machine (2) a motor vehicle drive strand with an electric machine.

Description

The present invention relates to a method for controlling an inverter by means of space vector modulation, in particular for driving an electrical machine, wherein the inverter has a plurality of controllable switches, and which is adapted to provide a multi-phase electric current, in particular to the electric machine polyphase with electrical Power supply, wherein the inverter is driven to set up a plurality of successive different switching states of the switches, and wherein the inverter is switched by means of two switching states in a de-energized switching state, the inverter is controlled by at least two predefined sequences of switching states, have the different on times of the switching states.

Furthermore, the present invention relates to a device for driving an inverter, in particular for driving an electrical machine, wherein the inverter comprises a plurality of controllable switches which are connected to provide a multi-phase electric current, in particular to supply the electric machine polyphase with electric current with a control unit, which is designed to control the inverter, that the inverter assumes a plurality of successive different switching states of the switches, wherein the inverter 2 can adopt voltage-free switching switching states, wherein the inverter can be controlled by means of at least two predefined sequences of switching states having different switch-on times of the switching states.

Finally, the present invention relates to a motor vehicle drive train with at least one electric machine for providing drive power, an alternating direction for driving the electric machine and a device for driving the inverter of the type described above.

State of the art

In the technical field of three-phase current consumers in general and three-phase electrical machines in particular, different driving methods are known. Currently, the method of space vector modulation for controlling the three-phase current consumer is currently preferred. In this driving method, a space pointer is formed by sequentially setting eight ground voltage hands. In order to provide the string voltage, the ground voltage phasors are switched pulse width modulated, so that a corresponding drive voltage is generated in the mean value.

In the known driving method, the electrical loads are controlled by means of an inverter with power semiconductor switches. The setting of the eight ground voltage pointer for generating the voltage space pointer is realized by alternately turning on and off certain power semiconductor switches of the inverter. At very low rotational speeds of the space hand or, if the three-phase load is an electrical machine, at low speeds of the driven electrical machines, individual power semiconductor switches are very common or switched on for a very long time and thus thermally loaded by a very long or very frequently flowing electric current. Therefore, the power semiconductor switches must be designed for very long turn-on times and for very large currents, whereby the inverter is generally technically complex.

In order to counteract one, in particular thermal, overloading of the power semiconductor switches is in the WO 2010 / 000548A2 proposed to omit one of two de-energized switching states in certain pulse width modulation periods in order to reduce the switching losses of the power semiconductor switches.

The disadvantage here is that one, in particular thermal, load distribution of the power semiconductor switch can not be targeted and thus peak loads of individual power semiconductor switch must be taken into account.

Disclosure of the invention

According to the invention, therefore, a method is provided for driving an inverter by means of space vector modulation of the aforementioned type, in which the inverter is driven for a first number of drive periods by means of a first of the sequences and is driven for a second number of drive periods by means of a second of the sequences the first number and the second number are determined in dependence on a predetermined load setpoint of the inverter.

Furthermore, an apparatus for driving an inverter of the type mentioned above is therefore provided according to the invention, wherein the controller is configured to drive the inverter for a first number of drive periods by means of a first of the sequences and for a second number of drive periods to control by means of a second of the sequences, wherein the control device is adapted to determine the first number and the second number in response to a predetermined load setpoint of the inverter.

Finally, according to the invention, a motor vehicle drive train is provided with an electric machine for providing drive power, an inverter for driving the electric machine and with a device for driving the inverter of the type described above.

Advantages of the invention

By means of the present invention, a method and a device can be provided in which a predetermined, in particular thermal, load setpoint value of the inverter can be adjusted by simple means in order to avoid corresponding overloading of the power semiconductor switches. Because the load values of the two drive sequences are known and can be changed as desired between the two sequences, the control effort is low and a virtually arbitrary load value of the power semiconductor switches can be set. Therefore, with the inventive method and the device according to the invention at low speeds, the power semiconductor switch of the inverter can be loaded evenly and overloading of individual power semiconductor switch can be avoided. As a result, the power semiconductor switches can thus be made smaller, making the manufacture of the inverter generally less expensive and less expensive. Furthermore, the control effort for driving the inverter is thereby reduced, since fewer parameters must be taken into account, whereby in particular the device for driving the inverter can be produced more cost-effectively and with technically simpler means due to the reduced computational complexity.

Preferably, both sequences set an identical voltage space vector.

As a result, the control of the connected load is not affected by changing the sequences.

It is of particular advantage if the two predefined sequences have different switch-on times of the two voltage-free-switching switching states.

This can be achieved with simple and effective means relieving one or more of the power semiconductor switches, since any change between these two switching states has no effect on the voltage space vector, but has a large effect on the load of each of the power semiconductor switches.

It is furthermore preferred if the first number and the second number are limited.

This allows a dynamic control of the load setpoint trouble-free.

It is further preferable that a ratio value is formed based on the load target value and reference load values of the two sequences.

As a result, a ratio of the two sequences can be determined with little effort to each other to achieve the predetermined load setpoint.

It is furthermore preferred if the two sequences are selected from a plurality of sequences whose load value has the least difference from the load setpoint.

As a result, the predetermined load setpoint can be set specifically with slight deviations.

It is further preferred if the first and second numbers are determined based on the ratio.

As a result, the load setpoint of the inverter can be set with little control effort and in a short time.

It is of particular advantage if the first or the second number is determined by rounding off a quotient which is determined on the basis of the ratio value.

As a result, it is possible with little computational effort to determine a whole-line value, on the basis of which the activation sequences are used and thus a load value is achieved which has only slight deviations from the load setpoint.

It is generally preferred if a binary number sequence is determined on the basis of the ratio value and the inverter is driven according to the numerical values of the binary number sequence by means of the first or the second predefined sequence.

This allows an optimal ratio of the drive by means of the two sequences be calculated in order to achieve the smallest possible deviations from the load setpoint.

It is understood that the features, properties and advantages of the method according to the invention also apply to the device according to the invention or are applicable.

Brief description of the drawings

1 shows in schematic form an inverter for driving a three-phase load, in particular a three-phase machine;

2 shows a complex phasor diagram for explaining the space vector modulation method for driving the inverter;

3 shows in schematic form the voltage curve of three phases of the inverter within a pulse width modulation period for setting a voltage space vector;

4 shows in schematic form a drive sequence by means of two control sequences; and

5a and 5b show in schematic form flowcharts for the determination of binary number sequences for driving the inverter.

Embodiments of the invention

In 1 is an inverter for driving a three-phase load, in particular an electric machine shown schematically and generally with 10 designated.

The inverter 10 is with a DC voltage source 12 connected and serves to the three-phase load 14 , which in this case as a three-phase electric machine 14 is designed to energize three-phase. The inverter has three half-bridges, which are parallel to the DC voltage source 12 are connected and each have two controllable switch S. Between the switches S is in each case a half-bridge tap 16 formed, each with a phase conductor of the phases U, V, W of the electric machine 14 are connected.

Parallel to the switches S, a freewheeling diode D is in each case connected, which allows a current flow in the opposite direction.

In 1 are the switches S corresponding to the phase U, V, W, they provide and according to the assignment to a high potential of the DC voltage source 12 or a low potential of the DC voltage source 12 denoted by SAH, SLA, SHB, SLB, SHC, SLC. Accordingly, the freewheeling diodes are designated DHA, DLA, DHB, DLB, DHC, DLC.

By alternately opening and closing the switch S, a drive voltage is applied in each case between the phase conductors U, V, W, so that in each case a phase current IU, IV, IW adjusts, which sets the electric machine 14 drives. The inverter 10 is preferably formed by means of semiconductor switches. The switches of the inverter are by means of a control unit shown schematically 18 alternately opened and closed, to provide the phase voltages with a certain gradient, and provide a corresponding voltage space pointer and the electric machine 14 to energize accordingly with the phase currents IU, IV, IW.

In 2 is a complex phasor diagram for explaining the space vector modulation for driving the three-phase current consumer 14 or the electric machine 14 represented and generally with 20 designated.

In the pointer diagram 20 is a voltage vector V * with a drive angle alpha of the electric machine 14 shown. In the pointer diagram 20 Also shown are six ground voltage indicators V1, V2, V3, V4, V5, V6, which result when one or two of the switches S of the inverter 10 be closed and the electric machine is controlled accordingly. In order to set the voltage vector V *, which in this example has the driving angle alpha between the ground voltage indicators V1 and V2, this is achieved by alternately driving the inverter 10 realized in accordance with the base voltage vector V1 and the ground voltage vector V2. The two ground voltage vectors V1, V2 are alternately set with a predefined switching frequency, so that at the same duty cycle of the ground voltage vector V1, V2, the voltage vector V * results in a phase angle alpha of 30 °. If a voltage vector V * has to be set with a larger activation angle alpha, the on-time of the basic voltage vector V2 is correspondingly extended and the on-time of the basic voltage vector V1 is shortened. Thus, the switch S of the inverter can be controlled by pulsed driving 10 the voltage space vector V * realize with an arbitrary drive angle alpha.

If the voltage vector V *, as in the in 2 illustrated case with a smaller amount (shorter length) than the basic voltage space vector V1, V2 is to be set, is set according to a zero voltage pointer V0, V7, in which the switch SHA, SHB, SHC on the upper side or SLA, SLB, SLC on the lower side of the inverter 10 are open. The respective other switches S are closed accordingly. Accordingly, the voltage vector V * can be realized by a combination of the basic voltage space vector V1 and V2 and one of the zero voltage V0, V7.

Depending on the voltage space vector V *, a current space vector I * is established. The current space vector I * has an amplitude and a phase angle, which depend on the driven electrical load 14 to adjust. The phase angle of the current space vector I * can be in phase with the phase angle α of the voltage space vector V * or can have a phase shift.

For energizing the three-phase current consumer 14 or the electric machine 14 the voltage space vector V * is provided by sequentially adjusting the different fundamental voltage space vectors V1-V6 and the zero voltage pointers V0, V7 in a fast sequence. As a result, the different switches S and the different freewheeling diodes D of the inverter 10 evenly loaded at a correspondingly fast rotating voltage space vector V *. If the rotational frequency of the voltage space vector V * is very low or zero, for example, at low speeds of the electric machine 10 , the corresponding switches S and the diodes D of the inverter 10 loaded over a long period of time, so that an overload of the corresponding switches S and the diodes D may occur and the switches S and the diodes D of the inverter 10 generally unevenly loaded. In order to prevent overloading of each of the switches S and diodes D, measures must be taken to distribute the load to different ones of the switches S and the diodes D.

In 3 are waveforms of the phase voltages of the three phases U, V, W within a pulse width modulation period T shown to adjust the voltage vector V0, V1, V2, V7 successively. Within the pulse-width modulation period T, the duration of the individual voltage vectors can be varied in order to be able to set the voltage space vector V * precisely.

wobei m der Belastungswert ist, t0 die Schaltdauer des Nullspannungszeigers V0 und t7 die Schaltdauer des Nullspannungszeigers V7. Der Belastungswert m kann folglich zwischen dem Wert 1 und dem Wert 0 eingestellt werden, wobei bei einem Wert 1 die oberen Schalter SHA, SHB, SHC bzw. die oberen Freilaufdioden DHA, DHB, DHC des Wechselrichters 10 stärker belastet werden und bei einem Wert 0 die unteren Schalter SLA, SLB, SLC bzw. unteren Freilaufdioden DLA, DLB, DLC des Wechselrichters 10 stärker belastet werden. Dies gilt speziell für die Berechnung des Belastungswertes m sofern zwei benachbarte Grundspannungsraumzeiger V1–V6 und wenigstens ein Nullspannungszeiger V0, V7 verwendet wird, wie beispielsweise für eine Ansteuerungssequenz nach 3.Since the zero voltage pointer V0 only the switches SLA, SLB, SLC, the low potential of the voltage source 12 are assigned, loaded, and there according to the zero voltage pointer V7, only the switches SHA, SHB, SHC, the high potential of the DC voltage source 12 are assigned, can be compensated by a targeted distribution of the zero voltage pointer V0, V7 uneven load between the upper switches THA, THB, THC and the lower switches TLA, TLB, TLC. A corresponding load value m is calculated by the formula

where m is the load value, t0 is the duty cycle of the zero voltage pointer V0 and t7 is the duty cycle of the zero voltage sense V7. The load value m can thus be set between the value 1 and the value 0, with a value 1, the upper switch SHA, SHB, SHC or the upper freewheeling diodes DHA, DHB, DHC of the inverter 10 more heavily loaded and at a value 0, the lower switches SLA, SLB, SLC and lower freewheeling diodes DLA, DLB, DLC of the inverter 10 be charged more heavily. This is especially true for the calculation of the load value m if two neighboring basic voltage space vector V1-V6 and at least one zero voltage vector V0, V7 is used, as for example for a drive sequence 3 ,

und für einen der Schalter SLA, SLB, SLC der unteren Seite des Wechselrichters 10 durch:

wobei t_min die minimal mögliche Einschaltdauer des Schalters S, t_max die maximal mögliche Einschaltdauer des Schalters S ist, um den aktuellen Spannungsraumzeiger V* einzustellen und t_on die aktuelle Einschaltdauer in der entsprechenden Pulsweitenmodulationsperiode T ist, die den aktuellen Spannungsraumzeiger V* einstellt. Dadurch entsteht ein weiterer Freiheitsgrad, um lsm einstellen zu können. Dabei ist anzumerken, dass der allgemeine Belastungswert lsm gleich bleibt, unabhängig davon welcher der Schalter zur Berechnung herangezogen wird. Dadurch kann die Belastung eines Schalters S des Wechselrichters 10 berücksichtigt werden, um zwischen den beiden Sequenzen A, B hin und her zu schalten. Dabei kann der Belastungswert m durch den allgemeinen Belastungswert lsm ersetzt werden.In general, a general load value lsm can be determined on the basis of the general switch-on time t of one of the switches S. In this case, one of the switches S is selected, such as the one that is heavily loaded. The load value lsm is for one of the switches SHA, SHB, SHC the upper side of the inverter 10 calculated by:

and for one of the switches SLA, SLB, SLC of the lower side of the inverter 10 by:

where t_min is the minimum possible duty cycle of the switch S, t_max is the maximum duty cycle of the switch S to set the current voltage space pointer V * and t_on is the current duty cycle in the corresponding pulse width modulation period T setting the current voltage space pointer V *. This creates another degree of freedom to set lsm. It should be noted that the general load value lsm remains the same, regardless of which of the switches is used for the calculation. This allows the load of a switch S of the inverter 10 be considered to switch between the two sequences A, B back and forth to switch. In this case, the load value m can be replaced by the general load value lsm.

According to the invention, two sequences of consecutive switching states V0-V7 are used to generate a given voltage space vector V * having a certain magnitude and angle alpha. The different sequences have different sequences of switching states V0-V7, so that the switches S of the inverter 10 be charged differently. For each of the two sequences, a corresponding load value m1, m2 is determined. The inverter is set to set the determined voltage space vector V * 10 controlled alternately with the first of the two sequences and the second of the two sequences, wherein both sequences A, B set the same voltage space vector V *. The inverter 10 is driven for a certain number of pulse width modulation periods T by means of the first of the two sequences A, B and driven for a second number of pulse width modulation periods T by means of the second sequence. As a result, a desired load value m, which lies between the load values m1, m2, can be set by the known different load values m1, m2 of the two sequences.

In 4 the inventive method is shown schematically and generally with 25 designated. This will be the inverter 10 or the electric machine 14 via a first number n_a of pulse-width modulation periods by means of a first sequence A and then for a number of n_b pulse-width modulation periods T by means of a second sequence B. The sum of the number n_a and the number n_b results in n_c. By virtue of the sequence A having a load value m1 and the sequence B having a load value m2, a load setpoint m can be set in a targeted manner by setting the numbers n_a and n_b, thereby providing a more uniform load on the switches of the inverter 10 can be achieved, where A and B are chosen so that m1 <m2.

To determine the number n_a and the number n_b, two different methods are preferably used.

wobei r der Verhältniswert, m der Belastungssollwert ist und m1, m2 die Belastungsreferenzwerte der beiden Sequenzen A, B sind.First, a ratio of the load values m, m1, m2 is determined by the formula

where r is the ratio value, m is the load setpoint and m1, m2 are the load reference values of the two sequences A, B.

In order to determine the number n_a, n_b, first the ratio value r is approximated by a rational number which corresponds to the quotient n_b / n_c of two integers. The number n_a is given by the equation n_a = n_c - n_b.

In order to avoid too long sequences of the periods n_c, the number of pulse-width modulation periods T can be limited within a total duration n_c, for example to 10 or 20 , Too long sequences n_c can disturb the dynamic regulation of the load setpoint m, whereby the switches S of the inverter 10 can be loaded unevenly.

The order of sequences A and B can be chosen freely. In order to distribute A and B homogeneously, the rounding function and from this a distribution of the two sequences A, B is determined. If n_a <= n_b, the number n_a is determined by a rounding function qa = floor (n_c / n_a), where qa is the number of pulse width modulation periods T, of which the inverter 10 is controlled in one of the pulse width modulation periods T by means of the first sequence A and for the remainder (qa-1) of the pulse-width modulation periods T by the second sequence B.

If, for n_a pulse width modulation periods T, the inverter 10 was controlled by the sequence A, the inverter is then driven exclusively by means of the sequence B to the inverter 10 in sum for n_c pulse width modulation periods T was driven by means of the sequence A and B, ie for n_a periods T by sequence A and n_c - n_a periods T by means of the sequence B.

If n_b <n_a, the number n_b is determined by the rounding function: qb = floor (n_c / n_b), where qb is the number of pulse width modulation periods T, of which the inverter 10 is driven in one of the pulse width modulation periods T by means of the sequence B, and wherein the sequence (A) is used for the remainder (qb-1) of the pulse-width modulation periods T.

If, for n_b pulse width modulation periods T, the inverter 10 was driven by the sequence B, then the inverter 10 exclusively controlled by the sequence A until the inverter 10 in sum for n_b pulse width modulation periods T was driven by the sequences A and B, ie for n_c periods by Sequence B and n_c - n_a periods T by means of the sequence A.

Alternatively, the distribution of sequences A, B may be determined by a binary sequence of numbers, as shown in FIG 5a and 5b is shown.

wobei n_a und p jeweils ganze Zahlen sind. Bei dem Beispiel aus 5a wird a mit 5 und n mit 3 angenommen. n_a wird mittels eines binären Baumes 30 zerlegt und es wird eine binäre Sequenz 32 bestimmt. Die ganze Zahl n_a wird in p Schritten in die binäre Sequenz 32 zerlegt. Dazu wird ein erster Wert durch Abrundung des Quotienten n_a/2 ermittelt und ein zweiter Wert durch die Differenz des ersten Wertes und des Ausgangswertes bestimmt. Der Abrundungswert wird in den Baum 30 links und der Restwert rechts vom Ausgangswert aufgetragen. Diese Berechnung wird auf die jeweiligen Ergebnisse erneut angewendet, bis die binäre Sequenz 32 erstellt ist. Similar to the method described above, a rational number is determined from the ratio r, so that

where n_a and p are each integers. In the example off 5a a is assumed to be 5 and n is 3. n_a is using a binary tree 30 decomposed and a binary sequence 32 is determined. The integer n_a is decomposed into binary sequence 32 in p steps. For this purpose, a first value is determined by rounding the quotient n_a / 2 and a second value is determined by the difference between the first value and the output value. The rounding value is in the tree 30 to the left and the residual value to the right of the initial value. This calculation is reapplied to the respective results until the binary sequence 32 is created.

Each digit of binary sequence 32 represents a pulse width modulation period T and determines whether sequence A or sequence B is being used. In this case, for example, a 0 in the binary sequence 32 corresponds to the sequence A and a 1 in the binary sequence 32 to the sequence B.

Alternatively, a second binary sequence 34 may be determined by the same method as described in US Pat 5b is shown. In this case, the binary tree is different 30 ' from the binary tree 30 out 5a in that the rounded result n_a / 2 is plotted on the right branch and the remainder on the left branch. This results in an alternative binary sequence 34.

QUOTES INCLUDE IN THE DESCRIPTION

This list of the documents listed by the applicant has been generated automatically and is included solely for the better information of the reader. The list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions.

Cited patent literature

• WO 2010/000548 A2 [0006]

Claims (11)

Procedure ( 25 ) for driving an inverter ( 10 ) by means of space vector modulation, in particular for driving an electrical machine ( 14 ), where the inverter ( 10 ) has a plurality of controllable switches (S) and which is designed to provide a polyphase electric current (IU, IV, IW), in particular around the electric machine ( 14 ) to supply polyphase with electric current (IU, IV, IW), whereby the inverter ( 10 ) is arranged to establish a plurality of successive different switching states (V0-V7) of the switches (S), and wherein the inverter ( 10 ) is switched by means of two switching states (V0, V7) in a voltage-free switching switching state, wherein the Wechslerichter ( 10 ) is controllable by means of at least two predefined sequences (A, B) of switching states (V0-V7) which have different switch-on times (t0-t7) of the switching states (V0-V7), characterized in that the inverter ( 10 ) is driven for a first number (n_a) of drive periods (T) by means of a first of the sequences (A) and is driven for a second number (n_b) of drive periods (T) by means of a second one of the sequences (B), the first one Number (n_a) and the second number (n_b) in dependence on a predetermined load setpoint (m) of the inverter ( 10 ) is determined. Method according to Claim 1, wherein the two predefined sequences (A, B) have different switch-on times (m1, m2) of the two voltage-free-switching switching states (V0, V7). The method of claim 1 or 2, wherein the two sequences (A, B) set an identical voltage space vector (V *). Method according to one of claims 1 to 3, wherein the first number (n_a) and the second number (n_b) is limited. Method according to one of Claims 1 to 4, wherein a ratio value (r) is formed on the basis of the load setpoint value (m) and reference load values (m1, m2) of the voltage-free-switching switching states (V0, V7) of the two sequences (A, B). Method according to one of claims 1 to 5, wherein the two sequences (A, B) are selected from a plurality of sequences whose load value (m1, m2) has the smallest difference from the load target value (m). The method of claim 6, wherein the first and second numbers (n_a, n_b) are determined based on the ratio (r). Method according to claim 6 or 7, wherein the first or the second number (n_a, n_b) is determined by rounding off a quotient (n_a / n_c) which is determined on the basis of the ratio value (r). Method according to one of claims 1 to 8, wherein on the basis of the ratio value (r) a binary number sequence ( 32 . 34 ) and according to the numerical values of the sequence of numbers ( 32 . 34 ) the inverter ( 10 ) is controlled by means of the first or the second predefined sequence (A, B). Contraption ( 18 ) for driving an inverter ( 10 ), in particular for driving an electrical machine ( 14 ), where the inverter ( 10 ) has a plurality of controllable switches (S) which are connected to provide a polyphase electric current (IU, IV, IW), in particular around the electric machine ( 14 ) to supply multiphase with electric current, with a control unit ( 18 ) which is adapted to the inverter ( 18 ) in such a way that the inverter ( 10 ) assumes a plurality of successive different switching states (V0-V7) of the switches (S), the inverter ( 10 ) can adopt two voltage-free-switching switching states (V0, V7), wherein the inverter ( 10 ) by means of at least two predefined sequences (A, B) of switching states which have different switch-on times (t0-t7) of the switching states, characterized in that the control unit ( 18 ) is adapted to the inverter ( 10 ) for a first number (n_a) of drive periods (T) by means of a first one of the sequences (A) and for a second number (n_b) of drive periods (T) by means of a second one of the sequences (B), the control unit ( 18 ) is adapted to the first number (n_a) and the second number (n_b) in response to a predetermined load reference value (m) of the inverter ( 10 ). Motor vehicle drive train with at least one electric machine ( 14 ) for providing drive power, an inverter ( 10 ) for driving the electric machine ( 14 ) and with a device ( 18 ) for driving the inverter ( 10 ) according to claim 10.

Images