Operating switching elements of an inverter

Abstract

Method for operating switching elements (22, 24, 26, 28, 30, 32, 34, 36, 38, 40) of an inverter (12) to which a stator winding of an at least three-phase electric machine (14) is connected, wherein the inverter (12) has at least one series connection of the switching elements (22, 24, 26, 28, 30, 32, 34, 36, 38, 40) for each of the phases (L1, L2, L3, L4, L5) of the electric machine (14) in order to electrically couple the electric machine (14) to a DC link (16) connected to the inverter (12), wherein switching signals for the switching elements are determined by vector control and provided at a clock rate greater than a frequency of a fundamental oscillation of the phase voltages in order to provide respective phase voltages by the inverter (12). is, wherein the switching signals are determined depending on at least one first reference pointer, which is determined bythat, for the switching states of the inverter (12) achievable by the switching elements (22, 24, 26, 28, 30, 32, 34, 36, 38, 40), first voltage state vectors are determined in a corresponding first space vector system using a Fortescue transformation adapted to the number of phases of the electrical machine (14), wherein the first reference vector is determined depending on a superposition of at least two of the first voltage state vectors, wherein, for determining the at least one first reference vector, at least two first voltage state vectors are chosen, one of which is a long first voltage state vector and the other a medium or a short first voltage state vector, and, for superimposing a harmonic of the phase voltage with respect to the frequency of the fundamental oscillation, the switching signals are additionally determined depending on at least one second reference vector.which is coupled to the first reference vector depending on an order number of the harmonics and which is determined by determining second stress state vectors in a corresponding second space vector system, wherein the second reference vector is determined depending on a superposition of at least two of the second stress state vectors and wherein a separate modulation level is determined for the harmonic and a total modulation level is maximized depending on the separate modulation level.

Description

[0001] The invention relates to a method for operating switching elements of an inverter to which a stator winding of an at least three-phase electric machine is connected, wherein the inverter has at least one series connection of the switching elements for each of the phases of the electric machine in order to electrically couple the electric machine to a DC link connected to the inverter, wherein switching signals for the switching elements are determined by the inverter using vector control and provided at a clock rate that is greater than a frequency of a fundamental oscillation of the phase voltages, wherein the switching signals are determined depending on at least one reference phasor, which is determined bythat, for switching states of the inverter achievable by the switching elements, first voltage state vectors are determined in a corresponding first space vector system using a Fortescue transformation adapted to the number of phases of the electric machine, wherein the first reference vector is determined depending on a superposition of at least two of the first voltage state vectors. The invention further relates to a control unit for operating switching elements of an inverter to which a stator winding, preferably configured in a star connection, of an at least three-phase electric machine is connected, wherein the inverter has at least one series connection of the switching elements for each of the phases of the electric machine in order to electrically couple the electric machine to a DC link connected to the inverter, wherein the control unit is configuredThe invention relates to the determination of switching signals for the switching elements by the inverter using vector control and to the provision of switching signals for the switching elements at a clock rate greater than the frequency of a fundamental oscillation of the phase voltages, wherein the control unit is configured to determine the switching signals depending on at least one first reference vector by determining first voltage state vectors in a corresponding first space vector system for switching states of the inverter achievable by the switching elements using a Fortescue transform adapted to the number of phases of the electric machine, wherein the control unit is configured to determine the first reference vector depending on a superposition of at least two of the first voltage state vectors. Finally, the invention also relates to a drive system with an electric machine with at least three phases.which has a stator winding preferably configured in a star connection, an inverter to which the stator winding is connected, wherein the inverter has at least one series connection of switching elements for each of the phases of the electrical machine in order to electrically couple the electrical machine to a DC link connected to the inverter, and a control unit for operating the switching elements of the inverter in order to provide respective phase voltages through the inverter.

[0002] Generic methods, generic control units, and corresponding drive systems are extensively known in the prior art, so no separate printed documentation is required. These are used in the prior art for a multitude of applications where a specific drive function is to be realized by means of a multiphase electric machine. The use of three-phase electric machines is particularly widespread; depending on the application, these can even be directly connected to a three-phase public power supply network to realize a predefined drive function. However, the control options in such a scenario are limited, which is why multiphase electric machines are now operated via individual inverters, usually assigned separately.This also allows the electrical energy required for the intended operation of the electric machine to be supplied from an electrical energy storage device, such as a battery, especially a high-voltage battery, but also a fuel cell, a wind generator, a photovoltaic module, and / or the like. Such a configuration makes it possible to provide a drive system that is not only suitable for stationary applications but also for mobile use, for example, in vehicles, especially motor vehicles. A multitude of other applications can also be realized in this way.

[0003] The use of an inverter makes it possible to control the multi-phase electric machine flexibly, even when the electric machine has more than three phases, for example, four, five, six, or even more. Although three-phase machines are currently the most common application, there are also specific applications where the electric machine has more than three phases, particularly five or six phases, as can be found, for example, in marine propulsion systems.

[0004] The use of an inverter to operate a multiphase electrical machine is now also widespread. The use of inverters opens up the possibility of operating the multiphase electrical machine in a very dynamic manner. For this purpose, vector control, sometimes also implemented as vector regulation, is frequently employed, which, in the three-phase case, is based on the use of a d / q transformation, also called the Park transformation. The Park transformation serves to transform three-phase quantities of the three-phase electrical machine into a two-axis coordinate system with reference axes d and q. This allows vector control to be implemented by using space vector representations to adjust the operating states of the inverter.The d / q transformation is related to the Clarke transformation and differs from the latter in particular in that the d / q coordinate system of the d / q transformation rotates with a rotor of the three-phase electric machine in a steady-state operating case, whereby in this case a respective pair of values ​​d, q represents a quantity that is constant over time for this operating state.

[0005] However, the aforementioned vector control cannot be directly used for electrical machines with more than three phases. For a five-phase electrical machine, a system and a device for optimizing an injected current of a third harmonic are disclosed in US 2011 / 0 221 365 A1.

[0006] The publication "Space Vector Modulation Schemes for a Five-Phase Voltage Source Inverter" by Igbal, A., Levi, E., European Conference Power Electronics and Applications, 2005, Dresden, pages 1-12, describes pulse-width modulation (SVPWM) schemes for five-phase inverters. The scheme involves combining two SVPWM modulations using vectors of different lengths (long and medium length). These modulations are applied sequentially within a single PWM period. This results in additional harmonics, which, according to the publication, are of low amplitude.

[0007] The publication “General Analysis of Multi-Phase Systems Based on Space Vector Approach” by Grandi, G., Serra, G. and Tani, A., 12th International Power Electronics and Motion Control Conference in Portoroz, 2006, pages 834-840, describes a space vector approach for multiphase systems. The Fortesque transform is extended here so that several space vectors arise in different d,q planes.

[0008] The publication "Study of Five-phase Space Vector PWM Considering Third Order Harmonics" by Jiang, D. and Qian, W., September 15-19, 2013, Denver, CO, USA, 2013 IEEE Energy Conversion Congress and Exposition, pages 4227-4232, describes the use of space vector modulation for five-phase electrical machines, which considers both the fundamental frequency and its third harmonic with regard to back EMF. When using SVPWM to generate the fundamental frequency and the third harmonic, the distribution of two different zero vectors results in a relationship between SVPWM and a carrier-based PWM.

[0009] Even with these approaches, disadvantages remain regarding performance, especially concerning high voltage utilization.

[0010] The invention is therefore based on the objective of improving the voltage utilization of a vector control for an electrical machine with at least three phases, in particular for an electrical machine with more than three phases.

[0011] The invention proposes a method, a control unit and a drive system according to the independent claims as a solution.

[0012] Advantageous further training opportunities arise from the characteristics of the dependent requirements.

[0013] With regard to a generic method, it is particularly proposed that, in order to determine the at least one first space vector, at least two first stress state vectors are chosen, one of which is a long first stress state vector and the other a medium or a short first stress state vector.

[0014] With regard to a generic control unit, it is specifically proposed that the control unit be configured to select two first stress state vectors, one of which is a long first stress state vector and the other a medium or short first stress state vector. It is proposed that a long first stress state vector be selected that is longer than the medium or short first stress state vector. In general, two first stress state vectors of different lengths are used.

[0015] With regard to a drive system of the generic type, it is particularly proposed that the drive system comprises a control unit according to the invention.

[0016] The invention is based on the idea that by choosing the first voltage state vectors more freely when determining the reference point, the modulation level can be increased to such an extent that voltage utilization with respect to the phase voltages can be significantly improved.

[0017] The phase voltages are alternating voltages at the respective phase terminals of the inverter or the electric machine, and specifically refer to the voltages at the respective phase terminals of the inverter to which the stator winding of the electric machine is connected. For each phase of the at least three-phase electric machine, the inverter provides at least one phase terminal to which the respective phase of the electric machine is connected. For each phase terminal, the inverter provides at least one series circuit of switching elements, which is generally connected to the DC link to which the inverter is also connected. The respective center taps of the series circuits provide the phase terminals of the inverter.

[0018] Typically, the phase currents of the stator winding are set by specific PWM patterns derived from switching signals provided by the control unit, which control the respective switching elements of the series circuits. The phase voltages can then be determined from these PWM patterns by, for example, filtering the respective PWM pattern using a suitable low-pass filter, calculating a Fourier transform, measuring the respective phase voltage with a power meter, and / or similar methods. The phase voltage therefore does not correspond directly to the PWM pattern provided at the respective phase terminal of the inverter. For this reason, the frequency of the fundamental oscillation of each phase voltage is generally lower than the clock rate of the switching signals used to provide the respective PWM patterns at the respective phase terminals.

[0019] An inverter is a type of electrical energy converter that couples the DC link to an AC network, allowing electrical energy to be exchanged between them. The AC network is defined by its phase voltages. For this purpose, at least one series circuit consisting of at least two switching elements is typically provided for each phase of the AC network, electrically coupled to the DC link. The respective phase of the AC network can be connected to the center taps of these series circuits.

[0020] In the case of coupling the DC link via the inverter to the at least three-phase electric machine, the respective phase terminals of the electric machine, referred to simply as phases, are connected to the respective center taps of the series circuits that provide the phase terminals of the inverter. By appropriately controlling the switching elements using the switching signals provided by the control unit, the desired conversion function of the inverter can be achieved, thus enabling the desired drive function of the electric machine.

[0021] For this purpose, the switching elements are usually operated with switching signals that have a predetermined clock rate significantly higher than the fundamental frequency of the phase voltages or phase alternating currents of the electric machine. The desired energy coupling can then be achieved using specific control methods such as pulse width modulation (PWM) or similar techniques. To this end, the control unit provides the respective specific switching signals for preferably each of the inverter's switching elements, enabling the switching elements to be operated as desired during switching operation. The switching signals can be provided individually for each switching element.

[0022] By using suitable pulse patterns of the switching signals for the switching elements provided by the control unit, a predetermined energy exchange can be achieved.

[0023] A switching element within the meaning of this disclosure is, in particular, a controllable electronic switching element, for example, a controllable electronic semiconductor switch such as a transistor operated in switching mode, a thyristor, combinations thereof, preferably with inverse diodes connected in parallel, a gate turn-off thyristor (GTO), an insulated-gate bipolar transistor (IGBT), combinations thereof, or the like. However, the switching element can also be formed by a field-effect transistor, in particular a metal oxide semiconductor field-effect transistor (MOSFET).

[0024] To provide the desired energy conversion functionality, the inverter's switching elements operate in switching mode. In the context of a semiconductor switch using a transistor, switching mode means that in the on state, a very low electrical resistance is present between the transistor's terminals forming the switching path, allowing for a high current flow at a very low residual voltage. In the off state, however, the transistor's switching path has a high resistance, meaning it presents a high electrical resistance. Therefore, even with a high voltage applied to the switching path, there is essentially no current flow, or only a very small, often negligible, current flow. This differs from linear operation in transistors, which is generally not used in inverters of this type.

[0025] To implement the control function, each switching element has at least one control terminal where it can be supplied with switching signals provided by the control unit, thus enabling the desired switching function of the switching element. The switching signal can be a binary signal, capable of assuming two state values ​​to achieve the desired switching functions of the switching element. For example, the switching element can be formed by a pulse train, which supplies the control terminal. This is particularly useful for thyristors, such as GTOs or similar devices. Furthermore, for transistors, the switching signal can be a square wave, where each switching state of the switching element can be assigned to one of the potentials of the square wave.Such a signal is useful, for example, for transistors, especially bipolar transistors, field-effect transistors, or the like.

[0026] The control unit provides the function for generating the switching signals. It can also implement other functions, particularly those related to the inverter, such as monitoring, safety, and / or similar functions. For this purpose, the control unit can comprise a hardware circuit and / or a program-controlled computer unit, or similar components. Naturally, the control unit can be a separate module. However, it can also be at least partially integrated into a higher-level control system for the drive system.

[0027] The inverter is an electronic hardware circuit comprising series circuits, preferably consisting of two switching elements each, corresponding to the number of phases of the at least three-phase electric machine. These series circuits are typically connected in parallel to the at least one DC link. However, alternative configurations may also allow some of the series circuits to be connected to separate DC links. The respective center taps of the series circuits provide the phase connections to which the respective phases of the at least three-phase electric machine can be connected. The desired power coupling between the DC link and the at least three-phase electric machine can be achieved through suitable pulse patterns of the switching signals.The basic operating procedures in this regard are known, therefore no further details will be provided here. For intended operation, the switching signal rate is generally considerably higher than the frequency of the fundamental oscillations of the phase voltages or the phase-alternating currents of the at least three-phase electrical machine.

[0028] The at least three-phase electric machine can be designed as a synchronous machine, an asynchronous machine, a doubly fed asynchronous machine, or the like. Combinations of these are also possible, for example, to realize specific applications that require influencing the magnetic flux.

[0029] The DC link can be coupled to an electrical power source, which may be an electrical energy storage device such as a battery or similar. Furthermore, it is of course possible for the DC link to be electrically coupled to a public power supply network, which is configured as an AC network, via a rectifier unit. Combinations of these configurations are also possible.

[0030] To represent the phase voltages that occur, the control unit uses a space vector system when processing the corresponding data, enabling the representation of the phase voltages of the at least three-phase system. The Fortescue transform serves this purpose. Using the Fortescue transform, the phase components of the multi-phase system can be represented by corresponding transformed components and a zero component. Since the zero components are usually zero, they can be neglected in further analysis. Further simplifications can be achieved by exploiting symmetries. In the Fortescue transform, the dimension of the matrix is ​​chosen depending on the number of phases. For example, in a five-phase system, the dimension can be reduced from five to four by utilizing symmetry properties.To visualize the corresponding space vectors, two subspaces are used, with the first subsystem being assigned to the fundamental oscillation. Initial stress state vectors are defined in this subspace.

[0031] The Fortescue transform determines the first stress state vectors in the corresponding first space vector system or subspace, whereby immediately adjacent first stress state vectors can form sectors in the space vector system. The space vector system is mapped in the complex plane. According to the invention, at least two stress state vectors are used to determine the first reference vector, one of which is a long first stress state vector. The other of the at least two first stress state vectors is a medium or a short first stress state vector. Preferably, exactly one long first stress state vector and / or exactly one medium or a short first stress state vector is provided.For example, in the five-phase case with three voltage levels, an increased number of voltage state vectors result, such as long, medium, and short first voltage state vectors. These can be used to determine the first reference vector. By combining a long with a medium or short first voltage state vector, a particularly favorable first reference vector can be determined. This makes it possible to adjust the modulation level much more finely and precisely, so that an overall higher modulation level compared to the prior art can be achieved, thereby significantly improving voltage utilization. Overall, the invention allows for further improvement of the performance of the drive system with at least a three-phase electric machine.

[0032] It is further proposed that immediately adjacent first stress state vectors form sectors in the space vector system, wherein the first reference vector is determined depending on a superposition of at least two of the first stress state vectors, and wherein, to determine the at least one first reference vector, at least two of the first stress state vectors are selected that are spaced apart from each other by at least two adjacent sectors. Preferably, no first stress state vectors, particularly immediately adjacent ones, are used that are only spaced apart from each other by one sector.

[0033] In particular, at least two sectors are positioned between at least two consecutive first stress state vectors. If more than two consecutive first stress state vectors are used, at least two of the first stress state vectors are separated from each other by at least two adjacent sectors. To determine the first reference vector in this case, instead of using at least two first stress state vectors that are directly separated by a sector, first stress state vectors are used where at least two are directly separated from each other by at least two adjacent sectors. This allows for a further improvement in the modulation level.

[0034] It is further proposed that, to superimpose a harmonic of the phase voltage with respect to the frequency of the fundamental oscillation, particularly an odd harmonic, the switching signals are additionally determined depending on at least one second reference phasor. This second reference phasor is coupled to the first reference phasor depending on an order number of the harmonic and is determined by calculating second voltage state vectors in a corresponding second space vector system. The second reference phasor is determined depending on a superposition of at least two of the second voltage state vectors. The Fortescue transform can also be used to determine the voltage state vectors. This allows for a further improvement, in addition to the improvements already discussed according to the invention, by additionally taking into account a harmonic, particularly an odd harmonic.This makes it possible to further improve voltage utilization. For this purpose, the second reference phasor can be coupled to the first reference phasor via the order number of the harmonics. This could mean, for example, that a rotating second reference phasor rotates at an angular velocity increased by the order number relative to the first reference phasor. In this way, it can be easily achieved not only that harmonics are superimposed to improve voltage utilization, but also that a correspondingly high modulation level can be realized simultaneously.

[0035] Preferably, the electrical potential of the star connection can change depending on the harmonics via the action of the stator winding. For this purpose, the neutral point is preferably not electrically coupled to any other electrical potential; that is, it is preferably insulated.

[0036] A further development provides that, in an electrical machine with at least four phases, at least one odd harmonic is used as the harmonic, the ordinal number of which is less than the number of phases of the electrical machine. In an electrical machine with at least six phases, for example, a third and / or a fifth harmonic can be used. It has been shown that in this case, the third and / or the fifth harmonic are particularly suitable for achieving a further improvement in voltage utilization, especially in electrical machines with more than five phases, particularly five or more. It can be provided that either only the third or only the fifth harmonic is used. However, it is also possible to provide that both harmonics are used simultaneously.Depending on the application and requirements, this can be adapted accordingly, in particular, it can be adjustable. However, this training is not limited to this and can of course also be used with higher-phase electrical machines, adapted accordingly.

[0037] It is proposed that a separate modulation level be determined for the harmonic and that an overall modulation level be maximized depending on this separate modulation level. To this end, it can further be provided that a modulation level is determined with respect to the frequency of the fundamental frequency, and that this modulation level with respect to the frequency of the fundamental frequency is then combined with the separate modulation level to form an overall modulation level. This can be achieved, for example, by determining the modulation levels using an approximation based on a Taylor series expansion or similar method. For instance, a maximization function can then be used to maximize the overall modulation level, thereby also maximizing the voltage utilization. Overall, this can lead to a further improvement.

[0038] Furthermore, it is proposed that at least one of the at least two first voltage state vectors be a medium or a short voltage state vector. This allows – in contrast to the prior art – further improvement in the flexibility regarding the determination of the first and / or the second reference vector. In particular, this can result in more favorable adjustment options, which can also be advantageous for the operation of the inverter and its switching elements.

[0039] Preferably, it is proposed that at least one of the phases be operated in unswitched mode for at least one switching period. Likewise, the switching period can, of course, also correspond to the period of one of the harmonics. This can reduce switching losses in the inverter as well as interference.

[0040] Furthermore, it is proposed that the sequence for utilizing the voltage state vectors be chosen to minimize the number of switching state changes of the switching elements. This refinement takes into account that several options exist for implementing a specific sequence for utilizing the voltage state vectors, from which the one resulting in the fewest possible switching state changes of the switching elements can be selected. This can further improve the functionality of the inverter because reducing the number of switching state changes not only reduces switching losses but also minimizes interference, particularly with regard to electromagnetic compatibility. In addition, reliability and interference immunity can, of course, also be improved.

[0041] The advantages and effects described for the method according to the invention naturally apply equally to the control unit according to the invention and to the drive system equipped with the control unit according to the invention, and vice versa. Consequently, method features can also be formulated as device features and vice versa.

[0042] The invention also includes further developments of the control unit and the drive system according to the invention, which have features already described in connection with the further developments of the method according to the invention. For this reason, the corresponding further developments of the method according to the invention are not described again here.

[0043] The invention also includes combinations of the features of the described embodiments.

[0044] The following are exemplary embodiments of the invention described. This is illustrated by: Fig. 1 in a schematic block diagram representation a drive system connected to a DC intermediate circuit; Fig. 2. A schematic circuit diagram showing the structure of an inverter designed for three-phase operation; Fig. 3 in a schematic circuit diagram a stator winding of a three-phase electric machine in a star point connection; Fig. 4 possible switching states of the inverter in a schematic circuit diagram according to Fig. 2; Fig. 5 in a schematic circuit diagram representation such as Fig. 2. A design for an inverter that is designed for five-phase operation; Fig. 6 in a schematic circuit diagram representation as Fig. 4 possible switching states of an inverter according to Fig. 5; Fig. 7 in a schematic circuit diagram a stator winding of a five-phase electric machine in a star point connection; Fig. 8 in a schematic diagram representation a three-phase space vector system for vector control of the inverter according to Fig. 2; Fig. 9 in a schematic diagram representation such as Fig. 10 a five-phase space vector system for vector control of the inverter according to Fig. 5, where a left sub-diagram is assigned to a fundamental frequency and a right sub-diagram to a third harmonic; Fig. 10 in a schematic representation phase sequences of the fundamental frequency and the third harmonic according to Fig. 9; Fig. 11 in a schematic diagram representation a three-phase space vector modulation in a geometric hexagon; Fig. 12 in a schematic diagram representation such as Fig. 9 a five-phase space vector modulation in a geometric decagon; Fig. 13 in a schematic diagram representation, determining a space vector in three-phase space vector modulation according to Fig. 11; Fig. 14 in a schematic diagram representation as Fig. 9. Determining a space vector in five-phase space vector modulation according to a 2L method, wherein a left sub-diagram is assigned to a fundamental frequency and a right sub-diagram to the third harmonic; Fig. 15 switching signals normalized in a schematic signal representation for the five phases according to vector control according to Fig. 14; Fig. 16 in a schematic diagram representation as Fig. 14. Determining a space vector according to a 2L2M method; Fig. 17 in a schematic signal representation as Fig. 15 standardized switching signals for the five phases according to vector control. Fig. 16; Fig. 18 in a schematic diagram representation a space vector system such as Fig. 14 for determining a switching sequence of switching state changes of the switching elements of the inverter according to Fig. 5; Fig. 19 in a schematic signal representation as Fig. 15 standardized switching signals for the five phases according to vector control. Fig. 16 according to a CPWM procedure; Fig. 20 in a schematic signal representation such as Fig. 19 standardized switching signals for the five phases according to vector control according to Fig. 16 according to a DPWMmin procedure; Fig. 21 in a schematic three-dimensional diagram representation a signal waveform for an oscillation at the fundamental frequency, to which a fifth harmonic is superimposed, depending on a phase difference between the oscillations; Fig. 22 in a schematic diagram representation a signal waveform for a normalized phase voltage; and Fig. 23 in a schematic diagram representation such as Fig. 14 Reference phasors for the fundamental frequency are shown in an upper diagram and reference phasors for the third harmonic are shown in a lower diagram, achieving maximum stress utilization by using stress state vectors that are spaced apart by more than one sector.

[0045] The exemplary embodiments described below are preferred embodiments of the invention. In these exemplary embodiments, the described components each represent individual features of the invention that can be considered independently of one another. Each of these features further develops the invention independently and can therefore be considered part of the invention individually or in a combination other than that shown. Furthermore, the described embodiments can also be supplemented by other features of the invention already described.

[0046] In the figures, functionally identical elements are each provided with the same reference symbols.

[0047] Fig. Figure 1 shows a schematic block diagram of a drive system 10, which comprises a multiphase, permanent magnet synchronous machine 14 as an at least three-phase electric machine with a stator winding 52 connected in a star configuration. The drive system 10 also includes an inverter 12 to which the stator winding 52 is connected. The inverter 12 has at least one series connection of switching elements 22, 24, 26, 28, 30, 32, 34, 36, 38, 40 for each of the phases L1, L2, L3, ... in order to electrically couple the synchronous machine 14 to a DC link 16 connected to the inverter 12. A high-voltage battery 20 is also connected to the DC link 16 as an electrical energy storage device, providing the electrical energy for the intended operation of the drive system 10.In the present embodiment, the drive system 10 is part of a drive unit of a motor vehicle (not shown) by means of which the motor vehicle can be electrically driven. In the present embodiment, the motor vehicle is designed as an electric vehicle. However, it can also be designed as a hybrid vehicle or the like.

[0048] For the intended operation of the drive system 10, in particular the inverter 12, the drive system 10 comprises a control unit 18 for operating the switching elements 22, 24, 26, 28, 30, 32, 34, 36, 38, 40 of the inverter 12, in order to provide the respective phase voltages for the phases L1, L2, L3, ... of the synchronous machine 14 by means of the inverter 12.

[0049] Fig. Figure 2 shows a simplified schematic circuit diagram of the inverter 12 for three-phase operation. For this purpose, the inverter 12 has a series connection for each of the three phases L1, L2, L3, with corresponding switching elements 22, 24, 26, 28, and 30, 32, which are connected in parallel to the DC link 16. The respective center taps 42, 44, 46 of the series connections provide corresponding phase terminals of the inverter 12, to which the respective phases L1, L2, L3 of the synchronous machine 14 can be connected. The series connections according to Fig. These are often also called half-bridges or half-bridge circuits.

[0050] In this embodiment, the synchronous machine is designed as a three-phase synchronous machine 14, whose stator winding 52 is in Fig. Figure 3 is shown in a schematic circuit diagram. It can be seen that the stator winding 52 has three windings, each connected at one of its terminals to the respective phases L1, L2, and L3. The opposite terminal of each winding is connected to a neutral point N, thus forming a star connection for the stator winding 52. The neutral point N is isolated from other electrical potentials.

[0051] Fig. Figure 4 shows in a schematic circuit diagram the eight possible different switching states that are achieved by switching elements 22, 24, 26, 28, 30, 32 of the inverter according to Fig. 2 can be taken.

[0052] Out of Fig. Figure 2 shows that the switching elements used are IGBTs with an integrated inverse diode. These IGBTs have an unlabeled gate terminal as their control connection, which is connected to the control unit 18. This allows the control unit 18 to individually control the switching elements 22, 24, 26, 28, 30, and 32 with respect to their respective switching states. For this purpose, the IGBTs are operated in switching mode by means of a suitable individual switching signal. By varying the switching signals or their pulse patterns using suitable modulation techniques, a square wave voltage waveform, for example according to pulse-width modulation or the like, can be provided at the center taps 42, 44, and 46.Since the corresponding phase windings of the synchronous machine 14 are connected to the phase terminals of the inverter 12, corresponding current waveforms of alternating phase currents are generated in the phase windings of the stator winding 52, resulting from the voltages provided by pulse width modulation. In this way, it is possible to operate the synchronous machine 14 with high flexibility in a wide variety of operating modes using the inverter 12.

[0053] In the present symmetrical three-phase case, the windings of the stator winding 52 are energized such that the phase angles between the respective phase alternating currents are approximately 120°. In the present synchronous machine 14, it is provided that each of the phases L1, L2, L3 exhibits essentially the same electrical and magnetic properties. This also applies to the following descriptions.

[0054] Fig. 5 shows in a schematic circuit diagram how Fig. 2 an inverter 12, which in this case is designed for five-phase operation, so that a corresponding five-phase electrical machine, which in this case is also designed as a permanent magnet synchronous machine 14, can be connected to it. Accordingly, the inverter 12 has, according to Fig. 5 Five series circuits of switching elements 22, 24, 26, 28, 30, 32, 34, 36, 38, 40 are connected in parallel to the DC link 16. Phase connections of the inverter 12 are provided at center taps 42, 44, 46, 48, 50 so that phases L1, L2, L3, L4, L5 of a five-phase synchronous machine 14, which has a stator winding 52 according to Fig. 7, can be connected. From Fig. Figure 7 also shows that the stator winding 52 is configured in a star connection.

[0055] The individual elements of the inverter 12 and the synchronous machine 14 essentially correspond to those described in the preceding embodiment, and reference is therefore made to the corresponding preceding explanations. Here too, the individual windings of the stator winding 52 are arranged symmetrically offset from one another.

[0056] Fig. Figure 6 shows an example accordingly Fig. 4 four of 32 possible achievable switching states of the inverter 12 according to Fig. 5.

[0057] It has been shown that, compared to three-phase electric machines, it is advantageous if the five-phase electric machine can be subjected to a trapezoidal magnetic flux in the air gap during normal operation. This can result in the following advantages: - Harmonic overtones, especially the third harmonic, contribute positively to torque generation. - A higher degree of modulation can be effectively controlled. - Better iron utilization can be achieved because all teeth, and not just one tooth per phase, are saturated. - Increased power density can be achieved. - A higher flow density can be achieved. - A higher torque density can be achieved.

[0058] Overall, it proves to be particularly advantageous to explicitly use a third harmonic for the operation of the synchronous machine 14 according to Fig. 7 to provide in order to approximate or improve trapezoidal magnetic flux patterns in the air gap.

[0059] To represent phase voltages in multiphase drive systems, a space vector representation can be used. Furthermore, other phase quantities can also be represented, which can be derived from phase components, for example, using a Clarke transform. First, however, a general transformation for N-phase systems will be presented and then applied to the three-phase and five-phase cases.

[0060] A Fortescue transformation can serve this purpose, enabling the simple representation and calculation of fault cases in multiphase systems, as well as the visualization of symmetrical drive systems using space vectors. The Fortescue transformation can be represented by the following equations: x=Tn⋅xph [ x0xα1 / β1xα2 / β2⋮xαn−1 / βn−1]=2n[αn0 αn0 αn0αn0⋯αn0αn0 αn1 αn2αn2⋯αn1⋅(n−1)αn0 αn2 αn4αn6⋯αn2⋅(n−1)αn0 αn3 αn6αn9⋯αn3⋅(n−1)⋮ ⋮ ⋮ ⋮⋱ ⋮αn0αn1⋅(n−1)αn2⋅(n−1)αn3⋅(n−1)⋯αn(n−1)⋅(n−1)]⋅[xph,L1xph,L2xph,L3 ⋮xph,Ln] αn=ejφn=ej2πn=cos(2πn)+jsin(2πn)

[0061] From the aforementioned equations, it can be seen that n phase components x can be transformed using this method. ph ,L v by n-1 components x αv / βv and represent a component x0. x can be any component, for example a current, a voltage, or the like. Since the zero components result in zero due to the star connection, they can be neglected in the following. α n is called the rotation coefficient and corresponds to a rotation by multiples of an angle. φn=2πn rad, which determines an offset of the phase voltages.

[0062] Applying equation 1.3 to the three-phase case (n = 3) yields the following equation: [xα1 / β1xα1 / β2]=[α30α31α32α30α32α31]⋅[xph,L1xph,L2xph,L3]

[0063] Equation 1.5 shows that a system with x α1 / β1 with ascending exponents (0 → 1 → 2) and a complementary system with descending exponents (2 → 1 → 0). In a balanced, symmetrical system, as is the case here, the components of both systems are complex conjugates of each other and thus represent the same information. Due to the resulting redundancy, it suffices to consider only one system to represent all the information. In a three-phase system, the complementary system is usually selected, resulting in equation 1.7: xα / β=23[α30 α31 α32]⋅[xph,L1xph,L2xph,L3]

[0064] If we now substitute in equation 1.7 according to equation 1.4 and split α3 into two separate components according to the real and imaginary parts, as shown in equation 1.6, the Clarke transformation is obtained according to the following equation: xα / β=[xαxβ]=23[cos(0φ3)cos(1φ3)cos(2φ3)sin(0φ3)sin(1φ3)sin(2φ3)]⋅[xph,L1xph,L2xph,L3] =23[1−12−12032−32]⋅[xph,L1xph,L2xph,L3] xα / β=xα+jxβ

[0065] The Clarke transform is thus a special case of the Fortescue transform for the three-phase case. The Clarke transform is used to determine the eight possible combinations of phase voltages for the case according to... Fig. 4. To transform by replacing x with u in the corresponding equations, the following results for the eight possible switching states according to Fig. 4 eight corresponding stress state vectors. These can be represented in a complex α / β plane, which is defined by u α / β= u α + ju β is defined. This is a stator-fixed coordinate system, which is defined such that, if one considers the star point N of the winding axes of the synchronous machine 12 according to Fig. 3, at whose origin the stress state vector 1 points in the direction of phase L1. This is shown in the schematic diagram according to Fig. Figure 8 shows the abscissa being assigned to the quantity α and the ordinate to the quantity β. The Arabic numerals in the diagram are as follows: Fig. 8 corresponds to the value of the binary represented switching matrices according to Fig. 4, where the switching state of the highest phase number, here phase L3, corresponds to the most significant bit.

[0066] This results in six active voltage state vectors, 1 to 6, and two zero-state vectors, namely 0 and 7. The zero-state vectors 0 and 7 arise because either all switching elements 22, 24, 30 or all switching elements 24, 28, 32 are in the switched-on state. The other switching elements 22, 24, 26, 28, 30, 32 are in the switched-off state. In these two switching states, the windings of the stator winding 52 are short-circuited by the inverter 12, so the resulting phase voltages are zero. The active voltage state vectors therefore result from combinations of different switching states of the switching elements 22, 24, 26, 28, 30, 32. These are represented in the two-dimensional plane according to Fig. 8 can be represented and together span six sectors I - VI.

[0067] If n=5 is substituted into formula 1.3, a corresponding transformation for the five-phase case results, as shown below: [uα1 / β1uα2 / β2uα3 / β3uα4 / β4]=25[α50α51α52α53α54α50α52α54α56α58α50 α53α56α59α512α50α54α58α512α516]⋅[uph,L1uph,L2uph,L3uph,L4uph,L5]

[0068] The matrix will also be referred to as T5 in the following. It results in four subsystems. Table 1 shows how these subsystems are derived from the rows of the transformation matrix of formula 1.9. Zeile in T5 Subsystem Exponentenfolge 1 Mitsystem 1 0→1→2→3→4 2 Gegensystem 2 0→-3→-6→-9→-12 3 Mitsystem 2 0→3→6→9→12 4 Gegensystem 1 0→-1→-2→-3→-4

[0069] When creating Table 1, a periodicity is taken into account. a5k=a5k+5 This is used for exponent conversion. This results in two corresponding positive and negative systems, which—as in the three-phase case—contain the same information for symmetric systems. Therefore, it is sufficient to consider one of the positive systems and one of the negative systems to represent the complete information. In this case, positive system 1 and positive system 2 are used for this purpose. This results in the following equation: [uα1 / β1uα2 / β2]=25[α50α51α52α53α54α50α53α51α54α52]⋅[uph,L1uph,L2uph,L3uph,L4uph,L5]

[0070] When the real and imaginary parts of the respective complex components are represented separately, the following equation results: [uα1uβ2uα3uβ3]=25[cos(0φ5)cos(1φ5)cos(2φ5)cos(3φ5)cos(4φ5)sin(0φ5)sin(1φ5)sin(2φ5)sin(3φ5)sin(4φ5)cos(0φ5)c os(3φ5)cos(6φ5)cos(9φ5)cos(12φ5)sin(0φ5)sin(3φ5)sin(6φ5)sin(9φ5)sin(12φ5)]⋅[uph,L1uph,L2uph,L3uph,L4uph,L5]

[0071] It should be noted that the transformation matrix of the five-phase case is not an extension of the Clarke transformation but rather another special case of the Fortescue transformation.

[0072] Here too, the stress state vectors of the phase stresses can be visualized using space vectors. This results in a four-dimensional vector, which would have to be represented as a hypercomplex number in a four-dimensional coordinate system: uα1 / β1 / α3 / β3=1uα1+j1uβ1+j2uα3+j3uβ3

[0073] For the representation, we therefore choose a representation in two two-dimensional subspaces that are linked together. Here, the first cosystem is represented in the first subspace or first subsystem, and the second cosystem is represented in the second subspace or second subsystem using complex quantities ( Fig. 9). In Fig. In 9, the left diagram is assigned to the first subspace and the right diagram to the second subspace.

[0074] The same reference symbols for stress state vectors 0 to 31 in the two subsystems correspond to identical stress state vectors. The overall system therefore exhibits, among other things, the following properties: - For the 32 possible switching states, 32 space vectors result as voltage state vectors, namely 30 active voltage state vectors 1 to 30 and two zero state vectors 0 and 31. - The space pointers each span 10 sectors I to X. - The two subsystems are orthogonal to each other and together form a four-dimensional space. - The first subsystem represents a fundamental oscillation. - A third harmonic is mapped in the second subsystem. Three types of stress state vectors allow for three stress levels. These are abbreviated as U. small, medium U medium and long U large designated. - The potentials are determined solely by the DC link voltage 16 U dc defined and stand in fixed relationships to each other. Long space vectors in the first subsystem correspond to short space vectors in the second subsystem. Medium space vectors are assigned to corresponding medium space vectors. Usmall=45⋅cos(2π5)⋅Udc=0.247⋅Udc Umedium=25⋅Udc=0.400⋅Udc Ularge=45⋅cos(π5)⋅Udc=0.647⋅Udc Usmall:Umedium:Ularge=1:1.618:1.6182

[0075] The first subsystem is oriented such that when the neutral point N of the stator winding 52 is placed at the origin, the abscissa α1 is oriented in the direction of phase L1. Since the third harmonic is represented in the second subsystem, the phase sequence differs here. The resulting sequence is: L1→L3→L5→L2→L4. This is in Fig. Figure 10 is shown. The left representation is assigned to the fundamental frequency and the right representation to the third harmonic.

[0076] One way to visualize the relationship between the two oscillations is through Fourier series expansion, specifically with respect to a square wave signal. If the series expansion is truncated after the second term, a function consisting of a fundamental oscillation and a third harmonic can be represented. With so few terms, a signal approximating a trapezoidal shape can be achieved, which closely approximates the trapezoidal magnetic flux in the air gap desired in the five-phase case.

[0077] To influence the phase voltages of the inverter, space vector modulation can be used. Here, a reference voltage vector is formed by geometrically adding voltage state vectors.

[0078] For the three-phase case, a modulation level can be defined. This specifies a factored ratio between the amplitude of the fundamental frequency and the DC link voltage and is defined as follows in the three-phase case: m=u^1(Udc3)

[0079] This is illustrated graphically using Fig. 11 is shown. An inner circle 56 has a radius of 12 VDC This neglects hardware limitations in the linear range, resulting in a maximum modulation level of max(m). linear ) = 1.0 can be achieved.

[0080] In the five-phase case, separate modulation levels are defined for the fundamental oscillation and the third harmonic, each relating to the amplitude of its oscillation: m1=u^1(Udc3) m3=u^3(Udc2)

[0081] This is in Fig. Figure 12 is shown. An inner circle 56 has a radius of 155+25Udc, which results in a maximum modulation level m max This results in a value approximately 6.6% greater than the maximum modulation level in the three-phase case.

[0082] Based on Fig. Section 13 describes how to determine the reference vector for the three-phase case. For this purpose, the stress state vectors 1 and 3, which define sector I, are considered. Fig. The reference vector 13 is denoted by u*. The reference vector u* is formed from the geometric sum of the adjacent voltage state vectors 1 and 3. For this purpose, the voltage state vectors 1 and 3 are scaled, meaning they are each active only for a portion of the respective switching period of the switching signal at the clock rate. For each switching period, the desired reference vector u* is formed by the short-term average of the used voltage state vectors 1 and 3. It is possible that the sum of the durations of the voltage state vectors 1 and 3 is shorter than the duration specified by the switching period or the clock rate. In this case, any remaining missing duration can be provided by one of the two zero vectors. In the other sectors II to VI, the corresponding adjacent space vectors can be selected accordingly. u*=tU1tsw⋅U1+tU3tsw⋅U3+tU0.7tsw⋅U0.7 with tU0.7=tsw−(tU1+tU2)

[0083] In the three-phase case, the determination of the reference vector is unique. This results in a constant reference vector that rotates around the origin of the coordinate system at a constant angular velocity.

[0084] Preferably, the reference vector is also limited by the inner circle 56 during intended operation. This prevents undesirable overmodulation.

[0085] In the five-phase case, the larger number of available stress state vectors (0 to 32) and the three voltage levels result in a correspondingly large number of possibilities for providing the reference phasor. Among other things, two methods can be used for this purpose: a 2L method, in which two long stress state vectors adjacent by a sector are used, and a 2L2M method, in which two long stress state vectors adjacent by a sector and two correspondingly adjacent medium stress state vectors are used. This is illustrated by the Fig. 14 and Fig. 16 shown. At the Fig. Figure 14, which shows the 2L method, is based on the voltage state vectors 1 and 3, which form sector I. The resulting switching signals for the first sector I are shown in Fig. Figure 15 illustrates this. It is noticeable that two pairs of phases switch simultaneously, and these are the middle, or first and last, switching operations of a respective switching period. This is significant for the use of discontinuous modulation methods because, in such a case, two phases should maintain their state for a given switching period, thus saving a corresponding number of switching state changes of the switching elements 22 to 40. Otherwise, a conventional, centrally symmetrical switching pattern results.

[0086] Here too, the corresponding maximum modulation level can be achieved. However, it should be noted that in this case, the third harmonic is always present, regardless of the magnitude of the reference vector. This can be explained by the fact that there is no scaling possibility for the space vectors used, so they can geometrically eliminate each other in the α3 / β3 subsystem. The second subsystem, which in Fig. 14 and Fig. The element shown on the right in Figure 16 is disregarded in the regulation according to this method.

[0087] Fig. Figure 15 shows a schematic signal diagram showing normalized signal waveforms of the switching signals for the modulation method according to Fig. 14. This refers to t sw on the leap period.

[0088] The Fig. 16 and Fig. 17 show according to the Fig. 14 and Fig. 15 describes the facts for the 2L2M method. In the case shown, the voltage state vectors 1, 3, 19, 13 are used to achieve the reference vector u1* for sector I. This modulation method makes it possible to selectively reduce or even eliminate the third harmonic. Furthermore, voltage and current waveforms that are approximately trapezoidal can be achieved. Fig. Figure 17 shows the corresponding switching signals.

[0089] Furthermore, it is possible to use the sequence of the voltage state vectors to reduce the number of switching state changes of the switching elements 22, 24, 26, 28, 30, 32, 34, 36, 38, 40. This can be achieved by selecting a suitably efficient sequence of the voltage state vectors 0 to 31 used.

[0090] Fig. Figure 18 shows a schematic diagram illustrating a switching sequence for a situation where the number of switching state changes of the switching elements 22, 24, 26, 28, 30, 32, 34, 36, 38, 40 is reduced. Here, only the course of the Fig. The arrows shown in section 18, connecting the stress state vectors 0 to 31, are to be followed. In sector I, this results in the following sequence for the 2L2M method: 0→1→3→19→23→31→23→19→3→1→0. This can also be used analogously for other modulation methods.

[0091] In the aforementioned methods, short stress state vectors are not initially used to determine the reference vector. One reason for this is that short stress state vectors generally do not yield a long reference phasor, resulting in a lower modulation level and potentially a larger amplitude at the third harmonic. However, depending on the application, it may be advantageous to also consider short stress state vectors.

[0092] In both three-phase and five-phase systems, discontinuous modulation methods can also be used. This is generally also called discontinuous space vector modulation, specifically when at least one of the phases L1, L2, L3, L4, L5 is not switched within a switching period.

[0093] Several possibilities exist for this. For example, the central zero-space vector can be eliminated. The time of the central zero-space vector can then be added equally to the two outer voltage state vectors, thus preserving all information. Such a method is called, for example, DPWMmin, because for each switching period, one of the phases L1, L2, L3, L4, L5 is permanently connected to the minimum potential of the DC link 16. By eliminating the outer zero-space vectors, it can be ensured that at least one of the phases L1, L2, L3, L4, L5 is permanently connected to the maximum potential of the DC link 16 for each switching period. This is also called DPWMmax. Combinations of these methods are also possible, for example, DPWM0, DPWM1, DPWM2, DPWM3, or the like. Fig. 19 and Fig. Figure 20 shows schematic signal diagrams for normalized switching signals, where Fig. 19 shows the switching signals according to a 2L2M method using CPWM, whereas Fig. Figure 20 shows a corresponding representation of the switching signals according to the 2L2M method corresponding to DPWMmin.

[0094] The on-times of selected voltage state vectors can be determined using the reference vector u*, according to the following formulas: M⋅t=u*⋅tswt=M−1⋅u*⋅tsw

[0095] Using the 2L2M method, the components for sector I are as follows: M1=[U1U3U19U23]=[U1,α1U3,α1U19,α1U23,α1U1,β1U3,β1U19,β1U23,β1U1,α3U3,α3U19,α3U23,α3U1,β3U3,β3U19,β3U23,β3]

[0096] The matrix M contains the four Cartesian coordinates of the selected vectors, with the switch-on times to be calculated being summarized in the vector t. t1=[t1t3t19t23], u*=[uα1uβ1uα3uβ3], t0,31=tsw−t1−t3−t19−t23

[0097] It should be noted that in the five-phase case, the vectors of the second subsystem rotate at three times the angular velocity of the first subsystem. This results from the fact that the second subsystem represents the third harmonic. Furthermore, the starting point of the rotation in the second subsystem can deviate from the 0° axis or abscissa. Therefore, a starting angle does not need to be considered.

[0098] In order to utilize the maximum possible phase voltage, that is, to achieve the highest possible modulation level, the third harmonic must be considered as a function of the angle between the first and third harmonics. This is shown in the three-dimensional diagram according to... Fig. Figure 24 illustrates this. For the five-phase case, a general time function can be assumed as follows: x1,3(t)=A1 sin(ωt)+A3 sin(3ωt+φ31)

[0099] It can be shown that a maximum occurs in the following case: φ31=2πk−3(ωt)max−arccos(−A1 cos((ωt)max)3A3),k∈Z

[0100] As can be seen, an analytical solution to the aforementioned equation is not possible. However, the result can be approximated by a Taylor series. The signal waveform shows that a peak value occurs in a range around ωt=π2 occurs. When expanding a series to the aforementioned point, the following formula can be obtained after terminating after the first term: (ωt)max=φ31−2πA13A3−3+π2, for A3≠0

[0101] If this result is inserted into equation 4.21, an approximate peak value of the phase voltage can be obtained according to the following formula: x^1.3≈A1 sin((ωt)max)+A3 sin(3(ωt)max+φ31)

[0102] During optimization, the maximum possible modulation levels in the linear range should also be observed. The modulation levels are defined by formulas 2.2. Using: U1=ud12+uq12 U3=ud32+uq32 The modulation levels in the d / q plane can be defined as follows. m1=ud12+uq12(Udc3) m3=ud32+uq32(Udc3)

[0103] When considering the five-phase case according to Fig. As shown in section 9 and its properties, it becomes clear that m1 and m3 cannot be defined independently of each other. However, this would be useful for defining an overall modulation level.

[0104] Additionally, a relationship between the modulation intensities m1 and m3 can be determined by the selected modulation method, i.e., by the maximum resulting phasor length of the stress state vectors that yields the selected reference phasor. The overall modulation intensity can therefore be determined from the partial modulation intensities using Taylor series expansion. Simulations can be used to determine this.

[0105] For the five-phase synchronous machine 14, due to the symmetrically distributed winding, a normalization matrix W5 results according to formula 3.5 in order to transform the phase voltages defined in the time domain: W5=[10000000−10010000000−100100]

[0106] The transformation to the d / q plane is performed via the α / β plane. For this purpose, the Fortescue transformation matrix T5 is extended. Following the convention of the Clarke transformation matrix, this will be referred to as C5 (3.6) in the following. C5=25⋅[cos(0φ5)cos(1φ5)cos(2φ5)cos(3φ5)cos(4φ5)sin(0φ5)sin(1φ5)sin(2φ5)sin(3φ5)sin(4φ5)cos(0φ5)cos(3φ5)cos(6φ5) cos(9φ5)cos(12φ5)sin(0φ5)sin(3φ5)sin(6φ5)sin(9φ5)sin(12φ5)12cos(0φ5)12cos(5φ5)12cos(10φ5)12cos(15φ5)12cos(15φ5)]

[0107] The power-invariant transformation with the prefactor is used. 25 applied. Furthermore, the zero component, here the last row of the matrix, is not neglected, and the following holds true: φ5=2π5. This results in a square matrix, and by further scaling the zero components, an orthogonal matrix is ​​obtained. In this case, the inverse matrix corresponds to the transpose. This is advantageous for the inverse transformation, saves computational time, and simplifies the implementation.

[0108] Furthermore, the transformation matrix P5 (3.7) is defined. It is an extension of a three-phase Park transformation matrix and involves a rotation of the third harmonic component. The angle ϑ el was previously calculated in (3.2) taking (3.1) into account. In this, n mech Zp is the mechanical speed in revolutions per minute of the synchronous machine 14, and Zp is its number of pole pairs. The indices 1 and 3 refer to the fundamental frequency and its third harmonic, respectively. ωel=[ωel,1ωel,3]=nmech⋅2π60⋅Zp⋅

[13] ϑel=[ϑel,1ϑel,3]=L{∫ωel dt}=1s⋅ωel

[0109] To utilize the maximum phase voltage, the following consideration is applied. To exceed the maximum phase voltage, the phases must be driven with a duty cycle greater than 1. If the duty cycle is limited to 1, this limit can be observed, and a definition of the overall modulation level is unnecessary. To prevent the limit from being exceeded due to analytical errors during the calculation, and to ensure that hardware limitations are not completely disregarded, the limit value can be set to, for example, 98% of the maximum. DuCylimit≤1(⋅0.98)

[0110] Limiting the duty cycle during optimization can be computationally intensive. In each iteration, current quintuples are calculated. These should then be converted into d / q voltage values ​​using formula 3.4, neglecting the zero component. These voltage values ​​can then be transformed into a time signal using formula 3.18. From this, the five-phase auxiliary signal can be generated, which can be used, for example, to determine the duty cycles using the DPWMmin method. These steps should be performed in each iteration and can therefore slow them down.

[0111] The following explains the specification using phase voltages. With this method, a five-phase reference voltage can be tuned by specifying the modulation rates m1 and m3, as well as a phase shift between the fundamental frequency and the third harmonic. It can contain both the fundamental frequency and the third harmonic. The following applies: u*=[uL1*uL2*uL3*uL4*uL5*]=∑k=1,k≠23Udc⋅mk3⋅cos(ϑel,k+ωel,k2fsw)+[0−72−144−216−218]⋅π180⋅k+ϑinit,k

[0112] This allows for specification by d / q stream quintuples. The method is based on formula 3.3, but includes both a starting angle ϑ. init,1 or ϑ init,3 as well as the amplitudes of the fundamental oscillation u^1* or the third harmonic u^3* can be defined on the d / q plane via the following equation: [ud1uq1ud3uq3]=Rs[id1iq1id3iq3]+ωel,1[0−Lq100Ld1000000−3Lq3003Ld30][id1iq1id3iq3]+ωel,1[0ϕ103ϕ3]

[0113] The voltage calculated in this way can be the input to the system, while the specified currents can represent the output. This allows for a comparison of target and actual values ​​and a plausibility check of the overall model. Finally, by converting the d / q voltages of each d / q subsystem into polar coordinates, the respective amplitudes (magnitude) and phase shifts of the individual oscillations—namely, the fundamental frequency and the harmonics, especially the third harmonic—can be determined.

[0114] The advantage is that, firstly, this can be used for control purposes because the new input parameters can be determined from the resulting currents or phase currents. Secondly, the input variables in the d / q plane are known, allowing a direct comparison with the simulation output variables.

[0115] The Fig. Figure 23 shows a flat-top signal curve in a schematic diagram representation using a graph 54 of one of the phase voltages.

[0116] The upper diagram in Fig. 23 is assigned to the fundamental frequency and the lower diagram to the third harmonic. It can be seen that stress state vectors, spaced more than one sector apart, are used to determine the reference phasors.

[0117] Although the invention has been explained using a three-phase and a five-phase case, the invention is not limited to this and can be adapted to other multi-phase cases as needed without abandoning the idea of ​​the invention. Reference symbol list 0 to 32 stress state vectors 10 Drive system 12 inverters 14 multiphase electrical machine 16 DC link 18 Control unit 20 high-voltage batteries 22 Switching element 24 switching element 26 switching element 28 switching element 30 switching element 32 switching element 34 Switching element 36 switching element 38 switching element 40 switching element 42 Center tap 44 Center tap 46 Center tap 48 Center tap 50 Center tap 52 Stator winding 54 Graph 56 inner circle 58 Graph Sectors I to X L1 Phase L2 Phase L3 Phase L4 Phase L5 Phase N star point t sw Leap period

Claims

[1] Method for operating switching elements (22, 24, 26, 28, 30, 32, 34, 36, 38, 40) of an inverter (12) to which a stator winding of an at least three-phase electric machine (14) is connected, wherein the inverter (12) has at least one series connection of the switching elements (22, 24, 26, 28, 30, 32, 34, 36, 38, 40) for each of the phases (L1, L2, L3, L4, L5) of the electric machine (14) in order to electrically couple the electric machine (14) to a DC link (16) connected to the inverter (12), wherein switching signals for the switching elements are determined by vector control and provided at a clock rate greater than a frequency of a fundamental oscillation of the Phase voltages, wherein the switching signals are determined depending on at least one first reference pointer, which is determined bythat, for the switching states of the inverter (12) achievable by the switching elements (22, 24, 26, 28, 30, 32, 34, 36, 38, 40), first voltage state vectors are determined in a corresponding first space vector system using a Fortescue transformation adapted to the number of phases of the electrical machine (14), wherein the first reference vector is determined depending on a superposition of at least two of the first voltage state vectors, wherein, for determining the at least one first reference vector, at least two first voltage state vectors are chosen, one of which is a long first voltage state vector and the other a medium or a short first voltage state vector, and, for superimposing a harmonic of the phase voltage with respect to the frequency of the fundamental oscillation, the switching signals are additionally determined depending on at least one second reference vector.which is coupled to the first reference vector depending on an order number of the harmonics and which is determined by determining second stress state vectors in a corresponding second space vector system, wherein the second reference vector is determined depending on a superposition of at least two of the second stress state vectors and wherein a separate modulation level is determined for the harmonic and a total modulation level is maximized depending on the separate modulation level. [2] Method according to claim 1, characterized by, that immediately adjacent first stress state vectors form sectors in the space vector system, wherein the first reference vector is determined depending on a superposition of at least two of the first stress state vectors, and wherein, to determine the at least one first reference vector, at least two of the first stress state vectors are chosen which are spaced apart from each other by at least two adjacent sectors. [3] Method according to claim 1 or 2, wherein the harmonic of the phase voltage is odd. [4] Method according to claim 3, characterized by , that the switching elements (22, 24, 26, 28, 30, 32, 34, 36, 38, 40) in a star-connected stator winding are controlled in such a way that, through the action of the stator winding, an electrical potential of a star point of the star connection changes depending on the harmonic. [5] Method according to one of claims 3 or 4, characterized by , that in an electrical machine with at least four phases, at least one odd harmonic is used as a harmonic, the ordinal number of which is smaller than the number of phases of the electrical machine with at least four phases. [6] Method according to any one of the preceding claims, characterized by , that at least one of the at least two of the first stress state vectors is a medium or a short stress state vector. [7] Method according to any one of the preceding claims, characterized by , that for at least one switching period at least one of the phases (L1, L2, L3, L4, L5) is operated without switching. [8] Method according to any one of the preceding claims, characterized by , that an order regarding the use of the voltage state vectors is chosen such that a number of switching state changes of the switching elements (22, 24, 26, 28, 30, 32, 34, 36, 38, 40) is minimized. [9] Control unit (18) for operating switching elements of an inverter to which a stator winding of an at least three-phase electric machine is connected, wherein the inverter (12) has at least one series connection of the switching elements (22, 24, 26, 28, 30, 32, 34, 36, 38, 40) for each of the phases (L1, L2, L3, L4, L5) of the electric machine (14) in order to electrically couple the electric machine (14) to a DC link (16) connected to the inverter (12), wherein the control unit (18) is configured to determine switching signals for the switching elements (22, 24, 26, 28, 30, 32, 34, 36, 38, 40) by means of vector control and to provide them at a clock rate greater than a frequency of a fundamental oscillation of the phase voltages, wherein the control unit (18) is configured,to determine the switching signals as a function of at least one first reference vector by determining first voltage state vectors in a corresponding first space vector system for switching states of the inverter (12) achievable by the switching elements (22, 24, 26, 28, 30, 32, 34, 36, 38, 40) using a Fortescue transformation adapted to the number of phases (L1, L2, L3, L4, L5) of the electrical machine (14), wherein the control unit (18) is configured to determine the first reference vector as a function of a superposition of at least two of the first voltage state vectors, wherein the control unit (18) is configured to select two first voltage state vectors for determining the at least one first reference vector, one of which is a long first voltage state vector and the other a medium or a short first voltage state vector, and wherein the control unit (18) is configuredTo superimpose a harmonic of the phase voltage with respect to the frequency of the fundamental oscillation, the switching signals are additionally determined depending on at least one second reference phasor, which is coupled to the first reference phasor depending on an order number of the harmonic, and the control unit (18) is configured to determine second voltage state vectors in a corresponding second space vector system, wherein the control unit (18) is configured to determine the second reference phasor depending on a superposition of at least two of the second voltage state vectors, and wherein the control unit (18) is configured to determine a separate modulation level for the harmonic and to maximize an overall modulation level depending on the separate modulation level. [10] Drive system (10) with - an at least three-phase electrical machine (14) having a stator winding (52), - an inverter (12) to which the stator winding (52) is connected, wherein the inverter (12) has at least one series connection of switching elements (22, 24, 26, 28, 30, 32, 34, 36, 38, 40) for each of the phases (L1, L2, L3, L4, L5) in order to electrically couple the electric machine (14) to a DC link (16) connected to the inverter (12), and - a control unit (18) for operating the switching elements (22, 24, 26, 28, 30, 32, 34, 36, 38, 40) of the inverter (12) in order to provide respective phase voltages through the inverter (12), wherein the control unit (18) is configured according to claim 9.

Images