OPERATION OF SWITCHING ELEMENTS OF A MULTI-LEVEL POWER CONVERTER

Description

[0001] The invention relates to a method for operating switching elements of a multi-level energy converter using at least three electrical potentials, to which a multi-phase electrical machine is connected, wherein switching elements of the multi-level energy converter are operated by means of corresponding switching signals in a predetermined clock mode, wherein the switching signals for the switching elements assigned to a respective one of the phases are determined as a function of a respective target voltage signal in order to apply a respective phase voltage to each phase of the multi-phase electrical machine.Furthermore, the invention relates to a control unit for controlling switching elements of a multi-level energy converter which uses at least three electrical potentials and is designed to be connected to a multi-phase electrical machine, wherein the control unit is designed to operate the switching elements of the multi-level energy converter by means of corresponding switching signals in a predetermined clock mode, for which purpose the control unit determines the switching signals for the switching elements assigned to a respective one of the phases as a function of a respective target voltage signal in order to apply a respective phase voltage to each phase of the multi-phase electrical machine.Furthermore, the invention relates to a multi-level energy converter designed for connection to a multi-phase electrical machine and utilizing at least three electrical potentials, comprising switching elements for applying a respective phase voltage to each phase of the multi-phase electrical machine, and a control unit for controlling the switching elements in a predetermined cycle mode using corresponding switching signals. Finally, the invention also relates to a drive device comprising a multi-phase electrical machine and a multi-level energy converter connected to the multi-phase electrical machine, which utilizes at least three electrical potentials.Methods for operating switching elements of a multi-level energy converter, control units therefor, multi-level energy converters with control units, and drive devices of the generic type are extensively known in the prior art, so that a separate written reference is not required. Multi-level energy converters are specific energy converters that utilize multiple electrical potentials to perform the energy conversion function, which can be provided, for example, by one or more DC intermediate circuits. The invention is particularly directed at such multi-level energy converters that utilize at least three electrical potentials to implement the desired energy conversion function.

[0002] Such multi-level energy converters are frequently used in the energy supply sector or in power drives. Preferably, such multi-level energy converters are operated bidirectionally, allowing electrical energy to be converted from the AC voltage side to the DC voltage side and vice versa.

[0003] A multi-level energy converter is known, for example, from WO 02 / 23703 A1.

[0004] US 2005 / 071090 A1 discloses the estimation of temperature changes in semiconductors.

[0005] EP 2 849 331 A1 aims to balance the voltage across a capacitive voltage divider of a multi-level energy converter. This compensates for a voltage difference at the neutral point.

[0006] The paper "An Improved Circulating Current Injection Method for Modular Multilevel Converters in Variable-Speed ​​Drives" (IEEE TRANSACTIONS ON INDUSTRIAL, ELECTRON I CS., Vol. 63, No. 11, November 1, 2016 (2016-11-01), pages 7215-7225, XP055600388, USA) deals with the suppression of ripple voltages in capacitors and the reduction of the required capacitance of a multilevel energy converter. The goal is to keep fluctuations in the capacitor voltage within a certain tolerance interval, which requires a corresponding power exchange. A corresponding common-mode voltage is only used up to a cutoff frequency.

[0007] WO2019137916 A1 discloses a method for controlling an electrical converter, wherein the electrical converter is designed to convert a direct voltage into a multi-phase voltage with at least two voltage levels.

[0008] Multilevel energy converters are typically used to couple one or more DC link circuits to an AC network, especially a multiphase AC network. The AC network is often provided by a corresponding multiphase electrical machine. The multilevel energy converter is adapted to the multiphase electrical machine so that each phase of the electrical machine can be supplied with electrical energy accordingly.

[0009] Multiphase electrical machines are also widely known in the prior art. These machines are used in a variety of applications in which specific drive functions are to be implemented using the multiphase electrical machine. Three-phase electrical machines are particularly widespread, and depending on the application, they can even be directly coupled to a three-phase public power grid to provide the desired drive function. However, the control options in such a scenario are limited, which is why multiphase electrical machines are now generally operated via individually assigned energy converters, particularly inverters.This also allows the electrical energy required for the intended operation of the electrical machine to be provided from an electrical energy storage device, for example, a battery, in particular a high-voltage battery, a photovoltaic device, and / or the like. This makes it possible, for example, to provide drive systems that are suitable not only for stationary applications but also for mobile use, for example, in motor vehicles. A multitude of other applications can also be realized in this way.

[0010] Particularly when the multi-level energy converter is designed as an inverter, it is possible to control the multi-phase electrical machine flexibly, even when the electrical machine has more than three phases, for example four phases, five phases, six phases, or even more phases. Even though three-phase machines currently predominate or are very widespread, it can be particularly useful for specific applications for the electrical machine to have more than three phases, such as ship propulsion or the like. The electrical machine can be designed, for example, as a synchronous machine, an asynchronous machine, or the like. Such electrical machines are generally designed to be supplied with alternating voltage.

[0011] The use of an inverter, in particular, to operate a multiphase electrical machine opens up the possibility of operating the multiphase electrical machine in a highly dynamic manner. The multiphase electrical machine is typically designed as a rotating electrical machine with a stator-shaped stand, which generally provides a substantially circular opening for accommodating a rotor-shaped armature. The armature is rotatably mounted in the opening, with an air gap formed between the armature and the stator.

[0012] A rotating electrical machine is a device that converts electrical energy into mechanical energy, particularly kinetic energy in motor operation and / or mechanical energy into electrical energy in generator operation. This movement is usually a rotary motion performed by the rotor. Unlike the rotor, the stator is usually non-rotatably mounted, meaning that a rotary motion is a rotational motion of the rotor relative to the stator.

[0013] The stator and rotor of a rotating electrical machine are linked by a magnetic flux, which generates the force or torque during motor operation, which drives the rotor to rotate relative to the stator. In the generator, the mechanical energy or torque supplied to the rotor is converted into electrical energy in the form of rotation. For this purpose, the stator has a corresponding multiphase winding through which the respective phase currents flow. In the stator or rotor, the winding can also be replaced or at least partially supplemented by a permanent magnet.

[0014] Rotating electrical machines of this type include, for example, induction motors connected to a corresponding multi-phase, particularly three-phase, AC electrical network, such as asynchronous motors, synchronous motors, synchronous motors with a damper cage, or the like. The AC network is preferably provided by the inverter.

[0015] The stator and rotor typically comprise a magnetically active unit, which, in addition to the respective windings, may also include at least respective laminated cores. A laminated core is typically made of individual, electrically insulated ferromagnetic sheets. Depending on their intended use, the sheets may be referred to as dynamo sheets, motor sheets, transformer sheets, or similar. Such sheets are subject to standards, for example, EM 10106, EM 10107, or similar.

[0016] When operating inverters, especially when they are designed as multi-level energy converters, losses arise, among other things, due to the switching operation of the switching elements.

[0017] In addition, conduction losses also occur in the switching elements when they are in the switched-on state and are subjected to a respective electrical current. The switching signals for controlling the switching elements are often generated and provided by a control unit within the framework of pulse width modulation (PWM). The control unit can be included in the respective multi-level energy converter. However, it can also be provided at least partially by a higher-level controller, for example, a controller of the drive device or the like.

[0018] In pulse-width modulation, two DC voltage potentials are provided as a switching signal in a time range shorter than a fundamental oscillation period, or a clock rate (also called a PWM period), at a temporal ratio (duty cycle) such that the voltage averaged over the PWM period approximately corresponds to a desired phase voltage at a particular point in time during the PWM period. The respective phase current is adjusted accordingly.

[0019] The temporal course of the aforementioned mean value of the respective phase voltage is decisive for the formation of a respective phase current in relation to a fundamental oscillation of the frequency of the phase voltage.

[0020] Multi-level energy converters, in which a respective phase connection can be coupled to one of three electrical potentials by means of respective switching elements, are also called three-level inverters or three-point inverters.

[0021] The phase current can develop according to the difference between the phase voltage provided by the energy converter and an internal voltage of the electrical machine. The closer a current value of a respective phase voltage is to one of the available electrical potentials, the longer the respective phase connection is connected to this electrical potential in a corresponding PWM period. If, for example, small phase voltages are provided, this can result in a permanent proximity with regard to the mean electrical potential in a three-level inverter. Therefore, in this type of operating state, a so-called midpoint path, i.e. the switching element between the phase connection and the middle of the electrical potentials of the inverter, is switched on for a relatively long time. This can lead to large conduction losses there.In addition, however, switching losses due to the necessary switching operations to the respective other electrical potentials must also be taken into account.

[0022] In many practical applications, for example, when using asynchronous machines, there is an approximately proportional relationship between a respective phase voltage with respect to the fundamental oscillation and the respective frequency of the fundamental oscillation. In particular, this can lead to a situation where, at low phase voltages, the corresponding frequency is also low, meaning that the duration of the fundamental oscillation period is correspondingly long. This can cause the conduction losses in the midpoint path to occur over a long period of time compared to a positive path and / or a negative path, thus leading to significant thermal stress and correspondingly large temperature fluctuations in the affected switching elements of the midpoint path.

[0023] A switching element within the meaning of this disclosure is preferably a controllable electronic switching element, for example, a controllable electronic semiconductor switch such as a transistor operated in a switching mode, a thyristor, combination circuits thereof, preferably with parallel-connected inverse diodes, a gate-turnoff thyristor (GTO), an insulated-gate bipolar transistor (IGBT), combinations thereof, or the like. In principle, 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 of the multi-level energy converter or inverter, the switching elements are operated in switching mode. In relation to a semiconductor switch using a transistor, switching mode means that, in an on-state, a very low electrical resistance is provided between the transistor terminals forming a switching path, allowing a high current flow with a very low residual voltage. In an off-state, however, the transistor's switching path has a high impedance, meaning it provides a high electrical resistance, so that even with a high electrical voltage applied across the switching path, there is essentially no current flow or only a very low, in particular negligible, current flow.This differs from linear operation in transistors, which, however, is generally not used in multi-level energy converters or inverters of this type. When switching from the on to the off state, the current flowing through the switching element must dissipate and the voltage applied to the switching element must build up. The simultaneous occurrence of voltage and current results in energy loss during this transition period.

[0025] Likewise, energy loss occurs during the transition from the switched off to the switched on state.

[0026] To implement the switching functionality, the switching element has at least one control terminal at which it can be supplied with the switching signal provided by the clock generator, so that the desired switching function of the switching element can be implemented. The switching signal can be a binary switching signal that can assume two state values ​​in order to provide the desired switching functions of the switching element. For example, the switching signal can be formed by a pulse sequence by means of which the control terminal is supplied. This is particularly useful for thyristors and GTOs. Furthermore, in the case of transistors, the switching signal can be designed as a square-wave signal, wherein a respective switching state of the switching element can be assigned to one of the electrical potentials of the square-wave signal.Such a signal is useful, for example, for transistors, especially bipolar transistors, field-effect transistors, or the like. A temporal relationship between the two electrical potentials usually determines the duty cycle.

[0027] The control unit provides the functionality related to generating the switching signals. Other functions, particularly related to the operation of the multi-level energy converter or inverter, can also be implemented by the control unit, for example, monitoring functions, safety functions, and / or the like. For this purpose, the control unit can comprise a hardware circuit and / or a program-controlled computer unit or the like. Naturally, the control unit can be designed as a separate module. However, the control unit can also be at least partially comprised by the higher-level controller, for example, a controller for the drive device or the like.

[0028] The multi-level energy converter is formed by an electronic hardware circuit that has switching elements corresponding to the number of phases of the multi-phase electrical machine, by means of which each of the phases of the electrical machine can be electrically coupled to the respective electrical potentials of the multi-level energy converter. For this purpose, the multi-level energy converter comprises corresponding phase terminals to which the electrical machine can be connected. The desired energy coupling can be achieved by suitable pulse patterns of the switching signals according to pulse width modulation. The basic operating procedures in this regard are known, which is why they will not be discussed further.For the intended operation of the energy conversion, the clock rate of the switching signals is usually considerably higher than the frequency or the fundamental frequency of the phase alternating currents or the phase voltages.

[0029] The switching elements should not exceed a maximum rated temperature during intended operation. Above the rated temperature, reliable operation of the switching elements can no longer be guaranteed. This could result in malfunctions and / or failures of the switching elements.

[0030] Furthermore, it should be noted that cyclic temperature changes can lead to aging of the switching elements, especially if they are designed as semiconductor switching elements, even below the rated temperature. Such temperature fluctuations can occur, for example, during the course of a respective fundamental oscillation period corresponding to the frequency of the fundamental oscillation of the alternating phase currents or phase voltages. This can also lead to failure of the switching elements. Particularly in a three-level inverter, the provision of phase voltages with a small amplitude and a simultaneously low fundamental frequency can lead to unfavorable thermal stress on the center point path, resulting in significant overall aging, at least with regard to the center point path.

[0031] The invention is based on the object of developing a generic method, a control unit, a multi-level energy converter and a drive device in such a way that an improved thermal load on the switching elements can be achieved, in particular in partial load operation.

[0032] As a solution, the invention proposes a method, a control unit, a multi-level energy converter and a drive device according to the independent claims.

[0033] Advantageous further training results from features of the dependent claims.

[0034] With regard to a generic method, it is proposed in particular that the switching signals are additionally determined as a function of a predetermined superimposed voltage signal in such a way that a superimposed voltage dependent on the superimposed voltage signal is superimposed on the respective phase voltages.

[0035] With regard to a generic control unit, it is proposed in particular that the control unit is further designed to determine the switching signals additionally as a function of a predetermined superimposed voltage signal in order to superimpose a superimposed voltage dependent on the superimposed voltage signal on the respective phase voltages.

[0036] With regard to a generic multi-level energy converter, it is particularly proposed that it comprises a control unit according to the invention.

[0037] With regard to a generic drive device, it is particularly proposed that it comprises a multi-level energy converter according to the invention.

[0038] The invention is based on the idea that a potential shift of the phase voltages better distributes the stress on the switching elements across all switching elements of the multi-level energy converter, particularly at small phase voltage amplitudes and low fundamental phase voltage frequencies, or in the partial load range. This is achieved by superimposing the superimposed voltage on each of the phase voltages. The superimposed voltage is preferably the same for all phase voltages. Since the multi-phase electrical machine is generally connected to the multi-level energy converter in a floating manner, this need not lead to any changes in the operating characteristics of the electrical machine.

[0039] This principle is neither limited to the use of three electrical potentials by the multi-level energy converter nor to application in three-phase electrical machines. The multi-level energy converter can therefore also use four different electrical potentials, five electrical potentials, or even more different electrical potentials. In the case of a multi-phase electrical machine, the number of phases is also irrelevant. The invention therefore provides a simple method for achieving a more even load on the switching elements without having to accept any further disadvantages.

[0040] For example, the clock rate could also be reduced. While lowering the clock rate allows for a reduction in switching losses of the switching elements, it also increases the so-called ripple current and its effects. Ripple current is caused by a difference between the phase voltage and the output voltage actually provided by the multi-level energy converter or inverter within a PWM period at the respective phase connection and is superimposed on the actual phase alternating current. This can lead to higher losses, higher temperatures and noise in the connected multi-phase electrical machine, as well as potentially other undesirable side effects.

[0041] Furthermore, the alternating phase current could also be reduced. This would reduce not only the switching losses but also the conduction losses of the switching elements. However, this would naturally also result in reduced power, which in turn would result in a corresponding reduction in the torque of the electric motor.

[0042] Furthermore, timing gaps in the switching signals could also be considered. Timing gaps are periods of time within a respective period of the fundamental oscillation or fundamental frequency during which no switching operations of the switching elements take place in a respective phase. However, to ensure even loading of the switching elements, the timing gaps would have to alternate cyclically between all switching elements. Such pulse patterns are also referred to as discontinuous. The disadvantage here, however, is that the phase alternating currents have less favorable spectra. For example, the effective frequency of a line-to-line voltage can alternate between twice the pulse frequency of a respective phase when both phases involved are switching, and once the pulse frequency when one of the two phases has a timing gap. This also increases the ripple current superimposed on the phase alternating current.

[0043] Furthermore, the common-mode voltage that typically arises in this case proves to be disadvantageous. For example, after one-sixth of a fundamental oscillation period, all phase voltages can jump relative to a reference potential, such as ground potential. Such voltage jumps occur alternately in both directions, their magnitude corresponding to a direct voltage, and their steepness generally depends on the switching speed of the switching elements. In conjunction with parasitic capacitances, for example, between the winding of the multi-phase electrical machine and a grounded housing of the electrical machine, comparatively large common-mode currents can result, which can also lead to corresponding disturbances in electromagnetic compatibility (EMC).

[0044] These disadvantages can be avoided by the method according to the invention. It takes advantage of the fact that the operating behavior of multi-phase electrical machines depends on the interlinked phase voltages, which is why these phase voltages can be shifted by an equal amount without impairing the intended function of the electrical machine.

[0045] Although a sinusoidal waveform is provided for the alternating phase currents or phase voltages in this case, the invention is not limited to this. The alternating phase currents or phase voltages can, of course, also have a different harmonic waveform, which may depend on the specific application.

[0046] According to the invention, the superimposed voltage signal is therefore additionally taken into account when determining the switching signals, wherein the switching signals are additionally determined depending on the predetermined superimposed voltage signal in such a way that a superimposed voltage dependent on the superimposed voltage signal is superimposed on the respective phase voltages.

[0047] This superimposed voltage is preferably the same for all phase voltages, so that the effects on the multi-phase electrical machine, in particular, can be kept largely small. The superimposed voltage signal can be used to influence the duty cycle of a switching signal unit of the control unit, which provides respective switching signals for the switching elements from predetermined values ​​for the phase voltages or alternating phase currents, such that the respective phase voltages are provided at the respective phase terminals of the multi-level energy converter superimposed with the superimposed voltage.This makes it possible to reduce the problem described at the beginning regarding the unfavorable load, in particular with regard to a midpoint path in a three-level inverter, because the load is at least partially shifted to the other paths by shifting the respective phase voltages due to the superimposed voltage.

[0048] This makes it possible to reduce the load on the midpoint path, particularly with small phase voltage amplitudes and low fundamental phase voltage frequencies. The other paths are correspondingly more heavily loaded, resulting in a more even thermal load on the switching elements. The principle is fundamentally independent of how many electrical potentials the multi-level energy converter uses to achieve the desired energy conversion function. Likewise, the principle is naturally independent of how many phases the multi-phase electrical machine has. In principle, therefore, the invention can at least partially shift the load on at least one of the potentials to other potentials.

[0049] The invention thus makes it possible to ensure that the switching elements of the multi-level energy converter are loaded more evenly overall, so that in addition to greater stress on the multi-level energy converter as a whole, aging of the switching elements as a whole and thus of the entire multi-level energy converter can be reduced. This can increase the load capacity of the multi-level energy converter as a whole and / or reduce its design with regard to structural stresses. Reliability can be increased and / or costs can be reduced.

[0050] The phase voltages are alternating voltages at the respective phase terminals of the multi-level energy converter or inverter or the electrical machine, and refer to alternating voltages at the respective phase terminals of the multi-level energy converter or inverter to which the multi-phase electrical machine is connected. For each of the phases of the multi-phase electrical machine, at least one phase terminal is preferably provided by the multi-level energy converter or inverter, to which the respective phase of the multi-phase electrical machine can be connected.For each phase connection, the multi-level energy converter or the inverter provides at least one circuit arrangement with a suitable number of switching elements, by means of which the respective phase connection can be selectively electrically coupled to at least one of the electrical potentials of the multi-level energy converter or the inverter.

[0051] Typically, the respective phase currents with respect to the multiphase electrical machine are adjusted by suitable pulse patterns of the switching signals provided by the control unit, which are used to control the respective switching elements. The associated phase voltages can then be determined from these pulse patterns, for example, by filtering the respective pulse pattern using a suitable low-pass filter, determining a Fourier transform, determining the respective phase voltage using a power meter, and / or the like. The phase voltage therefore does not correspond to the direct pulse pattern provided directly at the respective phase connection of the multi-level energy converter or inverter.For this reason, the frequency of the fundamental oscillation of the respective phase voltages is usually significantly lower than the clock rate of the switching signals according to which the respective pulse patterns are provided at the respective phase terminals.

[0052] According to a further development, it is proposed that the superimposition occurs at least when a frequency of the target voltage signals is lower than a predetermined comparison frequency. The predetermined comparison frequency can, for example, be a frequency of approximately 30 Hz or less. This can take into account that thermal time constants of the switching elements are relatively small and a corresponding temperature swing can already be significant for the respective switching element. Above the comparison frequency, the provision of the superimposed voltage signal can be reduced or even switched off. As a result, the control or operation of the switching elements according to the invention only needs to be provided if this proves expedient due to the operating characteristics of the multi-level energy converter, in particular its switching elements, during intended operation.Since the fundamental frequency of the target voltage signals usually corresponds essentially to the fundamental frequency of the phase alternating currents, the comparison frequency can of course also be related to the fundamental frequency of the phase alternating currents.

[0053] According to a further development, it is proposed that the superimposition occurs at least when an amplitude of the target voltage signals is smaller than a predetermined comparison amplitude. This can prevent the electrical potentials or the switching elements electrically coupled to them, which are relevant for the realization of such phase voltages, from being subjected to particular stress. These are preferably electrical potentials or their switching elements that are in a middle range between the maximum positive and the maximum negative electrical potential. In a three-level inverter, this is preferably the middle potential or the midpoint path. By superimposing the superimposed voltage, the stress can thus be distributed to other electrical potentials or the switching elements assigned to them.Above the reference amplitude, the control of the switching elements or the operation of the switching elements according to the invention therefore need not be implemented. Of course, this further development can also be combined with the previous development with regard to frequency.

[0054] It can be provided that a separate superimposed voltage signal is provided for each of the phases, by means of which amplitudes of the respective superimposed voltages are individually adjusted for each of the phases. This makes it possible to individually adjust the respective superimposed voltage for each of the phases. This has the advantage that tolerances, in particular with regard to the provision of the target voltage signals as well as with regard to the electrical properties of the respective phases of the multi-phase electrical machine, phase-specific peculiarities and / or the like can be better taken into account or even compensated for. In addition, other applications can of course also be provided, wherein different phases are subjected to different superimposed voltages in such a way that these can compensate for each other during normal operation of the drive device.This allows specific effects to be achieved that may be relevant for specific applications. Preferably, however, the superimposed voltage signals and the correspondingly assigned superimposed voltages are essentially the same magnitude.

[0055] It is further proposed that temperatures of the switching elements be determined, and that the superimposed voltage signal be adjusted depending on the determined temperatures. This development has the advantage that the control of the switching elements can be specifically adapted to a respective thermal load on the switching elements. This makes it easy to achieve a moderately uniform stress on the switching elements, particularly with regard to thermal stress. The superimposed voltage or the superimposed voltage signal therefore does not necessarily have to be constant, but can be varied as needed. The superimposed voltage signal can be a DC voltage signal, whereby the amplitude of the DC voltage can be adjusted as needed.In addition to the temperature dependence, the amplitude of the superimposed voltage signal can naturally also depend on the amplitude of the phase voltage and / or the phase alternating current, and possibly also on a frequency, in particular a fundamental oscillation, of the phase voltage or the phase alternating current. The temperatures of the switching elements can be determined in a known manner, for example, by detecting the switching properties of the switching elements or the like. Of course, it can also be provided that the switching elements have respective temperature sensors that are connected to the control unit, and whose sensor signals are evaluated accordingly by the control unit. Combinations of these can also be provided.

[0056] Furthermore, it can be provided that aging states of the switching elements are determined, and the superimposed voltage signal is adjusted depending on the determined aging states. In semiconductor switching elements, an increase in the respective temperature during normal operation proves to be a particularly indicative indicator of an aging state. This can be additionally taken into account for the provision of the superimposed voltage signal, so that, if necessary, different aging states of switching elements of the multi-level energy converter can also be used to adjust the superimposed AC voltage.

[0057] The superimposed voltage signal can, for example, be a DC voltage signal. It can also, of course, be formed by an AC voltage signal. This makes it possible to cyclically distribute the load across the available electrical potentials or their switching elements according to the AC voltage signal. Depending on the type of AC voltage or taking the frequency into account, an appropriate distribution of the thermal load can be achieved. The load capacity of the multi-level energy converter or inverter, especially its switching elements, can thus be further improved.

[0058] Furthermore, it is proposed that an amplitude and / or a frequency of the AC voltage signal be adjusted depending on a thermal time constant of the switching elements. This allows for better consideration of special properties of the switching elements as well as of the multi-level energy converter or inverter as a whole. For example, the frequency of the AC voltage can be selected such that the temperature variation of a particular switching element is as small as possible. Thermal time constants of the switching elements can also be taken into account.

[0059] The alternating voltage can be a sinusoidal alternating voltage. However, a trapezoidal alternating voltage is particularly advantageous.

[0060] Of course, these embodiments can also be combined with one another. The combining can include changing the signal shape as needed, for example, at a predetermined time interval, depending on detected temperatures of the switching elements, and / or the like. The signal shape of the alternating voltage can be variable, for example, depending on a detected load on the switching elements, an operating state of the multi-level energy converter, and / or the like. The alternating voltage signal is preferably a trapezoidal alternating voltage signal. This embodiment is particularly suitable when, for example, in a three-level inverter, the center path is designed to be less resilient than the other paths in terms of its load capacity.In such an inverter, operation according to the state of the art can therefore quickly lead to significant stress on the correspondingly small-sized switching elements of the medium-voltage path. By selecting a trapezoidal AC signal, the unfavorable load for the midpoint path can be better distributed among the other paths. Overall, the signal shape of the AC signal can, of course, also be adapted to the design of the multi-level energy converter or inverter in order to better accommodate different switching element designs. This can, of course, also include cooling of the switching elements.

[0061] The advantages and effects stated for the method according to the invention naturally also apply equally to the control unit according to the invention, the multi-level energy converter according to the invention, and the drive device according to the invention, and vice versa. Accordingly, device features can also be formulated for method features, and vice versa.

[0062] Further features, advantages, and effects will become apparent from the following description of exemplary embodiments with reference to the figures. In the figures, identical reference numerals denote identical features and functions.

[0063] They show: FIG 1 shows a schematic circuit diagram of a drive device with a three-phase asynchronous machine connected to a three-level inverter; FIG 2 shows a schematic circuit diagram of the three-level inverter according to FIG 1 ; FIG 3 a schematic diagram of target voltage signals for phase voltages at phase terminals of the inverter according to FIG 2 at a small voltage amplitude, FIG 4 a temporally enlarged section of area IV in FIG 3 , FIG 5 a schematic diagram of output voltages at phase terminals of the inverter according to FIG 2 , which according to FIG 4 have been determined, FIG 6 a schematic diagram representation as FIG 4 for a selected one of the phases, showing the effect of a superimposed voltage for a shift towards a positive potential, FIG 7 a schematic representation as FIG 5 for an output voltage at the phase connection according to FIG 6 , FIG 8a schematic diagram representation such as FIG 6 for a negative output voltage at the phase connection, showing the effect of a superimposed voltage for a shift towards a negative potential, FIG 9 a schematic diagram representation such as FIG 7 for determining the output voltage according to FIG 8 , FIG 10a schematic diagram representation such as FIG 4 , in which the phase voltages are superimposed by a superposition voltage for a shift to the positive potential, FIG 11 a schematic diagram representation such as FIG 5 for the switching signals for corresponding switching signals according to FIG 10 , FIG 12 a schematic diagram representation such as FIG 10 , whereby the superposition voltage now causes a shift to the negative potential, FIG 13 a schematic diagram representation such as FIG 11 for the scenario according to FIG 12 , FIG 14 another schematic diagram representation as FIG 10 , in which the superposition voltage is smaller than in the design according to FIG 10 , FIG 15 a schematic diagram of the switching signals for the scenario according to FIG 14 , FIG 16 a schematic diagram representation for individual superposition voltages for the phase voltages according to a first embodiment, FIG 17 a schematic diagram representation such as FIG 16 for a second embodiment, FIG 18 a schematic diagram representation as FIG 16 for a third embodiment, FIG 19 a schematic diagram of phase voltages superimposed by the superposition voltages, FIG 20 a schematic diagram of superposition voltage signals according to FIG 19 , and FIG 21 a schematic diagram in which a current carrying capacity of the switching elements is shown by means of graphs for different frequencies of the fundamental oscillation of the phase voltages.

[0064] FIG 1 shows a schematic circuit diagram of a drive device 10 according to the invention, which comprises a three-phase asynchronous machine 30 having respective phases U, V, W. Furthermore, the drive device 10 comprises a three-level inverter 28 connected to the three-phase asynchronous machine 30 as a multi-level energy converter that utilizes three electrical potentials 22, 24, 26. The three-level inverter 28 has phase terminals 32, 34, 36 to which the phases U, V, W of the asynchronous machine 30 are connected.

[0065] The three-level inverter 28 is further connected to a first DC intermediate circuit 12 and a second DC intermediate circuit 14, which are electrically coupled to one another at a connection point 20. This provides the electrical potentials 22, 24, 26 for the three-level inverter 28. The DC intermediate circuits 12, 14 provide respective DC link voltages U1, U2, which are approximately equal in this case. However, they can also differ from one another as needed.

[0066] During normal operation, the DC intermediate circuits 12, 14 are coupled to the asynchronous machine 30 for energy-related purposes by means of the inverter 28. In the present case, bidirectional energy exchange is enabled, so that the asynchronous machine 30 can be operated in four-quadrant mode. The asynchronous machine 30 can thus be operated both as an electric motor and as an electric generator. The inverter 28 establishes the corresponding energy-related coupling to the DC intermediate circuits 12 and 14. For its normal operation, the three-level inverter 28 comprises a control unit 38, which is communicatively coupled to a higher-level controller (not shown) of the drive device 10. In other embodiments, however, the control unit 38 can also be formed at least partially in one piece with the higher-level controller of the drive device 10.

[0067] FIG 2 shows a further schematic circuit diagram of the basic structure of the three-point inverter 28. The three-point inverter 28 comprises three circuit arrangements 58, 60, 62, which comprise respective switching elements 40 to 56. The circuit arrangement 58 comprises the switching elements 40, 42, 44. The circuit arrangement 60 comprises the switching elements 46, 48, 50. The circuit arrangement 62 comprises the switching elements 52, 54, 56. The switching elements 40, 42, 44 of the circuit arrangement 58 are each connected by one of their connections to the electrical potential 22. Accordingly, the switching elements 46, 48, 50 of the circuit arrangement 60 are each connected by one of their connections to the electrical potential 24. Finally, the switching elements 52, 54, 56 are each connected by one of their connections to the electrical potential 26. The switching elements in this case are formed by IGBTs.

[0068] The second terminals of the switching elements 40, 46, 52 are each connected to the phase terminal 32, the second terminals of the switching elements 42, 48, 54 are each connected to the phase terminal 34 and the second terminals of the switching elements 44, 50, 56 are each connected to the phase terminal 36.

[0069] Due to the aforementioned circuit topology, each of the phase terminals 32, 34, 36 can be electrically coupled to a respective one of the electrical potentials 22, 24, 26, depending on the switching state of the respective switching elements 40 to 56. The switching elements 40 to 56 electrically coupled to a respective one of the phase terminals 32, 34, 36 are generally electronically interlocked with one another, so that they cannot simultaneously assume the switched-on state. This prevents short circuits between the electrical potentials 22, 24, 26.

[0070] The three-level inverter 28 further comprises the control unit 38, which serves to control the switching elements 40 to 56 in a predetermined clock operation, for which purpose it generates respective switching signals in order to produce output voltages 70, 72, 74 at the respective phase terminals 32, 34, 36 ( FIG 5 ). The switching signals are generated in this case according to pulse width modulation, as will be explained further below.

[0071] FIG 3 shows, in a schematic diagram, nominal voltage signals 64, 66, 68, which are available for pulse width modulation to generate respective phase voltages 70, 72, 74 at the respective phase terminals 32, 34, 36. An ordinate is assigned to the electrical potential, whereas an abscissa is assigned to time. In the diagram according to FIG 3 A graph 76 illustrates a period of the fundamental oscillation of the target voltage signals 64, 66, 68. The target voltage signals 64, 66, 68 are each phase-shifted from one another, in this case by approximately 120°.

[0072] Out of FIG 3 It can be seen that the voltage amplitude of the setpoint voltage signals 64, 66, 68 is comparatively small, which is why a midpoint path of the three-level inverter 28, which is formed by the circuit arrangement 60, is subjected to particularly high loads. This operating state occurs in the present drive device 10 when the asynchronous machine 30 is operated at a low speed. Typically, the voltage amplitude of a respective one of the setpoint voltage signals 64, 66, 68 is approximately proportional to the frequency of the fundamental oscillation. This affects the inverter 28 in that the corresponding circuit arrangement 60, and in particular its switching elements 46, 48, 50, are subjected to particularly high thermal loads at low fundamental oscillation voltages.

[0073] FIG 4 shows a temporally enlarged view of an area IV in FIG 3 . The reference numerals 64, 66, 68 denote corresponding setpoint voltage signals for the phase voltages 70, 72, 74 ( FIG 5 ) as they are processed by the control unit 38. In order to generate the switching signals for the phase voltages 70, 72, 74 according to the pulse width modulation, the desired voltage signals 64, 66, 68 are fed to respective comparators (not shown) of the control unit 38, which compare these signals with corresponding triangular signals which are FIG 4 are shown by the graphs 80, 82. The graph 80 shows a triangular signal for nominal voltage signals 64, 66, whose potentials are between the electrical potentials 22, 24. The graph 82 shows a triangular signal which serves to generate nominal voltage signals such as the nominal voltage signal 68 according to FIG 4 to compare which is between the potentials 24, 26.

[0074] The comparison operation results in corresponding switching signals for the phase voltages 70, 72, 74, as shown in FIG 5 These switching signals lead to corresponding phase voltages at the phase terminals 32, 34, 36. In the FIG 4 und 5 Each graph 78 represents a single clock period of the clock rate of the switching signals. The ordinate here again represents the electrical potentials, whereas the abscissa represents time.

[0075] FIG 6 shows in a schematic diagram how FIG 4 the situation for a single phase, in this case for phase U, if a superimposed voltage is applied to the comparison function during switching signal generation by means of a superimposed voltage signal 86 ( FIG 20 ) is superimposed. As can be seen from FIG 6 As can be seen, the switching times of the switching signal for the phase voltage 70 are obtained taking into account the time periods t3, t4, as can be seen from FIG 7 which is a corresponding representation according to FIG 5 shows. Due to the effect of the superimposed voltage due to the superimposed voltage signal 86, the time periods shift, so that the switching times are now determined based on the time periods t5, t6, as shown by the switching signal for the phase voltage 108. This shifts the average electrical potential at the corresponding phase terminal 32.

[0076] The FIG 8 und 9 show according to the FIG 6 und 7 the conditions for the case where the target voltage signal has a current instantaneous potential between the electrical potentials 24 and 26. The superimposed voltage signal 86 leads to a negative superimposed voltage. Accordingly, the time periods shift according to FIG 9 . In the FIG 8 und 9 This is shown for phase W. Accordingly, FIG 9 the phase voltage 74 without the influence of the superimposed voltage signal 86, whereas FIG 9 shows the situation for the phase voltage 110 under the influence of the superimposed voltage signal 86.

[0077] Since the superimposed voltage signal 86 for the phases U, V, W is the same ( FIG 20 ), this has the effect that the setpoint voltage signals 64, 66, 68 are shifted by the same amount at the phase connections 32, 34, 36. This has no effect on the asynchronous machine 30, since only the interlinked voltages are relevant for the operation of the asynchronous machine 30.

[0078] In contrast, the superimposed voltage has the effect of reducing the load on the midpoint path of the three-level inverter 28. Thus, not only are the switching elements of the circuit arrangement 60 subjected to less load, but the load can be distributed among the switching elements of the other circuit arrangements 58, 62. This allows the switching elements 40 to 56 to be loaded more evenly overall, thereby increasing reliability and reducing aging. Furthermore, the overall load capacity of the three-level inverter 28 can be increased.

[0079] The FIG 10 und 11 show in schematic diagrams how the FIG 4 und 5 the conditions when a large superimposed voltage is superimposed on the respective setpoint voltage signals 64, 66, 68. The corresponding phase voltages 70, 72, 74 are shown FIG 11 . The FIG 12 und 13 show a corresponding situation where the superposition voltage is negative, that is, the electrical potentials of the setpoint voltage signals 64, 66, 68 are between the electrical potentials 24, 26. The corresponding phase voltages 70, 72, 74 are in FIG 13 shown.

[0080] With the FIG 14 und 15 an intermediate value of the superimposed voltage is shown, from which it can be seen that the superimposed voltage can also change continuously, whereby the load on the switching elements 40 to 56 can also be changed, preferably continuously. FIG 14 und 15 represent an intermediate value, as it relates to the scenarios according to the FIG 10 bis 13 has already been explained.

[0081] FIG 16 shows the target voltage signals 64, 66, 68 superimposed by a trapezoidal superimposed voltage signal 86, using corresponding graphs 94, 96, 98, showing how they can change over time due to the superimposed voltages. In this case, the superimposed voltages are formed by trapezoidal alternating voltages whose frequency is higher than the frequency of the phase voltages 64, 66, 68. The superimposed alternating voltage signal 86 is therefore, for example, trapezoidal.

[0082] FIG 17 shows a corresponding schematic diagram representation such as FIG 16 , in which the superimposed voltage signal 86 has an approximately rectangular signal shape, with the corners being rounded.

[0083] FIG 18 shows, in a further embodiment, the superposition voltage due to a superposition voltage signal 86, which has an approximately sinusoidal signal shape.

[0084] The frequency of the superimposed AC voltage signals need not be constant. It can, of course, be varied as needed, for example, depending on the design of the three-level inverter 28, a current load condition, and / or the like.

[0085] In the diagrams according to the FIG 16 bis 18 the abscissa is assigned to time, whereas the ordinate is assigned to the electrical potential.

[0086] FIG 19 shows in a further schematic diagram the time courses of the target voltage signals 64, 68, 66 as well as the courses of the respective target voltage signals superimposed by the superimposed voltage with correspondingly assigned graphs 88, 90, 92, as shown in FIG 16 It can be seen that all of the setpoint voltage signals 64, 66, 68 are potential-shifted according to the superimposed voltage signal 86, so that the corresponding circuit arrangements 58, 60, 62 are loaded more evenly.

[0087] FIG 20 shows in a voltage-time diagram the corresponding superimposed voltage signal 86, which has an amplitude 84.

[0088] The invention can achieve the following advantages: With a low frequency of the fundamental oscillation of the setpoint voltage signals, the permissible current of the switching elements can be increased; in electrical machines, the maximum torque is already available at a low speed; ripple current and thus heating and noise of the electrical machines can be reduced; common-mode currents and the resulting interference in the area of ​​electromagnetic compatibility can be reduced while maintaining the same parasitic capacitances.

[0089] The switching element, where the highest temperature occurs, usually limits the current carrying capacity of the entire inverter or multi-level energy converter. If all phase voltages are shifted by the same amount, the line-to-line voltages remain constant, so any effects on the electrical machine are negligible.

[0090] At the same time, however, the distribution of conduction losses among the various switching elements of the multi-level energy converter can be influenced. By shifting the losses away from the switching element subject to maximum load to other switching elements, its temperature can be reduced at the same current. The maximum permissible temperature of this switching element is therefore only reached at a higher current.

[0091] A particular advantage is that the shifting of the target voltage signals can be carried out continuously, which can also result in a continuous change in the loss distribution. This allows the losses or maximum temperature of the switching elements to be adjusted despite possible differences in the conduction and cooling properties of the switching elements. The ripple current can be reduced because no timing gaps are required. The frequency relevant for the ripple current, for example, remains twice the pulse frequency throughout the entire period of the fundamental oscillation.

[0092] Common-mode currents are generally proportional to the capacitance in the common-mode circuit and the rate of change of a common-mode voltage. This rate can be very large when using the timing gap, because it usually corresponds to the switching speed of the switching elements, which should switch as quickly as possible to achieve low switching losses. In the method described here according to the invention, however, the rate of change of the common-mode voltage is decoupled from the switching behavior of the switching elements and is freely adjustable.

[0093] Due to the freely selectable distribution of losses among the switching elements, a lower overall temperature of the switching elements and a more even load on the switching elements can be achieved. The lower thermal stress on individual switching elements can lead to greater reliability at a constant current. Conversely, a higher current can be achieved at the same permissible maximum temperature, thus enabling an increase in performance.

[0094] The first DC link 12 provides the first DC link voltage U1, and the second DC link 14 provides the second DC link voltage U1. In this case, the first and second DC link voltages U1, U2 are equal. Depending on the application, however, the first and second DC link voltages may also differ from each other.

[0095] FIG 21 A schematic diagram shows the effects that can be achieved with the operating method according to the invention. The abscissa represents the frequency, whereas the ordinate represents the electrical current.

[0096] With the FIG 21 The graphs 100, 102, 104, 106 shown show the maximum current carrying capacities of the switching elements 40 to 56 and thus also of the multi-level energy converter or the three-point inverter 28.

[0097] A graph 104 shows a derating for the intended operating case without using an overlay voltage, whereby temperature ripple is disregarded. Graph 104 shows that as the frequency decreases, the permissible current carrying capacity also decreases. Graph 106 illustrates the situation for the maximum permissible junction temperature of each of the switching elements 40 to 56 in conjunction with a temperature ripple permissible for the specified service life. The temperature ripple refers to a temperature swing during one period of the fundamental oscillation of the nominal voltage signals 64, 66, 68. The longer the overload times, the greater the derating required. This can be shown using graphs 104 and 106.

[0098] Graphs 100 and 102 now show the situation accordingly using a superimposed voltage. With regard to the operating conditions, graph 100 is assigned to graph 104, and graph 102 to graph 106. It can be seen that by using the superimposed voltage, the load capacity of the multi-level energy converter or three-level inverter 28 can be significantly increased, particularly in the range of frequencies lower than approximately 30 Hz. However, this can also vary depending on the design of the multi-level energy converter or three-level inverter 28, because the effect of the invention can also depend, among other things, on the actual design of the multi-level energy converter or three-level inverter 28.

[0099] The embodiments serve solely to explain the invention and are not intended to limit it.

Claims

1. Method for operating switching elements (40 to 56) of a multilevel energy converter (28) using at least three electric potentials (22, 24, 26), to which a multiphase electrical machine (30) is connected, wherein switching elements (40 to 56) of the multilevel energy converter (28) are operated in a predefined clock mode by means of corresponding switching signals (70, 72, 74), wherein the switching signals (70, 72, 74) for the switching elements (40 to 56) that are assigned to a respective one of the phases (U, V, W) are determined as a function of a respective desired voltage signal (64, 66, 68), in order to apply a respective phase voltage (70, 72, 74) to each phase of the multiphase electrical machine (30), wherein the switching signals are determined in addition as a function of a predefined overlaying voltage signal (86) in such a way that the respective phase voltages (70, 72, 74) are overlaid with an overlaying voltage dependent on the overlaying voltage signal (86), wherein an AC voltage signal is used as the overlaying voltage signal (86), wherein the AC voltage signal is formed in such a way that the stress is distributed cyclically to the electrical potentials (22, 24, 26) and to their switching elements (40 to 56), wherein by displacement of the respective phase voltages (70, 72, 74) as a result of the overlaying voltage, a stress on the switching elements (40 to 56) is distributed more evenly to all switching elements (40 to 56) of the multilevel energy converter (28), wherein the AC voltage signal is a sinusoidal or trapezoidal AC voltage signal.

2. Method according to claim 1, characterised in that the overlaying occurs when a frequency of the desired voltage signals (64, 66, 68) is less than a predetermined comparison frequency.

3. Method according to one of the preceding claims, characterised in that the overlaying occurs when an amplitude of the desired voltage signals (64, 66, 68) is less than a predetermined comparison amplitude.

4. Method according to one of the preceding claims, characterised in that temperatures of the switching elements (40 to 56) are determined and the overlaying voltage signal (86) is set as a function of the temperatures determined.

5. Method according to one of the preceding claims, characterised in that aging states of the switching elements (40 to 56) are determined and the overlaying voltage signal (86) is set as a function of the aging states determined.

6. Method according to one of the preceding claims, characterised in that an amplitude and / or a frequency of the AC voltage signal is set as a function of a thermal time constant of the switching elements (40 to 56).

7. Method according to one of the preceding claims, characterised in that the AC voltage signal is a trapezoidal AC voltage signal.

8. Control unit (38) for controlling switching elements (40 to 56) of a multilevel energy converter (28) using at least three electric potentials (22, 24, 26), which is embodied for connection to a multiphase electrical machine (30), wherein the control unit (38) is embodied to operate the switching elements (40 to 56) of the multilevel energy converter (28) by means of corresponding switching signals (70, 72, 74) in a predefined clock mode, for which purpose the control unit (38) determines the switching signals (70, 72, 74) for the switching elements (40 to 56) assigned to a respective one of the phases as a function of a respective desired voltage signal (64, 66, 68), in order to apply a respective phase voltage (70, 72, 74) to each phase (U, V, W) of the multiphase electrical machine (30), wherein the control unit (38) is further embodied to determine the switching signals (70, 72, 74) additionally as a function of a predefined overlaying voltage signal (86), in order to overlay the respective phase voltages (64, 66, 68) with an overlaying voltage dependent on the overlaying voltage signal (86), characterisedin that the control unit (38) is embodied to perform a method according to one of claims 1 to 7.

9. Multilevel energy converter (28), which is embodied for connection to a multiphase electrical machine (30) and which uses at least three electric potentials (22, 24, 26), with: - Switching elements (40 to 56) for applying a respective phase voltage (70, 72, 74) to each phase (U, V, W) of the multiphase electrical machine (30), and - A control unit (38) for controlling the switching elements (40 to 56) in a predefined clock mode by means of corresponding switching signals, characterised in that the control unit (38) is embodied in accordance with claim 8.

10. Drive facility (10) with: - A multiphase electrical machine (30), and - A multilevel energy converter (28) connected to the multiphase electrical machine (30), which uses at least three electric potentials (22, 24, 26), characterised in that the multilevel energy converter (28) is embodied in accordance with claim 9.