Method and apparatus for determining a switching pattern for the operation of an inverter

Abstract

A method (400) for operating an inverter 1 is described. The method (400) comprises determining (401) a modulation index (252) of an output voltage vector (Va) of the inverter (1), as well as determining (402) a switching pattern (320) with relative switching states (262) of switching devices (S1, S2, S3) of the switching sections (3a, 3b, 3c) of the inverter (1), based on the modulation index (252). Furthermore, the method (400) comprises determining (403) a sector (K2) from a plurality of different sectors (K1, ..., K6) of a space vector diagram (4) of the inverter (1) in which the output voltage vector (Va) is arranged, and converting (404) the switching pattern (320) with relative switching states (262) into a corresponding switching pattern (320) with absolute switching states (261), based on the determined sector (K2).The method (400) further comprises operating (405) the inverter (1) using the switching pattern (320) with absolute switching states (261).

Description

[0001] The invention relates to a method and a corresponding device for operating an inverter, in particular a multi-stage inverter.

[0002] An electrically powered vehicle can have one or more electric drive motors powered by electrical energy from an electrical energy storage device. The one or more drive motors can be used to drive one or more wheels of the vehicle. The phase voltages or phase currents for operating a drive motor can be generated using an inverter, in particular a multi-stage inverter.

[0003] The inverter typically comprises a plurality of switching sections for generating a corresponding plurality of phase voltages or phase currents. The individual switching sections each comprise a set of switching devices, wherein the individual switching devices can each be selectively closed to generate the respective phase voltage or phase current. Within each period, a full wave of the phase voltage or phase current is generated. The period thus extends over a specific angular range, in particular 0° to 360° or 0 to 2π.

[0004] To generate a phase voltage or phase current using a switching section, a specific pulse pattern with a specific number N of switching angles can be considered. Each individual switching angle indicates the angle within a period at which a switching device of the switching section is closed to generate a full wave of the phase voltage or phase current using the pulse pattern. Due to the symmetry of a full wave, it is typically sufficient to consider switching angles for a half-wave or a quarter-wave, which can then be used in mirror symmetry to generate the subsequent half-wave or quarter-waves.

[0005] The pulse pattern with switching angles for a period can be determined in advance using an optimization method, e.g., optimal PWM. A specific optimization criterion, such as the total harmonic distortion (THD) of the phase current, can be considered, and in particular optimized. For different operating states of the electrical (drive) machine, especially for different modulation indices, different pulse patterns with switching angles can be determined and stored in a memory unit. It is typically sufficient to store the pulse patterns for one phase, since the pulse patterns for the other phases are determined by a phase shift (typically by 120° each).

[0006] This document deals with the technical problem of storing the pulse patterns with switching angles for the operation of the switching sections of an inverter, in particular a multi-stage inverter, in a storage unit in a particularly advantageous manner in order to enable particularly efficient operation of the inverter.

[0007] The problem is solved by each of the independent claims. Advantageous embodiments are described, among other things, in the dependent claims. It should be noted that additional features of a patent claim dependent on an independent patent claim can form a separate invention, independent of the combination of all features of the independent patent claim, without the features of the independent patent claim or only in combination with a subset of the features of the independent patent claim, which invention can be made the subject of an independent claim, a divisional application, or a subsequent application. This applies equally to technical teachings described in the description, which can form an invention independent of the features of the independent patent claims.

[0008] According to one aspect, a method for operating a (possibly multi-stage) inverter is described. The inverter can have multiple switching sections, each with a phase tap for electrically connecting to a respective phase of an electrical (drive) machine. The individual switching sections can each have a set of switching devices.

[0009] The inverter can have an intermediate circuit with three connections in the form of a high-side connection, a low-side connection, and a center-point connection. Furthermore, the inverter can have at least three switching sections connected to the intermediate circuit, each with a phase tap for electrically connecting to a respective phase of the electric (drive) machine. Each switching section can have three switching devices in the form of a high-side switching device connected to the high-side connection, a low-side switching device connected to the low-side connection, and a center-point switching device connected to the center-point connection.

[0010] In order to approximate an operating point-specific output voltage (i.e. an output voltage vector or vector), a switching pattern for the switching devices can be determined, by means of which the switching devices are transferred into different switching states and in doing so connect the phase tap of the respective switching section to one of the terminals of the intermediate circuit, generating a voltage vector.

[0011] A multi-level inverter, particularly a 3-level inverter, can thus provide three voltage levels for the phases of the electrical machine per switching section, which consists of a half-bridge and a midpoint switching device. In addition to the eight switching states of a conventional 2-level inverter, it has 19 additional switching states for approximating an operating-point-specific output voltage. Using these 27 switching states, a desired sinusoidal output voltage with a particularly low harmonic content can be generated, thereby reducing losses in the electrical machine. Furthermore, a multi-level inverter typically has relatively low switching losses.

[0012] The inverter can be designed to convert the direct voltage provided by the electrical energy storage device, which is applied as input voltage to the intermediate circuit, into a multi-phase alternating voltage and to provide it as output voltage for the phases of the electrical machine.

[0013] The inverter can have the intermediate circuit on the input side, which in particular has at least one series connection of at least two capacitors. The switching sections or switching bridges can each have the high-side switching device with at least one high-side switch, the low-side switching device with at least one low-side switch, and the midpoint switching device with at least one midpoint switch. The high-side switching device and the low-side switching device of a switching section can be connected to form a half-bridge. Each switching section has a phase tap connected to the associated phase of the electrical machine. The terminals of the series connection of capacitors, which form the high-side terminal and the low-side terminal of the intermediate circuit, can be connected to poles of an electrical energy storage device and to the half-bridges of the switching sections.The center point connection, or center point for short, can be provided between the capacitors of the at least one series circuit and is connected to the center point switching devices of the switching sections. The half-bridges each form a vertical bridge arm of the respective switching section, and the center point switching devices form a horizontal bridge arm of the respective switching section. The high-side switches and the low-side switches can each be designed as IGBTs or power MOSFETs, for example. To provide bidirectional blocking capability, the center point switching devices can each have two anti-serially connected center point switches or one bidirectional center point switch each. The center point switches can also be designed as IGBTs or power MOSFETs.

[0014] Due to the three switching devices per switching bridge, a three-stage inverter with three switching bridges has 27 switching states, with each switching state generating a voltage vector that can be represented as a space vector in a (space) vector diagram. When the high-side switching device of a switching section is closed, the phase tap of this switching section is electrically connected to the high-side terminal, thus applying a high-side potential to the associated phase. When the low-side switching device of a switching section is closed, the phase tap of this switching section is electrically connected to the low-side terminal, thus applying a low-side potential to the associated phase. When the mid-point switching device of a switching section is closed, the phase tap of this switching section is electrically connected to the mid-point terminal, thus applying a mid-point potential to the associated phase.

[0015] For example, three switching states, in which either all high-side switches are closed and thus all phase taps and phases are connected to the high-side terminal of the DC link, or all low-side switches are closed and thus all phase taps and phases are connected to the low-side terminal of the DC link, or all midpoint switches are closed and thus all phase taps and phases are connected to the midpoint terminal of the DC link, generate zero-voltage vectors, which are represented as zero-voltage space vectors in the space vector diagram. The zero-voltage space vectors can thus be generated redundantly.

[0016] Six additional switching states, which do not involve the midpoint switching devices and thus the midpoint switches, generate six voltage vectors, which are represented as long, outer voltage space vectors in the space vector diagram and form the corner points of an outer hexagon. These six long, outer voltage space vectors correspond to the basic voltage space vectors of a 2-level inverter and can also be referred to as basic voltage space vectors. Six additional switching states involve the midpoint switching devices and generate six additional voltage vectors, which are represented as middle, outer voltage space vectors in the space vector diagram and are located on the side edges of the outer hexagon between two long, outer voltage space vectors each.

[0017] Twelve additional switching states, which also involve the center-point switching devices, generate twelve short, internal voltage space vectors, with each two switching states providing two identical voltage space vectors, i.e. voltage space vectors with the same length and the same phase angle. This means that each of the short, internal voltage space vectors can be generated redundantly. The short, internal voltage space vectors span an inner hexagon, with two redundant voltage space vectors forming a corner point of the inner hexagon. The mutually redundant voltage space vectors have an opposite, complementary effect on the current in the center-point connection of the intermediate circuit. The two mutually redundant voltage space vectors are differentiated in this document by the letters "n" (for negative current in the center-point connection) and "p" (for positive current in the center-point connection).

[0018] The outer hexagon of the space vector diagram can be divided into K=6 triangular sectors, with the vertices of each sector being formed by the zero-voltage space vectors and two adjacent long, outer voltage space vectors. Each sector can, in turn, be divided into four triangular segments, with two redundant voltage space vectors located at at least one vertex of each segment.

[0019] The method described in this document involves determining a modulation index of an (approximate) output voltage vector of the inverter. The modulation index can correspond to the magnitude or length of the output voltage vector.

[0020] The output voltage vector can be determined as part of a control and / or regulation of the electrical machine, e.g., using a control loop. In particular, output voltage vectors can be determined at a specific repetition rate to ensure that the electrical machine exhibits a specific target behavior, e.g., a specific target torque. The output voltage vectors can be determined in an α / β coordinate system or in a d / q coordinate system.

[0021] The method further comprises determining a switching pattern with relative switching states of the switching devices of the switching sections based on the modulation index. The switching pattern with relative switching states can be determined using a look-up table (LUT), which specifies a corresponding switching pattern with relative switching states for a plurality of different values ​​of the modulation index.

[0022] The LUT can be determined in advance using an OPP (optimum pulse pattern) method. Pulse patterns with switching angles can be determined for each individual phase, which can then be combined into a switching pattern with switching states for the switching sections of all phases. Furthermore, the switching states can be transformed into relative switching states. Using the OPP method, for example, absolute switching angles (for the individual phases) can be determined. These can then be converted into relative angles and / or switching states to provide the relative switching states for the LUT.

[0023] The individual sectors of the plurality of different sectors of the space vector diagram can each extend over an angular range of 2π / K, with K>1, in particular K≥6 (where K corresponds to the number of sectors of the space vector diagram). The look-up table can be limited to switching patterns with relative switching states for (exactly) an angular range of 2π / K. Thus, a relatively compact LUT can be provided. If necessary, two different variants of the look-up table can be provided: a first variant for sectors with an even-numbered sector index and a second variant for sectors with an odd-numbered sector index.

[0024] The individual sectors of the plurality of different sectors of the space vector diagram can each have a set of different reference points. The individual reference points can each correspond to a corner point of a segment of a sector. The set of reference points can be designed such that the set of reference points of any sector can be converted into the set of reference points of any other sector simply by rotation around the origin point of the space vector diagram (which typically corresponds to the zero-voltage vector). The possible rotations can be limited to a rotation by the angle 2π / K or a multiple thereof.

[0025] The individual possible relative switching states can each correspond to a reference point from the set of reference points, in particular to a corner point of a segment of the space vector diagram. The reference points and / or the possible relative switching states can each be identified by an index, wherein the index is independent of the different sectors and / or wherein the index is the same for all sectors.

[0026] Thus, a switching pattern of relative switching states can be determined that is independent of the sector in which the output voltage vector to be approximated is located.

[0027] The method further comprises determining the sector from the plurality of different sectors of the inverter's space vector diagram in which the output voltage vector is located. Furthermore, the method comprises converting the determined switching pattern with relative switching states into a corresponding switching pattern with absolute switching states based on the determined sector. The switching pattern with relative switching states can be converted into the corresponding switching pattern with absolute switching states using a (predetermined and / or determined) assignment table.

[0028] The assignment table can specify, for each individual sector of the plurality of different sectors of the space vector diagram, the corresponding one or more absolute switching states from the set of possible relative switching states. Different one or more absolute switching states can be assigned to a possible relative switching state in each sector.

[0029] The method makes it possible to efficiently determine a switching pattern of absolute switching states for the operation of the inverter. The inverter can then be operated using the switching pattern with absolute switching states.

[0030] A switching pattern can specify a sequence of switching states for a quarter, a half, or a whole period. If necessary, the switching states can be specified for a quarter or half period, and the switching states for the other quarters or half of the period can be derived based on symmetry. A period can correspond exactly to one revolution of a phasor in the space phasor diagram.

[0031] Furthermore, the switching pattern with relative switching states and the corresponding switching pattern with absolute switching states can each specify switching angles. The individual switching angles can each specify the angle within the space vector diagram for the transition from one switching state to the directly subsequent switching state. The conversion of the determined switching pattern with relative switching states into the corresponding switching pattern with absolute switching states can be carried out in such a way that the (in particular, all) switching angles of the switching pattern with relative switching states and the corresponding switching angles of the corresponding switching pattern with absolute switching states are identical in pairs.

[0032] Alternatively, in the switching pattern with relative switching states, the individual switching angles can be specified relative to the sector in which the transition between directly consecutive switching states occurs. The switching angles in the switching pattern with relative switching states can be limited to the angular range between 0 and 2π / K.

[0033] In particular, in the above-mentioned case, the switching angles of the switching pattern with relative switching states and the corresponding switching angles of the corresponding switching pattern with absolute switching states can differ from each other in pairs by exactly k*2π / K, with k=0,...,5.

[0034] The process can be repeated for Q consecutive sections of a period, with Q>1, e.g., Q=6 or Q=12, where one period corresponds to an angular range of 2π or 360° of the inverter's space vector diagram. Q is preferably identical to or an integer multiple of the number K of sectors.

[0035] Thus, an output voltage vector and a corresponding switching pattern with absolute switching states can be determined Q times per period. The switching pattern determined in each section can indicate relative switching states or absolute switching states for a portion of a period, e.g., for an angular range of π or 180° of the inverter's space vector diagram (e.g., for a quarter period or a half period).

[0036] The inverter can be operated in a current section using the switching pattern with absolute switching states determined in the respective, particularly immediately preceding, section. Repeatedly updating the switching pattern enables consistently precise and efficient operation of the inverter.

[0037] The assignment table can specify several variants of the corresponding absolute switching state for at least one possible switching state (e.g. a p-variant and an n-variant of the absolute switching state).

[0038] Within the scope of the method, state information relating to the operating state of the inverter, in particular relating to the current and / or the voltage at the center of the inverter, can be determined. Depending on the state information, at least one or exactly one variant of the absolute switching state can then be selected for the switching pattern with absolute switching states. This can ensure particularly efficient and reliable operation of the inverter.

[0039] As already explained above, the switching pattern with relative switching states comprises a sequence of relative switching states, e.g., relative switching states for a portion of a period, such as a quarter period or a half period. The switching pattern can be limited to the portion of the period. The portion of the period can be so large that the output voltage vector traverses several sectors of the space vector diagram (e.g., two or three sectors) in that portion of the period.

[0040] The method comprises determining the sector in which a specific relative switching state from the sequence of relative switching states of the switching pattern is arranged. This can be determined, for example, based on the switching angle of the switching pattern. As already explained, the switching pattern with relative switching states can specify a switching angle within the space vector diagram for the change from one relative switching state to the directly following relative switching state for two directly consecutive relative switching states. The sector in which the specific relative switching state is arranged can be determined based on the switching angle for the change to the specific relative switching state, in particular based on all switching angles of the switching pattern arranged before the specific relative switching state.

[0041] The method may further comprise converting the determined relative switching state into the corresponding absolute switching state using the assignment table for the determined sector. In other words, the assignment table for the sector in which a relative switching state is located can always be determined. Thus, the switching pattern with absolute switching states can be determined particularly precisely.

[0042] The switching angle for changing to a specific relative switching state can always be specified relative to the sector in which the specific relative switching state is located. The individual switching angles can thus be limited to a range between 0 and 2π / K. This allows for particularly efficient storage of switching patterns with relative switching states.

[0043] According to another aspect, a software (SW) program is described. The SW program can be configured to be executed on a processor (e.g., on a vehicle control unit) and thereby to carry out the method described in this document.

[0044] According to a further aspect, a storage medium is described. The storage medium can comprise a software program configured to be executed on a processor and thereby to carry out the method described in this document.

[0045] According to a further aspect, a device for operating an inverter, in particular a multi-stage inverter, is described. The inverter comprises a plurality of switching sections, each with a phase tap for electrically connecting to a respective phase of an electrical machine, wherein the individual switching sections each have a plurality of switching devices.

[0046] The device is configured to determine a modulation index of an output voltage vector of the inverter and, on the basis of the modulation index, to determine a switching pattern with relative switching states of the switching devices of the switching sections.

[0047] The device is further configured to determine a sector from a plurality of different sectors of the space vector diagram of the inverter in which the output voltage vector is arranged, and, based on the determined sector, to convert the switching pattern with relative switching states into a corresponding switching pattern with absolute switching states. Each individual relative switching state can be replaced by a corresponding absolute switching state for the determined sector. The inverter can be operated using the switching pattern with absolute switching states.

[0048] According to a further aspect, a possibly multi-stage inverter is described which comprises the device described in this document.

[0049] According to a further aspect, a (road) motor vehicle (in particular a passenger car or a truck or a bus or a motorcycle) is described which comprises the device described in this document and / or the inverter described in this document.

[0050] According to a further aspect, a look-up table is described that specifies a switching pattern with relative switching states for a plurality of different modulation indices. The look-up table can be limited to switching patterns with relative switching states for (exactly) an angular range of 2π / K. Optionally, the look-up table can have a first variant for sectors with an even-numbered index and a second variant for sectors with an odd-numbered index.

[0051] It should be noted that the methods, devices, and systems described in this document can be used both alone and in combination with other methods, devices, and systems described in this document. Furthermore, any aspects of the methods, devices, and systems described in this document can be combined in a variety of ways. In particular, the features of the claims can be combined in a variety of ways. Furthermore, features listed in parentheses are to be understood as optional features.

[0052] The invention will be described in more detail below using exemplary embodiments. Fig. 1a exemplary components of a vehicle with an electric drive motor; Fig. 1b shows a circuit diagram of an exemplary embodiment of a 3-level inverter or a three-stage inverter; Fig. 2a a vector diagram of the 3-level inverter; Fig. 2b an example switching pattern; Fig. 2c relative switching states in the first sector of the vector diagram; Fig. 3a an exemplary look-up table with switching patterns with relative switching states; Fig. 3b a mapping table between relative switching states and absolute switching states for different sectors; and Fig. 4 a flowchart of an exemplary method for determining a switching pattern for the operation of an inverter.

[0053] In the figures, identical and functionally identical elements are provided with the same reference numerals.

[0054] As stated at the beginning, this document deals with the efficient storage of switching angles for the operation of an inverter, in particular a multi-level inverter, where the inverter can be operated, for example, to supply an electric drive motor of a vehicle. In this context, Fig. 1a shows exemplary components of a vehicle 140 having an electric machine 133 for driving the vehicle 140. The electric machine 133 is coupled to one or more wheels 141 of the vehicle 140 in order to drive the one or more wheels 141 and thus the vehicle 140. The electric machine 133 is operated with electrical energy from an electrical, in particular an electrochemical, energy storage device 130. The energy storage device 130 can be configured to provide a direct current with a specific direct voltage.

[0055] The vehicle 140 has an inverter 1 configured to generate phase voltages and phase currents for the different phases of the electric machine 133 based on the direct current from the energy storage device 130. The inverter 1 can be operated by a (control) device 131.

[0056] Fig. Figure 1b shows a 3-level inverter 1 for an electrified motor vehicle 140, as an example of an inverter, in particular a multi-level inverter. The 3-level inverter 1 has an intermediate circuit 2 on the input side, which here has three intermediate circuit branches 2a, 2b, 2c, each with a series connection of two capacitors C1, C2. A high-side connection DC+ and a low-side connection DC- of the intermediate circuit 2 are connected to poles of the electrical energy storage device 130 of the motor vehicle 140, which supplies the intermediate circuit 2 with a DC voltage V DCEach capacitor C1, C2 has half the DC voltage V DC / 2. The 3-level inverter 1 also has three switching sections 3a, 3b, 3c, which are each electrically connected to one of the intermediate circuit branches 2a, 2b, 2c. However, it can also be provided that the intermediate circuit 2 has only one intermediate circuit branch with two series-connected capacitors C1, C2, to which the switching sections 3a, 3b, 3c are connected. Each switching section 3a, 3b, 3c has a phase tap u, v, w, which is connected to a phase Phu, Phv, Phw of the electric drive motor 133 of the motor vehicle 140.

[0057] Each switching section 3a, 3b, 3c has three switching devices S1, S2, S3, each of which can have at least one switch embodied as a semiconductor switch, for example an IGBT. The switching devices S1 and S2 are interconnected to form a half-bridge HB of the respective switching section 3a, 3b, 3c. The first switching devices S1 are high-side switching devices and are connected to the high-side terminal DC+ of the intermediate circuit 2, which is connected to a positive pole of the electrical energy storage device 130, and to a tap u, v, w of the respective half-bridge HB. The second switching devices S2 are low-side switching devices and are connected to the tap u, v, w of the respective half-bridge HB and to the low-side terminal DC- of the intermediate circuit 2, which is connected to a negative pole of the electrical energy storage device 130.The third switching devices S3 are center-point switching devices and are connected to the taps u, v, w and to a center-point terminal N of the respective intermediate circuit branch 2a, 2b, 2c between the two capacitors C1, C2. The center-point switching devices S3 each comprise two anti-serially connected center-point switches S3a, S3b.

[0058] The 3-level inverter 1 can assume 27 switching states, by means of which a desired, operating point-specific output voltage can be modulated for application to the phases Phu, Phv, Phw. Each switching state generates a voltage vector, which is in the Fig. 2 with the axes α, β. In the space vector diagram 4, the different switching states (u, v, w) of the corresponding phase taps are plotted at discrete points, with u, v, w = 0, 1, 2, where the switching state u indicates which of the switching devices S1, S2, S3 in the first switching section 3a is closed, the switching state v indicates which of the switching devices S1, S2, S3 in the second switching section 3b is closed, and the switching state w indicates which of the switching devices S1, S2, S3 in the third switching section 3c is closed. The letters “u, v, w” can thus, depending on the context, indicate a specific phase tap or the switching state of the switching section 3a, 3b, 3c for the corresponding phase tap.

[0059] The state value "0" means that the respective low-side switching device S2 is closed (and S1 and S3 are open) and thus connects the respective phase tap u, v, w to the low-side connection DC-. The state value "1" means that the respective mid-point switching device S3 is closed (and S1 and S2 are open) and thus connects the respective phase tap u, v, w to the mid-point connection N. The state value "2" means that the respective high-side switching device S1 is closed (and S2 and S3 are open) and thus connects the respective phase tap u, v, w to the high-side connection DC+.

[0060] The discrete points in the space vector diagram 4 mark an arrowhead of a voltage space vector generated by the respective switching state, which has a specific length and a specific phase angle. It is evident from the space vector diagram 4 that switching states exist that generate identical voltage vectors. For example, the switching states "221" and "110" generate identical voltage vectors v2p and v2n. The switching states "000", "111", and "222" also generate identical voltage vectors, namely a zero-voltage vector v01, v02, and v03, respectively.

[0061] The space vector diagram 4 is divided into six sectors K1, K2, K3, K4, K5, K6, where each sector K1, ..., K6 is divided into four segments Q1, Q2, Q3, Q4 (shown here only for the first sector K1). If a specific, operating point-specific output voltage, for example the one shown in Fig. If the output voltage vector Va shown in Figure 2a is to be modulated, it is first determined in which sector K1, ..., K6 and in which segment Q1, ..., Q4 of the respective sector K1, ..., K6 the respective output voltage vector Va is located. For example, the output voltage vector Va is located in the first (inner) segment Q1 of the first sector K1 and can thus be modulated or approximated by those voltage vectors that form the corner points of the first (inner) segment Q1 of the first sector K1 and thus surround the output voltage vector Va to be generated. In the example shown, these are the following voltage vectors: • v01, generated by switching state “000”, • v02, generated by switching state “111”, • v03, generated by switching state “222”, • v2p, generated by switching state “221”, • v2n, generated by switching state “110”, • v1p, generated by switching state “211”, and • v1n, generated by switching state “100”.

[0062] In the case of relatively low drive levels of the 3-level inverter 1, especially when the drive level falls below a predetermined threshold, for example 0.6, the internal short voltage vectors v1p, v1n, v2p, v2n and the zero voltage vectors v01, v02, v03 can be set alternately within a switching pattern in order to achieve a uniform load on the switching devices. For example, the Fig. The switching pattern SM shown in Figure 2b can be specified.

[0063] One control method for inverters or power inverters 1 is the optimized pulse pattern method. In this method, switching angles for a phase are determined offline according to specific evaluation and / or optimization criteria and stored in one or more lookup tables (LUTs). The switch-on angles of the other phases can be derived from these (by phase shifting). Typically, the switching state of all three phases of power inverter 1 cannot be directly derived from the switch-on angles for the individual phases.

[0064] In an offline process, N switching angles (e.g. N≥2 or N≥3) can be determined for a phase in the angular range 0 to π (i.e. for one half-wave) in such a way that, for example, the THD of the phase current is reduced, in particular minimized (if necessary while simultaneously maintaining the desired modulation level). This results in one or more lookup tables which, depending on the modulation level (i.e. the modulation index), specify the switching angles of a phase in the angular range 0 to π. The switching angles for the angular range from π to 2π (i.e. for the second half-wave) result from the symmetry. The switching angles for the other phases can be determined by ensuring that the individual phases are 120° out of phase with one another.

[0065] In an angle-synchronous control method, during operation of the inverter 1 with a specific repetition rate, e.g. every 30°, it can be determined on the basis of the one or more stored LUTs which one or more phases are to be switched in the respective subsequent 30° section with which switching angle.

[0066] It can be particularly advantageous in a multi-stage inverter 1 if the overall state of all three phases of inverter 1, i.e., the switching state of inverter 1, is known. This makes it possible, for example, to determine which current flows into the center point N of inverter 1 and for how long, thereby changing the center point voltage.

[0067] As already explained above, the space vector diagram 4 of the inverter 1 can be divided into several different sectors 253 (K1 to K6). The individual sectors 253 can be identified by a sector index i=1, ..., 6, and each have a set of switching states of the inverter 1, with the individual switching states each corresponding to a corner point of a segment Q1 to Q4 of the respective sector 253. Individual corner points can possibly have several different switching states (each resulting in the same phasor in the space vector diagram 4).

[0068] The corner points of the first sector K1 can be, as shown in the example Fig. 2c, can be used to define relative switching states 262. The relative switching states 262 are shown in Fig. 2c identified by the numbers 0 to 5 (possibly with a p or n variant).

[0069] The sectors K2 to K6 following the first sector K1 are each derived from the first sector K1 by a rotation around the origin of the space vector diagram 4 with a specific rotation angle. The rotation angle corresponds to 60° for the second sector K2, 120° for the third sector K3, etc. All sectors K1 to K6 each have the same relative switching states 262 (which can be identified by the state index z=0, ..., 5). On the other hand, the individual relative switching states 262 in the individual sectors 262 are each associated with different absolute switching states 261. The assignment between relative switching states 262 and absolute switching states 261 in the individual sectors K1 to K6 is shown in Table 1 as an example for a three-level inverter 1. Table 1 Relativer Schaltzustand K1 K2 K3 K4 K5 K6 0 000, 111, 222 000, 111, 222 000, 111, 222 000, 111, 222 000, 111, 222 000, 111, 222 1 211, 100 221, 110 121, 010 122, 011 112, 001 212, 101 2 221, 110 121, 010 122, 011 112, 001 212, 101 211, 100 3 200 220 020 022 002 202 4 220 020 022 002 202 200 5 210 120 021 012 102 201

[0070] Within the scope of an offline optimization method, optimized switching patterns SM can be determined, by means of which possible output voltage vectors Va arranged in the first sector 253 can be approximated. Different optimized switching patterns SM can be determined for different vector angles 251 and / or for different vector lengths 252 (ie, for different modulation indices).

[0071] The optimized switching patterns SM can each comprise a sequence of switching states (e.g., for a half or a full period, i.e., for an angular range from 0 to π or from 0 to 2π). The relative switching states 262 can be specified. These optimized switching patterns can be stored in a look-up table (LUT).

[0072] Fig. 3a shows an exemplary LUT 310 that specifies an optimized (relative) switching pattern 320 for a plurality of different values ​​of the modulation index 252. A switching pattern 320 comprises a sequence of relative switching states 262 and switching angles 322 at which a change from a specific relative switching state 262 to a directly following relative switching state 262 occurs. In Fig. Figure 3a shows an example of a switching pattern 320 with four consecutive relative switching states 262 (each having a different hatching). Furthermore, the switching angle 322 for the transition from the first to the second switching state 262 is displayed.

[0073] A plurality of different LUTs 310 can be provided for a corresponding plurality of different values ​​of the voltage angle 251, resulting in an overall LUT 300.

[0074] In addition to the one or more LUTs 300, 310, an assignment table 340 may be provided (see Fig. 3b), which indicates for the different sectors K1, ..., K6 the relationship between the individual relative switching states 262 and the corresponding absolute switching states 261 in the respective sector (as shown by way of example in Table 1 for a three-stage inverter 1).

[0075] During operation of the inverter 1, the LUTs 300, 310 and the assignment table 340 can be accessed to repeatedly determine the optimized switching pattern SM with a sequence of absolute switching patterns 361. The period duration of a period results from the speed of the electric machine 133 and / or a specific offset value of the voltage angle 251 results from the load of the electric machine 133. The latter makes it possible to select a specific LUT 310 for a specific value of the voltage angle 251.

[0076] The sector K1, ...., K6 in which the output voltage vector Va is located can be determined at a specific repetition rate during a period (approximately Q=12 times per period), e.g., for each 30° section (generally for each angle section). In particular, the sector index i=1, ..., 6 can be determined. Furthermore, the modulation index 252 of the output voltage vector Va can be determined. The angular position of the output voltage vector Va within the respective sector K1 to K6 results from the offset value of the voltage angle 251 and the respective angle section.

[0077] Based on the value of the voltage angle 251 and the value of the modulation index 252, an optimized switching pattern 320 can be determined using the LUT 300, 310, wherein the switching pattern 320 specifies a sequence of relative switching states 262. The relative switching states 262 can be converted into corresponding absolute switching states 261 based on the sector K1, ..., K6 in which the output voltage vector Va is located, in particular based on the sector index i=1, ..., 6, using the assignment table 340. For one or more relative switching states 262, several possible variants of different absolute switching states 261 may be available, which can be selected to further optimize the operation of the inverter 1.

[0078] Thus, relative switching states 262 (e.g., 0, 1p, 1n, 2p, 2n, 3, 4, 5, or 0, 1, 2, 3, 4, 5) can be defined within the individual sectors K1, ..., K6. In an offline process, the switching angles of the optimized pulse patterns of all phases can be correlated to define corresponding switching patterns 320. For each modulation index 252, an overall vector (i.e., a switching pattern 320) is thus obtained, which includes all switching operations of all switching sections 2 of the inverter 1.

[0079] If the individual absolute switching states 261 in the individual sectors are assigned the corresponding relative switching state 262, it can be seen that the relative switching states 262 resulting in the individual sectors are the same. Thus, it is sufficient to store the switching patterns 320 (with the switching angles 322) for the range 0 to 2π / 6 (i.e., for the first sector K1, 253). The individual (relative) switching states 262 are typically the same over relatively wide ranges of the modulation index 252 and therefore only need to be stored for relatively few sections.

[0080] During operation, the sector K1, ..., K6 in which the output voltage vector Va to be approximated is located can be determined in order to convert the relative switching states 252 into corresponding absolute switching states 251. This repeatedly results (e.g., for 30° sections) in a set 320 with switching states 261, 262 and the respective associated angles 322 at which the individual switching states 261, 262 become active.

[0081] Due to the fact that only the switching patterns 320 are stored for exactly one of the six sectors (or possibly for two different sectors, in particular for a sector with an even-numbered index and for a sector with an odd-numbered index), a storage-efficient operation of the inverter 1 can be achieved. Furthermore, from the knowledge of the individual switching states 261, it can be deduced which midpoint currents will flow in the individual 30° sections. This knowledge can be used as the basis for a balancing method.

[0082] During operation of the electric machine 133, a PLL (phase-locked loop) circuit can be activated from a certain speed of the electric machine 133, which locks the control frequency to the rotational frequency of the electric machine 133 and to the voltage angle of the electric machine 133. Once synchronization is achieved, a switch can be made from a space vector modulation method to an OPP (optimal pulse pattern) method for operating the inverter 1, in which optimized switching patterns 320 determined in advance are used. In this case, the switching states 261 and the associated switching angles 322 can be determined in successive 30° sections (depending on the respective modulation index 252), which must be set in the next (calculation) period. The sequence of the absolute switching states 261 and the switching angle 322 can be determined as described in this document.

[0083] Fig. 4 shows a flowchart of a (possibly computer-implemented) method 400 for operating an inverter 1, in particular a multi-stage inverter. The inverter 1 comprises a plurality of switching sections 3a, 3b, 3c, each with a phase tap u, v, w for electrically connecting to a respective phase Phu, Phv, Phw of an electrical machine 133. The individual switching sections 3a, 3b, 3c each have a set of switching devices S1, S2, S3.

[0084] The method 400 may be repeated at a certain repetition rate, e.g., Q times per period (e.g., Q≥12).

[0085] The method 400 comprises determining 401 a modulation index 252 of an output voltage vector Va of the inverter 1. For example, using a control loop for the operation of the electrical machine 133, in particular for the control of one or more phase currents of the electrical machine 133, an output voltage vector Va can be determined as a manipulated variable. The output voltage vector Va can have a specific voltage angle 251 relative to the α-axis of the space vector diagram 4. Furthermore, the output voltage vector Va can have a specific magnitude or a specific length within the space vector diagram 4, which can correspond to the modulation index 252.

[0086] The method 400 comprises determining 402 a switching pattern 320 with relative switching states 262 of the switching devices S1, S2, S3 of the switching sections 3a, 3b, 3c based on the modulation index 252. For this purpose, a (predetermined) LUT 300, 310 can be used, which specifies the corresponding switching pattern 320 with relative switching states 262 for a plurality of different possible modulation indices (and for a plurality of different voltage angles). The switching patterns 320 can be determined based on an OPP method and stored in the LUT 300, 310. The LUT 300, 310 preferably covers only an angular range of 2π / K, where K is the number of different sectors K1, ..., K6 of the space vector diagram 4 of the inverter 1.

[0087] The individual possible relative switching states 262 can each correspond to a reference point within a sector K1, ..., K6, wherein the reference points can each be corner points of the individual segments Q1, ..., Q4 of the sector K1, ..., K6.

[0088] The method 400 further includes determining 403 the sector K2 from the plurality of different sectors K1, ..., K6 of the space vector diagram 4 of the inverter 1 in which the output voltage vector Va is arranged. The sector K2 can be determined based on the voltage angle 251 of the output voltage vector Va.

[0089] Furthermore, the method 400 includes converting 404 the switching pattern 320 with relative switching states 262 into a corresponding switching pattern 320 with absolute switching states 261 based on the determined sector K2. For this purpose, an assignment table 340 can be used, which, for each sector, specifies an assignment between the individual possible relative switching states 262 and one or more absolute switching states 261 in the respective sector.

[0090] The inverter 1 can be operated using the switching pattern 320 with absolute switching states 261 (steps 405).

[0091] The present invention is not limited to the embodiments shown. In particular, it should be noted that the description and figures are intended only to illustrate the principle of the proposed methods, devices, and systems by way of example.

Claims

[1] Method (400) for operating an inverter (1); wherein the inverter (1) has a plurality of switching sections (3a, 3b, 3c), each having a phase tap (u, v, w) for electrically connecting to a respective phase (Phu Phv, Phw) of an electrical machine (133); wherein the individual switching sections (3a, 3b, 3c) each have a set of switching devices (S1, S2, S3); wherein the method (400) comprises, - determining (401) a modulation index (252) of an output voltage vector (Va) of the inverter (1); - determining (402) a switching pattern (320) with relative switching states (262) of the switching devices (S1, S2, S3) of the switching sections (3a, 3b, 3c) on the basis of the modulation index (252); - determining (403) a sector (K2) from a plurality of different sectors (K1, ..., K6) of a space vector diagram (4) of the inverter (1) in which the output voltage vector (Va) is arranged; and - converting (404) the switching pattern (320) with relative switching states (262) into a corresponding switching pattern (320) with absolute switching states (261), based on the determined sector (K2); and - Operating (405) the inverter (1) using the switching pattern (320) with absolute switching states (261). [2] Method (400) according to claim 1, wherein the switching pattern (320) with relative switching states (262) is determined using at least one look-up table (300, 310) which specifies a corresponding switching pattern (320) with relative switching states (262) for a plurality of different values ​​of the modulation index (252). [3] Method (400) according to claim 2, wherein - the individual sectors (K1, ..., K6) of the plurality of different sectors (K1, ..., K6) of the space vector diagram (4) each extend over an angular range of 2π / K; - K>1, in particular K≥6, corresponds to a number of sectors (K1, ..., K6); and - the look-up table (300, 310) is limited to switching patterns (320) with relative switching states (262) for an angular range of 2π / K. [4] Method (400) according to one of the preceding claims, wherein - the switching pattern (320) with relative switching states (262) is converted into the corresponding switching pattern (320) with absolute switching states (261) using an assignment table (340); and - the assignment table for the individual sectors (K1, ..., K6) of the plurality of different sectors (K1, ..., K6) of the space vector diagram (4) indicates the corresponding one or more absolute switching states (261) for each individual possible relative switching state (262) of a set of possible relative switching states (262). [5] Method (400) according to claim 4, wherein - the assignment table specifies several variants of a corresponding absolute switching state (261) for at least one possible switching state (262); and - the method (400) comprises - Determining status information relating to an operating state of the inverter (1), in particular relating to a current and / or to a voltage at a center point (N) of the inverter (1); - selecting at least one variant of the absolute switching state (261) for the switching pattern (320) with absolute switching states (261) depending on the state information. [6] Method (400) according to one of claims 4 to 5, wherein - the switching pattern (320) with relative switching states (262) comprises a sequence of relative switching states (262); - the switching pattern (320) with relative switching states (262) has relative switching states (262) for a part of a period; - the output voltage vector (Va) completes exactly one revolution within the space vector diagram (4) in the period; - the part of the period is so large that the output voltage vector (Va) passes through several sectors (K1, K2) of the space vector diagram (4) in that part of the period; and - the method (400) comprises - determining the sector (K2) in which a specific relative switching state (262) from the sequence of relative switching states (262) is arranged; and - converting the determined relative switching state (262) into a corresponding absolute switching state (261) using the assignment table (340) for the determined sector (K2). [7] Method (400) according to claim 6, wherein - the switching pattern (320) with relative switching states (262) for two directly consecutive relative switching states (262) indicates a switching angle (322) within the space vector diagram (4) for the change from one relative switching state (262) to the directly following relative switching state (262); and - the sector (K2) in which the determined relative switching state (262) is arranged is determined on the basis of the switching angle (322) for the change to the determined relative switching state (262), in particular on the basis of all switching angles (322) of the switching pattern (320) arranged before the determined relative switching state (262). [8] Method (400) according to claim 7, wherein the switching angle (322) for changing to the determined relative switching state (262) is specified relative to the sector (K2) in which the determined relative switching state (262) is arranged. [9] The method (400) of any one of claims 6 to 8, wherein the portion of the period is a quarter period or a half period. [10] Method (400) according to one of the preceding claims, wherein - the individual sectors (K1, ..., K6) of the plurality of different sectors (K1, ..., K6) of the space vector diagram (4) each have a set of different reference points; - the set of reference points of any sector (K1) can be converted into the set of reference points of any other sector (K1) solely by rotation around an origin point of the space vector diagram (4); - the individual reference points each correspond in particular to a corner point of a segment (Q1, ..., Q4) of a sector (K1, ..., K6); and - a relative switching state (262) corresponds to a reference point from a set of reference points. [11] Method (400) according to one of the preceding claims, wherein - the switching pattern (320) with relative switching states (262) and the corresponding switching pattern (320) with absolute switching states (261) each specify switching angles (322); - a switching angle (322) indicates an angle within the space vector diagram (4) for a change from one switching state (261, 262) to a directly subsequent switching state (261, 262); and - the switching angles (322) of the switching pattern (320) with relative switching states (262) and the corresponding switching angles (322) of the corresponding switching pattern (320) with absolute switching states (261) are identical in pairs; or - the switching angles (322) of the switching pattern (320) with relative switching states (262) and the corresponding switching angles (322) of the corresponding switching pattern (320) with absolute switching states (261) each differ from each other in pairs by exactly k*2π / K, with k=0,...,5; where K is a number of different sectors (K1, ..., K6) of the space vector diagram (4). [12] Method (400) according to one of the preceding claims, wherein - the method (400) is repeated for Q consecutive sections of a period, with Q>1, e.g. Q=6 or Q=12; - one period corresponds to an angular range of 2π or 360° of the space vector diagram (4) of the inverter (1); - the switching pattern (320) determined in the respective section indicates relative switching states (262) or absolute switching states (261) for a part of the period; and - the inverter (1) is operated in a current section using the switching pattern (320) with absolute switching states (261) which was determined in a, in particular directly, previous section. [13] Method (400) according to one of the preceding claims, wherein - the inverter (1) is a multi-level inverter (1); and - the plurality of different sectors (K1, ..., K6) comprise K≥6 different sectors (K1, ..., K6), each covering a different angular range of 2π / K of the space vector diagram (4) of the inverter (1). [14] Device (131) for operating an inverter (1); wherein the inverter (1) has a plurality of switching sections (3a, 3b, 3c), each having a phase tap (u, v, w) for electrically connecting to a respective phase (Phu Phv, Phw) of an electrical machine (133); wherein the individual switching sections (3a, 3b, 3c) each have a set of switching devices (S1, S2, S3); wherein the device (131) is configured, - to determine a modulation index (252) of an output voltage vector (Va) of the inverter (1); - to determine a switching pattern (320) with relative switching states (262) of the switching devices (S1, S2, S3) of the switching sections (3a, 3b, 3c) on the basis of the modulation index (252); - to determine a sector (K2) from a plurality of different sectors (K1, ..., K6) of a space vector diagram (4) of the inverter (1) in which the output voltage vector (Va) is arranged; and - on the basis of the determined sector (K2), to convert the switching pattern (320) with relative switching states (262) into a corresponding switching pattern (320) with absolute switching states (261); and - to operate the inverter (1) using the switching pattern (320) with absolute switching states (261). [15] Look-up table (300, 310) for the operation of an inverter (1); wherein the inverter (1) has a plurality of switching sections (3a, 3b, 3c), each having a phase tap (u, v, w) for electrically connecting to a respective phase (Phu Phv, Phw) of an electrical machine (133); wherein the individual switching sections (3a, 3b, 3c) each have a set of switching devices (S1, S2, S3); wherein a space vector diagram (4) of the inverter (1) is divided into K different sectors (K1,...,K6), each covering a different angular range of 2π / K of the space vector diagram (4) of the inverter (1), with K>; wherein - the look-up table (300, 100) specifies a switching pattern (320) with relative switching states (262) of the switching devices (S1, S2, S3) of the switching sections (3a, 3b, 3c) for a plurality of different modulation indices (252); and - the look-up table (300, 310) is limited to switching patterns (320) with relative switching states (262) for an angular range of 2π / K. [16] Look-up table (300, 310) according to claim 15, wherein the individual relative switching states (262) have been determined, - using an optimum pulse pattern, OPP, method; and / or - by converting absolute switching angles into relative angles and / or switching states.

Images