DE102016208801A1 - Control of a rotating field machine - Google Patents

Abstract

A polyphase inverter (100) for a rotary field machine (100) has a half bridge with two current valves (S1-S6) for each phase (U, V, W). A method (310) for controlling the inverter (100) comprises steps of determining (120) a voltage to be set on a phase (U, V, W) of the inverter (100); determining (125) a PWM signal to drive current valves (S1-S6) of a half-bridge associated with the phase (U, V, W) based on the determined voltage; determining (130) a voltage error of the phase (U, V, W) resulting from dead times of the current valves (S1-S6) of the half-bridge; determining (130) a correction voltage based on the voltage error; and compensating (130) the PWM signal for the determined correction voltage. In this case, a time profile of the correction voltage is approximated by means of Fourier synthesis to a square wave voltage.

Description

The invention relates to the control of a rotating field machine. In particular, the invention relates to a compensation of blocking periods of current valves of a bridge circuit to a field-oriented control.

The rotational behavior of a rotary field machine can be controlled by means of a field-oriented control (FOR), which is also called vector control. One problem relates to the control of an inverter, which sets predetermined voltages to strings or phases of the induction machine by means of half-bridges. In each half bridge, two flow valves are in series between different potentials of a DC link voltage. Where the flow valves are connected, the corresponding phase is also contacted. If the two flow control valves are alternately opened and closed with a predetermined duty cycle, the desired voltage is established at the phase. It is important to avoid having both flow control valves open at the same time to avoid a short circuit. The flow control valves are usually designed as semiconductors having a predetermined dead time, which elapses between the start of a signal for closing the flow control valve and the complete closing. When controlling the flow control valves, these deadlocks or dead times must be taken into account to avoid that the flow control valves are opened overlapping in time.

DE 3 844 847 C2 relates to a device and a method for dead time compensation for a three-phase motor with a pulse-controlled inverter.

DE 198 08 104 A1 Proposes to perform a dead time compensation on an inverter control device on the basis of set and actual values of phase currents.

It is an object of the present invention to provide an improved technology for the compensation of dead times. The invention solves this problem by means of the subjects of the independent claims, subclaims give preferred embodiments again.

A polyphase inverter for a rotary field machine has a half bridge with two flow control valves for each phase. A method for controlling the inverter includes steps of determining a voltage to be set at a phase of the inverter; determining a PWM signal to drive current valves of a half-bridge associated with the phase based on the determined voltage; determining a voltage error of the phase resulting from dead times of the current valves of the half-bridge; determining a correction voltage based on the voltage error; and compensating the PWM signal for the determined correction voltage. In this case, a time profile of the correction voltage is approximated by means of Fourier synthesis to a square wave voltage.

It has been recognized that the approximation of the correction voltage to the square wave signal by means of Fourier synthesis reduces a steepness of the correction voltage in the region of a zero crossing, ie a sign change, so that an erroneous determination of the time of the zero crossing has a reduced damaging influence on the blocking time compensation.

The Fourier synthesis may consider a fundamental and one or more odd harmonics. For a perfect Fourier synthesis of a square wave, infinite harmonics have to be considered. The more harmonics are taken into account, the more perfectly does the particular course approximate a square wave signal. By taking into account only a limited number of harmonics is waived a perfect replica. The imperfectly simulated signal can advantageously be used for blocking time compensation.

The highest order of a considered harmonic wave may be 3 in particular. Since the harmonics always have odd-numbered orders, only the fundamental harmonic must be combined with the third harmonic. In further embodiments, the highest order of a considered harmonic may also be 5 or 7. By taking into account the harmonics of the said orders, an improved compensation of the blocking times can be achieved. Harmonics with even higher orders are preferably not taken into account in order to prevent the steepness of the correction voltage in the area of the sign change from increasing too much.

The voltage to be set at the phase of the inverter is preferably determined by field-oriented control or field-oriented control. In this case, a discrete embodiment of the method can be used.

The voltage to be adjusted is preferably determined based on a phase current through the phase, the phase current being modeled on the basis of a predetermined desired current. The replication preferably comprises delaying an input signal specified as a space vector by a predetermined time and transforming the d, q system into the UVW system. The simulated phase current can be better noise-free compared to a measured phase current, so that in particular a zero crossing of the phase current can be determined more accurately. In a further embodiment, however, a measured phase current can also be used, for example in the context of field-oriented regulation.

A computer program product comprises program code means for carrying out the method described above when the computer program product runs on an executing device or is stored on a computer-readable medium. In particular, the method can be carried out by means of a programmable microcomputer or microcontroller, wherein at least part of the method is present as a computer program product.

An apparatus for controlling a rotating field machine by means of an inverter, wherein the inverter for each phase has a half-bridge with two current valves, comprises: a first means for determining a voltage to be set at a phase of the inverter; second means for determining a PWM signal for driving current valves of a half-bridge associated with the phase based on the determined voltage; and lock-out compensation for determining a voltage error of the phase resulting from dead times of the current valves of the half-bridge, for determining a correction voltage based on the voltage error, and for compensating the PWM signal for the determined correction voltage. In this case, the blocking time compensation is set up to approximate a time profile of the correction voltage by means of Fourier synthesis to a square-wave voltage. The device can be advantageously used in particular in the context of a field-oriented control or a field-oriented control.

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

1 an inverter;

2 the inverter of 1 with a rotating field machine;

3 a circuit diagram of a field-oriented control (FOR);

4 an αβ coordinate system with a current vector

5 Fig. 12 is an illustration of a prediction of a stream of a processing step for a following processing step;

6 Phase currents of the induction machine of 2 and rectangular equalizing voltages;

7 Phase currents of the induction machine of 2 and proposed compensation voltages; and

8th a flowchart of a method for determining corrected PWM signals for the FOR of 3
represents.

1 shows an example and schematically an inverter, in particular designed as a pulse inverter 100 with a voltage intermediate circuit 105 , which by means of an intermediate circuit capacitor 110 is formed. The inverter 100 is formed in B6 bridge circuit with three half-bridges, each having two power switches S1 and S2 or S3 and S4, or S5 and S6, each in the form of semiconductors such as MOSFETs, IGBTs, thyristors, GTOs or other current valves. The indicated diodes are optional. Three circuit breakers S1, S3, S5 of the inverter 100 are arranged as a high-side circuit breaker, three switches S2, S4, S6 as a low-side circuit breaker.

The center taps of the half bridges are each with a phase U or V or W of a rotary field machine 115 , which may be designed as a permanent-magnet synchronous machine (PMSM), electrically connected. The induction machine 115 , which may be provided, for example, for use in a motor vehicle, for example in a hybridized or electric drive train, in a power steering or as Servomotor. The induction machine 115 has in the embodiment as PSM a stator with a stator winding, in particular with the three strands U, V, W, which are preferably symmetrically wound and arranged offset by 120 degrees, and further in particular a rotor equipped with permanent magnets.

2 shows an equivalent circuit diagram of the inverter 100 and the induction machine 115 out 1 , wherein in each case two switches S1, S2 or S3, S4 or S5, S6 of a half-bridge as a switch SU, SV, SW according to the functionality of the inverter 100 are shown. Depending on a respective PWM default value for one of the phases U, V, W, each switch SU, SV, SW can switch between high-side and low-side potential for a certain duration per PWM period from -1 to +1 be (or vice versa), where in 2 the high-side potential is illustrated by the switch position +1, the low-side potential by the position -1. A voltage at a phase U, V, W is determined by what time portion of a PWM period the associated switch spends in the + 1 position. This percentage is also called the duty cycle and expressed as a percentage of the PWM period. Thus, the usual voltage vectors in a space vector modulation method at the phases U, V, W of the induction machine 115 be set freely.

By means of different modulation methods can by the inverter 100 PWM values are generated to the desired voltages for the induction machine 115 adjust. The generated PWM values from the controller or controller to produce a particular voltage vector will vary with the desired voltage frequency and amplitude.

The switches S1 to S6 are usually realized as semiconductors. Switching off a semiconductor is not sudden, but it requires a so-called dead time t ₀ until all charges are eliminated in the stopband of the semiconductor and the semiconductor is completely switched off. This dead time depends on the type of semiconductor used and, for example, is less than 1 μs for a MOSFET and between 1 μs and 5 μs for the IGBT. In the inverter 100 allow the two complementary semiconductors in one phase, z. B. S1 and S2 in Figure 1, never be turned on at the same time to a short circuit in the voltage link 105 to avoid, leading to a destruction of the DC link capacitor 110 or the semiconductor can lead. For this reason, a latching time (also called a dead time or blocking time) is usually inserted between switching off a semiconductor (until all charges have been eliminated) and switching on the complementary semiconductor at the same phase U, V, W.

When controlling the induction machine 115 arises due to this blocking time a change in the desired voltage vector, causing distortion of the voltages on the induction machine 115 can lead. In order to compensate for this blocking time, it is known to compensate the locking time depending on the measured phase current in a phase U, V, W. An accurate determination of the zero crossing of the current is very important, otherwise the compensation can lead to wrong results.

A correct compensation of the blocking times can also be important for other reasons. For controlling the induction machine 115 Other functions or models that control the voltages on the induction machine can be used 115 as input parameter. For example, the functionality of a position sensor can be monitored via a sensorless method, the state of the machine (short circuit, phase break in the machine, etc.) can be monitored by means of a voltage diagnosis, the flow model of the induction machine 115 can be calculated via a flow model, or the stator resistance of the induction machine 115 can be determined, for example, for tracking or temperature monitoring. In all these models, the voltages of the phases U, V, W of the rotary field machine 115 required, which are usually determined on the basis of the desired voltages from the scheme. With missing or incorrect lock time compensation, the specified setpoint voltages may be faulty, so that the models can also operate inaccurately or incorrectly.

In the following, a technique is proposed for compensating the blocking time with reduced error even if the detection of the zero crossing of the phase current through one of the phases U, V, W is inaccurate. The technique can be applied to a field-oriented control (FOS) or field-oriented control (FOR) in conjunction with different induction machines 115 (eg a permanently excited synchronous machine or an asynchronous machine), wherein the blocking times of different semiconductors (eg MOSFET, FET, IGBT, GTO, thyristor etc.) can be taken into account and the zero crossings of the phase currents alternatively measured or by means of a Model can be determined.

For a more detailed explanation of the proposed technique, a FOR is described by way of example. 3 shows an exemplary field-oriented control for the induction machine 115 from the 1 and 2 , The field-oriented control can be used as a control device 305 be executed, which is a control of the rotational behavior of the induction machine 115 performs on the basis of a space vector pointer in the manner of a vector control. For this purpose, parts of the control device can in particular be comprised by a programmable microcomputer, wherein the processing takes place digitally. The representation of 3 However, it can also be used as a process flowchart 310 for controlling the induction machine 115 be understood.

A space vector, which is given in d, q representation and includes the components IsdRef and IsqRef, is present as input. The d component of the space vector is a magnetic flux, and the q component a torque of the induction machine 115 assigned. The components of the space vector are transmitted to a transformation device via optional proportional-integral elements PI 120 given further, which transforms the input variables into three voltages Us1, Us2, Us3, which are connected to the strings of the induction machine 115 are to be adjusted. Optionally, the determined voltages may then be limited to valid values by means of a limiter before a PWM generator 125 based on the determined voltages PWM signals PWM1, PWM2, PWM3 for the switches S1 to S6 of the inverter 100 determines the desired voltages on the strands U, V, W of the induction machine 115 provide.

For the control, it is necessary to determine currents Is1, Is2, Is3 flowing through the strands U, V, W. For this, different approaches are possible. In the illustrated embodiment, the strand currents are detected by current sensors 135 sampled. The phase currents Is1, Is2, Is3 are based on an electrical rotation angle Θ _{el of} the induction machine 115 by means of a further transformation device 170 transformed into the d, q-coordinate system, wherein the electrical rotation angle Θ _{el can} also be determined in different ways. In the present case, for example, the mechanical angle of rotation Θ _mech by means of a position _sensor 140 sampled, which may be formed for example as an arrangement of Hall sensors or incremental encoders. The electrical rotation angle Θ _el can then be multiplied by the mechanical rotation angle Θ _mech with the number of poles Zp of the induction machine 115 be determined.

The transformed values Isd, Isq of the phase currents are added to the values of the predetermined space vector, before these are sent to the PI elements or to the transformation device 120 be guided. The DC link voltage Udc used in various determination steps may be determined in any known manner.

Optionally, the components Isd, Isq fed back additively to the space vector can be separated by means of a decoupler 145 decoupled from each other and as EMKd and EMKq additive to the input of the transformation device 120 be coupled. For this purpose, the electrical angular velocity ω _el can be used by multiplying the mechanical angular velocity ω _mech by the number of poles Zp of the induction machine 115 can be determined. The mechanical angular velocity ω _mech can be determined by deriving the mechanical rotation angle Θ _mech after the time. An optional position estimation model 150 on the basis of the PWM signals and the phase currents, an estimated angular velocity ω ^ and an estimated rotation angle Θ ^ and the induction machine 115 ready.

It is proposed to use the PWM signals PWM1, PWM2, PWM3 of the PWM generator 125 by voltage errors due to blocking times of the switches S1 to S6, by means of a blocking time compensation 130 if possible, to compensate such that a state in which two switches of a half-bridge are simultaneously closed (S1 and S2, S3 and S4, or S5 and S6) is avoided, and at the same time voltage errors between the voltages applied to the phases U, V, W and intended voltages are as low as possible.

The determination of the voltage errors takes place with respect to the phase currents flowing through the individual lines U, V, W of the induction machine 115 flow. In one embodiment, the actual phase currents Is1, Is2, Is3 are detected by means of the current sensors 135 sampled. In another embodiment, replicated strand streams are used, which may have a reduced noise component. For this purpose, the space vector is preferably in a filter designed, for example, as a low-pass filter 160 delayed by a predetermined amount, which reflects the dynamics of the FOR or FOS as well as possible. Curves of the delayed setpoint values of the d, q currents should as far as possible follow the same course as the actual currents of the induction machine 115 to have. The inverse transformation of the delayed setpoint values of the d, q currents yields the simulated phase currents in the UVW coordinate system without noise. As a result, the use of measured and superimposed with a noise phase currents can be avoided, in particular to be able to determine a zero crossing of the phase currents improved. If the phase currents are measured, for example in the context of a FOR, the measured values can also be used directly; a replica, for example by means of the filter 160 and the transformation device 165 , then is not required. The means of transformation means 165 certain reference currents IsαF, IsβF are then preferably by means of yet another transformation means 170 transformed into the Cartesian coordinate system ("Cartesian To Polar" transformation).

The blocking time compensation 130 corrects the PWM signals PWM1, PWM2, PWM3 on the basis of the in the transformation means 170 certain phase and the specific amount of the space pointer to lock time-related voltage error, and provides the inverter 100 correspondingly recalculated PWM signals PWM1n, PWM2n, PWM3n ready.

If the phase currents of the phases U, V, W are determined periodically at regular intervals, as in a microprocessor-controlled device 305 is usual, then in particular an error in the determination of a voltage error can result if the phase current within a current period, ie between two determination times, its sign changes. Especially with long periods, such an error in the compensation can be prevented. Long periods occur especially when per electric period of the induction machine 115 only relatively few sampling periods of the currents take place, for example, less than 1000, less than 100 or less than 10. It is therefore proposed to take account of a change of sign occurring during a period in the compensation of the voltage errors.

des gefilterten Sollwertes und für jede Phase x des gefilterten Sollwerts I_sx(= I_sxRefF) des Phasenstroms (Nachbilden von I_sx) für die Berechnungen des Spannungsabfalls wegen der Sperrzeit der Stromventile S1–S6 im Wechselrichter 100 verwendet.The phases U, V, W are hereinafter also referred to as phase x with x ε {1; 2; 3}. For the following calculations, the current pointer

the filtered setpoint and for each phase x of the filtered setpoint I _sx (= I _sxRefF ) of the phase _current (simulating I _sx ) for the calculations of the voltage drop due to the blocking time of the current valves S1-S6 in the inverter 100 used.

mit:

U_sx:

Spannungsfehler wegen der Sperrzeit des Wechselrichters in Phase x

t₀:

Sperrzeit des Wechselrichters 100

T_s:

Schaltperiode des Wechselrichters 100 (T_s = 1/f_s; f_s ist die Schaltfrequenz des Wechselrichters 100)

U_dc:

Zwischenkreisspannung 105

sign(I_sxRefF):

Vorzeichen des gefilterten Sollwerts des Phasenstroms (Nachbilden von I_sx) in Phase x

When compensating the inverter blocking time, a voltage error ΔU _{sx is} compensated by default for each phase:

With:

U _sx :

Voltage error due to the blocking time of the inverter in phase x

t ₀ :

Blocking time of the inverter 100

T _s :

Switching period of the inverter 100 (T _s = 1 / f _s ; f _s is the switching frequency of the inverter 100 )

U _dc :

Intermediate circuit voltage 105

sign (I _sxRefF ):

Sign of the filtered setpoint value of the phase current (simulation of I _sx ) in phase x

In a discrete control device 305 will be with respect to 3 periodically repeated procedure described. A current processing step is denoted by k; the previous one is then k-1 and the following k + 1. At the beginning of each step, the PWM times PWM1, PWM2, PWM3 are determined which in the next sampling step are sent to the inverter 100 be handed over. Therefore, the current vector is rotated forward or backward depending on the sign of the electrical angular velocity. The current vector from the filtered nominal currents or the measured phase currents are used.

Die Phase (der Richtungswinkel) des Stromzeigers

fällt in einen von sechs Bereichen, die jeweils 60° messen und in 4 jeweils zwischen zwei Strich-Punkt-Punkt-Linien liegen. In jedem Bereich sind die Vorzeichen der drei Phasenströme I₁, I₂, I₃ eingezeichnet. Beisielsweise ist in dem Bereich, in dem in 4 der Stromzeiger

exemlarisch dargestellt ist, I₁ = I_U > 0; I₂ = I_V > 0 und I₃ = I_W < 0. Nach der Bestimmung des Stromzeigers

im Polarkoordinatensystem (Betrag und Phase) können anhand der Phase des Stromzeigers

der zutreffende Bereich und daraufhin die Vorzeichen der einzelnen Phasenströme I₁, I₂, I₃ bestimmt werden. 4 shows an αβ coordinate system with a current vector

The phase (the direction angle) of the current vector

falls into one of six areas, each measuring 60 ° and in 4 each lie between two dash-dot-dot lines. In each area, the signs of the three phase currents I ₁ , I ₂ , I _{3 are} drawn. For example, in the area where in 4 the current pointer

is shown exemlarisch, I ₁ = I _U >0; I ₂ = I _V > 0 and I ₃ = I _W <0. After determining the current vector

in the polar coordinate system (Amount and phase) can be determined by the phase of the current pointer

the true range and then the signs of the individual phase currents I ₁ , I ₂ , I _{3 are} determined.

By means of the transformation device 165 For example, the filtered currents IsdRefF and IsqRefF can be transformed into the αβ coordinate system in the following way:

By transforming the two currents I _sαF and I _sβF from the Cartesian coordinate system into the polar representation, the current vector value ∥I _s ∥ and the current vector phase (or the current vector angle) θ _{I (k) are obtained} .

5 shows a time course of a phase current of a phase U, V, W with zero crossing within a sampling period T _A. In the horizontal direction is a time and in the vertical direction a current plotted. A sampling period T _A denotes the duration of a processing step k, k + 1, etc.

auf den gleichen Zeitpunkt der Einstellung der PWM-Werte korrigiert werden. Dazu wird die Lage des Stromzeigers

für die Mitte des folgenden Schritts k – 1 als Schätzung vorhergesagt. Da die mechanische Zeitkonstante der Drehfeldmaschine 115 üblicherweise größer als die elektrische ist, kann die Drehzahl der Drehfeldmaschine 115 über aufeinander folgende Abtastperioden T_A als näherungsweise konstant betrachtet werden. Damit besteht die Möglichkeit, die Lage des Stromzeigers

zur Mitte der folgenden Abtastperiode θ_I(k+1) mithilfe der aktuellen Phase θ_I(k) des Stromzeigers

und der elektrischen Winkelgeschwindigkeit ω_el bzw. der Drehzahl der Drehfeldmaschine 115 vorausberechnen: θ_I(k+1) = θ_I(k) + ω_el(k)· 3 / 2T_A (Gleichung 3) mit:

θ_I(k):

Phase des Stromzeigers

zum Zeitpunkt k·T_A

θ_I(k+1):

Phase des Stromzeigers

zum Zeitunkt (k + 3/2)·T_A

ω_el:

die Winkelgeschwindigkeit der Drehfeldmaschine 115 (entspricht Drehzahl·2π·Polpaarzahl/60)

T_A:

die Dauer der Regel-Abtastperiode

PWM values determined by the FOR or FOS in step k usually do not arrive until approximately the middle of the next step k + 1 at the inverter 100 set. Therefore, the location of the current pointer must be

be corrected to the same time of setting the PWM values. To do this, the location of the current pointer

for the middle of the following step k-1 is predicted as an estimate. As the mechanical time constant of the induction machine 115 Usually larger than the electrical, the speed of the induction machine can 115 over consecutive sampling periods T _{A are} considered to be approximately constant. This gives the possibility of the position of the current indicator

to the middle of the following sampling period θ _{I (k + 1)} using the current phase θ _{I (k) of} the current vector

and the electrical angular velocity ω _el or the rotational speed of the induction machine 115 predict: θ _{I (k + 1)} = θ _{I (k)} + ω _{el (k)} · 3 / 2T _A (Equation 3) With:

θ _{I (k)} :

Phase of the current pointer

at the time k · T _A

θ _{I (k + 1)} :

Phase of the current pointer

at the time point (k + 3/2) · T _A

ω _el :

the angular velocity of the induction machine 115 (corresponds to speed · 2π · pole pair number / 60)

T _A :

the duration of the rule sample period

6 shows phase currents I _su , I _sv , I _{sw of} the induction machine 115 as well as rectangular correction voltages ΔUs ₁ Tot, ΔUs ₂ Tot, ΔUs ₃ Tot of the blocking time compensation 130 , each as a temporal course. The specified scale values are purely exemplary.

The correction voltages behave like rectangular alternating voltages, which are shifted by 120 ° from each other. Incorrect zero phase current compensation can result in a voltage error that is up to twice the original voltage error. If the zero crossing is thus determined incorrectly, the compensation can increase the voltage error instead of reducing it.

It is proposed to approximate the rectangular correction voltages by means of a superposition of sinusoidal oscillations of a fundamental and at least one harmonic. As a result, an error that results from an incorrect determination of the zero crossing can be kept small, while at the same time a good compensation of the blocking times can be achieved. In general, the decomposition of the square-wave signal results 6 by Fourier synthesis an equation with a fundamental and infinitely many harmonics odd orders, the amplitudes of the harmonics decrease with increasing atomic number. If only finitely many harmonics are taken into account, the course of the correction voltage assumes a more harmonic course, with the zero crossing in particular being less steep. It has been found that a Fourier synthesis with harmonics of orders of at most 5, preferably at most 3, is sufficient to provide a correction voltage which combines a good blocking time compensation with a weakened slope in the region of the zero crossing. Due to the reduced slope, an incorrectly determined zero crossing may have less disturbing influence on the blocking time compensation.

7 shows the phase currents I _su , I _sv , I _{sw of} the induction machine 115 and proposed correction voltages ΔUs ₁ Tot, ΔUs ₂ Tot, ΔUs ₃ Tot of lock-time compensation 130 as temporal courses. The specified scale values are again purely exemplary. In contrast to the representation of 6 Here, the correction voltages are shown as sinusoidal fundamental and superimposed with sinusoidal harmonics of the 3rd order fundamental vibrations. In the area of the zero crossings of the phase currents, the magnitudes of the correction voltages are very small, both in the fundamental and in the superimposed fundamental oscillations. An incorrectly determined zero crossing of a phase current therefore causes only a very small voltage error. Nevertheless, the effect of the blocking time, such as a sixth-order harmonic in the torque or a voltage distortion, can already be greatly attenuated by compensating the blocking time voltages by means of fundamental oscillations. The shorter the sampling period T _A , the smaller the error due to a possible wrong determination of the zero crossing.

Diese Drehfeldspannung der Grundschwingungen der drei Spannungskorrekturen ist in Phase mit dem Drehfeldstrom der Drehfeldmaschine 115.The fundamental oscillations of the three voltage corrections in the three phases U, V, W, result in a field voltage with a constant amount:

This field voltage of the fundamental oscillations of the three voltage corrections is in phase with the rotating field current of the induction machine 115 ,

If the fundamental and the third order harmonic are taken into account, then a curve is obtained which approximates the rectangular curve and likewise contains no abrupt change of the sign when the phase current changes. In other words, the slope of the approximated curve in the region of the zero crossing of the phase current (or its profile) is still significantly smaller than that of the rectangular signal. By superimposing the sinusoidal fundamental wave with the 3rd harmonic, the lock-up time can be compensated for better because it has a larger average amplitude over one period. Around the zero crossings of the phase currents (or the correction voltage) around the error, resulting in inaccurate or incorrect determination of the zero crossing, continues to be small.

Due to the symmetrical rotating field in a three-phase induction machine 115 the 3rd harmonics of the correction voltages in the voltage vector are canceled altogether. In the individual phases U, V, W, the blocking-time-dependent voltage drop is better compensated by the curve approximated by the third harmonic than by the fundamental alone. The induction machine 115 is not affected by the third harmonic, that is, there are no additional harmonics in the phase current.

für Phase 2:

und für Phase 3:

Equations 4 through 6 are used to calculate the voltage errors. For phase 1:

for phase 2:

and for phase 3:

für Phase 2:

und für Phase 3:

Equations 7 to 9 are used for the calculation of the pulse width modulation signals PWM1n, PWM2n and PWM3n. For phase 1 you get:

for phase 2:

and for phase 3:

8th shows a flowchart of a method 800 , The procedure 800 is set up for running on a processing device, which may in particular comprise a programmable microcomputer or microcontroller. Usually the procedure becomes 800 go through periodically and forms part of the processing described above with reference to 3 is described in more detail. The procedure 800 is suitable for use with a FOS or FOR.

The procedure 800 starts in one step 805 , Here, the phase currents Isu, Isv, Isw and the currents IsdRef, Isqref, the electrical angle θel and the mechanical angle ωel are determined. In one step 810 the streams IsdRef and IsqRef become optional in the filter 160 filtered. Then they are transformed by means of the transformation device 165 is transformed into the αβ system of the stator on the basis of the electrical angle θel (see Equation 2), so that the phase currents Isu, Isv and Isw result.

zum Zeitpunkt k bestimmt. In einem Schritt 620 wird auf der Basis der Phase des Phasenstromzeigers

zum aktuellen Verarbeitungsschritt θ_I(k) und dem mechanischen Winkel ωel der Phasenstromzeiger zum folgenden Verarbeitungsschritt θ_I(k+1) geschätzt bzw. vorhergesagt (vgl. Gleichung 3 und 5).In one step 615 By means of the transformation "Cartesian To Polar" from the currents Isα and Isβ the magnitude and the phase of the phase current vector

determined at time k. In one step 620 is based on the phase of the phase current vector

to the current processing step θ _{I (k)} and the mechanical angle ωel of the phase current phasors are estimated to the following processing step θ _{I (k + 1)} (see Equation 3 and Figs 5 ).

In one step 625 For each phase U, V, W, the blocking-time-dependent voltage drop Δus ₁ , Δus ₂ or Δus _{3 is} determined (compare equations 4, 5, 6 and 6 ). In this case, the course of each voltage drop is approximated as a superimposition of a sinusoidal fundamental and a sinusoidal harmonic of 3rd order (cf. 7 ).

In a subsequent step 630 For this purpose suitable PWM signals PWM1n, PWM2n, PWM3n are formed (compare Equations 7, 8, 9). These can be done in one step 835 to the inverter 100 be passed on.

LIST OF REFERENCE NUMBERS

100

inverter

105

Voltage link

110

Link capacitor

115

Induction machine

120

transformation means

125

PWM generator (optional limiter)

130

Blocking time compensation

135

current sensor

140

position sensor

145

decoupler

150

Position estimation model

160

filter

165

transformation means

170

transformation means

170

transformation means

S1-S6

Switch, flow control valve

AND MANY MORE

Phase or strand

305

control device

310

method

QUOTES INCLUDE IN THE DESCRIPTION

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

Cited patent literature

• DE 3844847 C2 [0003]
• DE 19808104 A1 [0004]

Claims (8)

Procedure ( 310 ) for controlling a polyphase inverter ( 100 ) for an induction machine ( 115 ), where the inverter ( 100 ) for each phase (U, V, W) has a half bridge with two flow control valves (S1-S6) and the method comprises the following steps: determining ( 120 ) of a voltage, which at a phase (U, V, W) of the inverter ( 100 ) should be set; Determine ( 125 ) a PWM signal for driving current valves (S1-S6) of a half-bridge associated with the phase (U, V, W) on the basis of the determined voltage; Determine ( 130 ) a voltage error of the phase (U, V, W), which results from dead times of the flow control valves (S1-S6) of the half-bridge; Determine ( 130 ) a correction voltage based on the voltage error; and Compensate ( 130 ) of the PWM signal to the specific correction voltage, characterized in that a time course of the correction voltage is approximated by means of Fourier synthesis to a square wave voltage. Procedure ( 310 ) according to claim 1, wherein the Fourier synthesis takes into account one fundamental and one or more odd harmonics. Procedure ( 310 ) according to claim 2, wherein the highest order of a considered harmonic wave is 3. Procedure ( 310 ) according to one of the preceding claims, wherein the voltage applied to the phase (U, V, W) of the inverter ( 100 ), is determined by means of field-oriented control or field-oriented control. Procedure ( 310 ) according to one of the preceding claims, wherein the voltage to be set is determined on the basis of a phase current through the phase (U, V, W) and the phase current is simulated on the basis of a predetermined desired current. Procedure ( 310 ) according to claim 4 and 5, wherein the replicating comprises delaying an input signal given as a space vector by a predetermined time and transforming the d, q system into the UVW system. Computer program product with program code means for carrying out a method ( 310 ) according to one of the preceding claims, when the computer program product runs on an executing device or is stored on a computer-readable medium. Contraption ( 305 ) for controlling an induction machine ( 115 ) by means of an inverter ( 100 ), where the inverter ( 100 ) for each phase (U, V, W) has a half-bridge with two flow control valves (S1-S6) and the device comprises: a first device ( 120 ) for determining a voltage which is connected to a phase (U, V, W) of the inverter ( 100 ) should be set; a second facility ( 125 ) for determining a PWM signal for driving current valves (S1-S6) of a half-bridge associated with the phase (U, V, W) on the basis of the determined voltage; and a blocking time compensation ( 130 ) for determining a voltage error of the phase (U, V, W) resulting from dead times of the current valves (S1-S6) of the half-bridge, for determining a correction voltage based on the voltage error, and for compensating the PWM signal by certain correction voltage, characterized in that the blocking time compensation ( 130 ) is adapted to approximate a time course of the correction voltage by means of Fourier synthesis to a square wave voltage.

Images