WO2019008242A1

WO2019008242A1 - Method for controlling a three-phase inverter

Info

Publication number: WO2019008242A1
Application number: PCT/FR2018/051289
Authority: WO
Inventors: Mohamad KOTEICH; Najib ROUHANA
Original assignee: Renault S.A.S; Nissan Motor Co.Ltd
Priority date: 2017-07-05
Filing date: 2018-06-05
Publication date: 2019-01-10
Also published as: CN110999067A; EP3649730A1; KR20200039664A; CN110999067B; FR3068845A1; KR102595777B1; FR3068845B1

Abstract

The invention relates to a method for controlling a three-phase voltage inverter receiving a direct input voltage, the method comprising a step (40) of correcting the reference voltage vector comprising: - a step (44) of computing at least one trigonometric value (Sk, Ck) of an angle formed by the bisector of the sector in which the reference voltage vector is located with the abscissa of the first two-dimensional reference frame as a function of the signs of the two components of the reference voltage vector in the first two-dimensional reference frame; - a step (47) of computing components of a corrected reference three-phase voltage as a function of said at least one computed trigonometric value (Sk, Ck).

Description

Method of controlling a three-phase inverter

The present invention relates to a method for controlling a three-phase inverter, in particular used in the powertrains of electric and hybrid vehicles.

More specifically, this invention relates to the management of the reference voltage sent to the inverter modulation strategy for controlling the power switches in the case where the control voltage is greater than the maximum voltage of the storage battery. In the automotive field, an electric power train (abbreviated

GMPE) consists in particular of a three-phase electric machine, controlled by a three-phase inverter.

The inverter is a static electrical circuit, composed of a plurality of semiconductor configurations, also called power switches, controlled by a digital computer implementing a control algorithm.

The purpose of the control method of an inverter is to provide a "close" control to enslave the torque generated by the electric machine powered by the inverter to the required value.

Thus, the voltage inverter, using a pulse width modulation strategy (abbreviated PWM), transforms the voltage supplied by the DC source Vdc into an alternating voltage of variable amplitude and frequency. The role of a modulation strategy is to provide the load with an output voltage whose waveform is close to a sinusoid.

A MLI control technique known as the space vector method, also known in English as Space Vector Modulation, is known.

In the spatial vector method, three-phase control voltages, provided by a closed-loop control command, are transformed into a single reference voltage vector Vref in a two-dimensional reference ( ^a fi). This transformation is generally carried out by a Clarke transform, well known to those skilled in the art. The linear operation of the inverter, in this reference (° ^) is delimited

Vdc

by a centered circle of radius - = -, Vdc being the DC voltage supplied in

V3

input of the voltage inverter.

However, when the reference voltage vector has a standard value greater than the radius of the circle, the operation of the inverter becomes non-linear, and is said overmodulation zone. In this overmodulation mode, a well-known problem is the distortion of the output signals with respect to the expected sinusoids.

However, the over-modulation zone provides a useful extension of the operating range of the inverter without requiring an increase in the amplitude of the DC voltage of the source Vdc.

Also, a known problem is to improve the control voltages of the inverter to allow optimal operation of the inverter in overmodulation area.

In the prior art, control methods that are generally complex to implement and expensive in computing time are known because they require the use of trigonometric operators. In particular, the method described in the publication "On continuous control of PWM inverters in the overmodulation range including the six-step mode" is known in IEEE Transactions on Power Electronics 8.4 (1993): 546-553, by Holtz, Joachim, Wolfgang Lotzkat, and Ashwin M. Khambadkone.

Overmodulation control methods are also known, such as that described in the publication "Overmodulation strategy for high-performance torque control." In IEEE Transactions on Power Electronics 13.4 (1998): 786-792, by Seok, Jul-Ki, Joohn -Sheok Kim and Seung-Ki Sul.

There is therefore a need for a method for optimizing the operation of the inverter in a zone of over-modulation, which is reliable and requires low computing times for a digital computer, which is not bulky in storage space and is easy to implement.

There is provided a method of controlling a three-phase voltage inverter receiving an input DC voltage; the method comprising:

a step of receiving a three-phase voltage setpoint; a step of transforming said three-phase voltage setpoint into a reference voltage vector, defined by two components in a first two-dimensional reference; a plurality of sectors being defined in said first two-dimensional mark, each sector corresponding to a portion of space of the two-dimensional mark formed between two instantaneous voltages of the inverter angularly adjacent;

a step of correcting the reference voltage vector.

a step of controlling the inverter as a function of a corrected three-phase reference voltage during a correction step,

the correction step comprising:

a step of determining three variables with three possible values according to the signs of the components of the reference voltage vector in the first two-dimensional reference, and a comparison of the absolute values of said components,

a step of calculating trigonometric values of an angle made by the bisector of a sector in which said reference voltage vector is located, with the abscissa of the first two-dimensional reference mark as a function of the variables determined previously,

a step of calculating components of the corrected three-phase reference voltage as a function of said previously calculated trigonometric values.

Thus, it is possible to determine a corrected reference three-phase voltage that does not need to store precalculated tables of values, and that do not involve trigonometric calculation, which is generally expensive in calculation time, the calculated trigonometric values being obtained by calculations based on simple logical operators (addition, multiplication, subtraction and division) and on comparisons of values.

It is advantageous to determine the trigonometric values associated with the bisector of the sector in which the vector reference voltage is located, without it being necessary to determine this sector. In other words, the step of calculating the trigonometric values is performed independently of the detection of the sector. This makes it possible to obtain a process that is particularly simple and quick to implement. Advantageously and in a nonlimiting manner, said step of calculating the components of a corrected reference three-phase voltage comprises a step of determining a transformed vector in a second two-dimensional coordinate system as a function of said at least one trigonometric value. Thus, the calculations for obtaining the components of the corrected three-phase voltage are simplified.

Advantageously and in a nonlimiting manner, said step of calculating the components of a corrected reference three-phase voltage comprises the saturation of the transformed vector. Thus, it is ensured that the reference vector is corrected within the limits of linear operation of the inverter in a relatively simple and fast manner by applying the saturation directly to the transformed vector.

Advantageously and in a nonlimiting manner, said variable calculation step comprises comparing the absolute value of a component of the reference voltage vector in the first two-dimensional coordinate system with another component of the reference voltage vector in the first two-dimensional coordinate system multiplied by V3. In this way the trigonometric values are obtained relatively accurately and by simple logic operations.

Advantageously and in a nonlimiting manner, said step of calculating the trigonometric values comprises calculating a value of the cosine and a sine value of said angle. In this way, the correction values can be determined more simply.

Advantageously and in a nonlimiting manner, said calculation of said sine value is performed regardless of the determined sector, by a single equation depending on the component of the reference voltage vector in the first two-dimensional reference and of said comparison. Thus, one can optimize the execution time of the process.

Advantageously and in a nonlimiting manner, said calculation of said cosine value is performed regardless of the determined sector, by a single equation function of said other component of the vector reference voltage in the first two-dimensional reference and of said comparison. Thus, one can optimize the execution time of the process. The invention also relates to a control device, for example a microcontroller, a microprocessor, a DSP, a computer, for example embedded in a motor vehicle, an inverter comprising:

means for receiving a three-phase voltage setpoint;

means for transforming said three-phase voltage setpoint into a reference voltage vector, defined by two components in a first two-dimensional reference frame; a plurality of sectors being defined in said first two-dimensional mark, each sector corresponding to a portion of space of the two-dimensional mark formed between two instantaneous voltages of the inverter angularly adjacent; and

means for correcting said reference voltage vector;

control means of the inverter as a function of a corrected reference three-phase voltage supplied by the correction means;

the correction means of said reference voltage vector being adapted to:

determining three variables with three possible values according to the signs of the components of the vector reference voltage in the first two-dimensional reference, and a comparison of the absolute values of said components,

calculating trigonometric values of an angle made by the bisector of a sector in which said reference voltage vector is located, with the abscissa of the first two-dimensional reference mark as a function of said variables,

calculating components of the corrected three-phase reference voltage in said first two-dimensional reference mark as a function of said calculated trigonometric values.

The invention also relates to an electrical assembly comprising a three-phase voltage inverter and a control device as described above.

The invention also relates to a motor vehicle comprising an electrical assembly as described above.

Other features and advantages of the invention will emerge on reading the description given below of a particular embodiment of the invention, given as an indication but without limitation, with reference to the accompanying drawings in which:

FIG. 1 is a representation of the reference voltages and operating voltages of a three-phase inverter in the two-dimensional reference frame obtained by the Clarke transform, known from the prior art;

FIG. 2 is a schematic view of a three-phase inverter known from the prior art; and

FIG. 3 is a flowchart of the correction steps of the control method according to one embodiment of the invention.

With reference to FIG. 2, a three-phase voltage inverter 20 comprises three switching arms A, B, C, each having two series power switches, A +, A-, B +, B- and C +, C-, respectively.

For each arm of the inverter A, B, C, respectively corresponding to a phase of the three-phase voltage to be generated, the two switches A +, A-, B +, B- and C +, C- can not be in the same state, open or closed at the same time. In other words, when one of the switches of one arm is closed, the other switch of the same arm is necessarily open, otherwise a short circuit would occur.

In the following description, for each arm A, B, C, when the upper switch, respectively A +, B +, C + is closed, describes the state of the arm considered by convention by the binary value 0, while when the lower switch respectively A-, B-, C- is closed, the arm is described by the binary value 1.

Thus, in order to describe the general configuration of the inverter at a given time, a three-bit binary write is employed. For example, the value 01 1 can be read as:

the most significant bit 0 for the arm A = A + closed;

the middle bit 1 for the B = B-closed arm; and

the least significant bit 1 for the closed arm C = C-.

In order to control the openings and closures of these switches so as to produce a sinusoidal signal, from a source of direct current VDC, a method of controlling the inverter by modulation is implemented. pulse width, PWM, also known as the Anglo-Saxon Pulse Width Modulation, PWM.

Here, the MLI is performed by the technique known to those skilled in the space vector art, also known by the Anglo-Saxon name of Space Vector Modulation.

This technique, with reference to FIG. 1, models the three-phase system of voltages to be generated for the current sampling duration in the form of a single reference voltage vector Vref.

The vector reference voltage Vref is obtained in a first reference

two-dimensional by a Clarke transform, well known to those skilled in the art, which provides the coordinates ^ and ^ of the reference voltage vector Vref, as a function of the voltages of the three phases received:

With Va the fundamental component of the output voltage of the arm A, Vb the fundamental component of the output voltage of the arm B and v _c fundamental component of the output voltage of the arm C.

In this reference (° ^) the eight possible configurations of the switches of the inverter are represented by the instantaneous voltage vectors V0 to V7 with, respectively, in binary writing as described previously:

V0: 000

V1: 100

V2: 1 10

V3: 010

V4: 01 1

V5: 001

V6: 101

V7: 1 1 1

Here, V0 corresponds to the case where all the upper switches A +, B +, C + are closed, and V7 in the case where all the lower switches A-, B-, C- are closed. The instantaneous voltage vectors VO and V7 are called freewheeling vectors.

The instantaneous voltage vectors V1 to V6 are configurations well known to those skilled in the art, as part of a Space Vector Modulation.

Since the inverter 20 has only a limited number of possible configurations, namely the only generation of instantaneous voltage vectors V0 to V7, it is well known to those skilled in the art that as a function of the reference voltage vector Vref desired, is applied for brief moments a succession of instantaneous voltage vectors VO to V7, in order to obtain on average the desired reference voltage vector Vref.

The convex hull 1 1 of the instantaneous voltage vectors V1 to V6 in the two-dimensional coordinate system (° ^) forms a hexagon 1 1. This convex hull 1 1 corresponds to the set of points from which the reference voltage vector can produce a full wave operation of the inverter 20.

The hexagon 1 1 is subdivided into six sectors S1 -S6, each limited by two vectors of non-zero instantaneous voltages V1 -V6 and the outer segments of the hexagon 1 1.

In the hexagon 1 1 is written a circle 10 defining the amplitude of the reference voltage vector Vref in which the linearity of the modulation is ensured.

The radius of the inscribed circle 10 is a value of Vdc / 3 3; Vdc being the DC voltage supplied at the input of the inverter 20.

When the reference voltage vector Vref remains within the limits of the inscribed circle 10, no problem of linearity is to be noted, these voltages being directly attainable by the inverter 20.

However, when the reference voltage vector Vref exceeds the limits of the inscribed circle 10, the inverter 20 then goes into over-modulation. Here, the linearity between the setpoint voltages, at the input of the pulse width modulation strategy (PWM), and the fundamental component of the voltage actually produced by the inverter 20, is no longer ensured. The reference voltage Vref is then brought back to a value corresponding to a projection on the hexagon 1 1 so as to reduce the control to a value of allowable voltage.

Whatever the sector S, between 1 and 6, and bisector respective di to d6, in which is the vector of the reference voltage Vref, also called control voltage Vref, the variables Sk and Ck, which are respectively equal to the sine and cosine of the angle λ that bisects the corresponding sector S with the abscissa axis a (axis of the first phase of the machine, phase a, V1), without using trigonometric functions, for limit the computational load, or pre-calculated value tables, to limit the memory space used.

Calculation of the variables Sk and Ck is therefore performed without it being necessary to determine in advance, or to know, the sector S in which the reference voltage is located, which improves the simplicity of the correction method. In other words, it is possible to determine the angular values of the bisector of the corresponding sector S, from the equations described below, without it being necessary to determine the corresponding sector S in which the reference voltage Vref is located.

To do this, starting from the equation (1) above, variables with three possible values (-1, 0, 1) are determined as a function of the signs sgn _a , sgn _p of the values v _a and v _p of the vector voltage reference Vref in the first two-dimensional coordinate ^system P>, and the comparison 43 of the absolute values sgn _aP of the same values ν _α , ν _β , as follows:

1 if \ ν _β \> 3 | vJ

Then, the values of sine Sk and cosine Ck are calculated by the following equation according to the three variables sgn _a , sgn _p and sgn _afj obtained previously:

¾ = + sgn _ap ) (5) c _k = ^ (l - sgn _ap ) (6)

In particular, it is possible to precalculate and store in memory the value V3 / 4 and VA so as to make the calculation of the cosine Ck and of the sinus Sk even faster.

Next, a reference voltage transformed in the sector mark S is calculated (this mark being obtained by a rotation of angle λ of the mark (αβ)) from the values ν _α , ν _β , s _k and c _k as follows :

dk = ¾ν _α + s _k v _p (7)

v _qk = ¾ν _β - s _k v _a (8)

The values obtained v _dk , v _qk correspond to the projections of Vref in the reference obtained by a rotation of angle λ of the two-dimensional reference point ^), and are then saturated 46 so as to obtain saturated reference voltage values ^v dk ^{P and v} qk ^P in the reference of the corresponding sector S:

P = min { _¾J ¾ (9)

^V ^P qk = ^min (l _r ^S 9 ^U q} (10)

These equations (9) and (10) make it possible to apply a maximum over-modulation rate (1 10% of the maximum linear voltage Vd _C / V3).

Equation (10) allows in particular to increase the amplitude of the output voltage up to 2 Vd _C / ft, and therefore allows to apply the full wave control.

In a variant which does not have this advantage, but which makes it possible to limit the distortion in phase of the voltage, equation (10) is replaced by the following equation (while keeping equation (9)):

app

app _ v _{dk m} ,

- max { _W } ^V « ^k

In which ε is a small quantity (for example ε = 0.0001) which eliminates the risk of a division by zero during the numerical resolution of equation (10) .: ^S gn _qk

With sgn _qk =

The three components of the final three-phase voltage to be sent to the inverter

20 are obtained by numerical application of the following equations:

= ¾ -¾ (12) vapp ₌ _ _v app _ app _{) These components of the final three-phase voltage v ^ ^pp , v ^ ^pp , v ^ ^pp are thus within the limits of the hexagon 1 1 of voltage, c that is, in the permissible voltage region at the output of the inverter 20.

The equations from (1) to (13) can easily be implemented and can be quickly solved by calculators, as they involve only elementary computational functions, addition, multiplication and comparison.

In addition, if the reference voltage Vref is lower than this voltage is not changed, which ensures continuity between the two linear and non-linear operating areas and thus improves the stability of the control.

Claims

1. A method of controlling a three-phase voltage inverter (20) receiving an input DC voltage (VDC);

the method comprising:

a step of receiving a three-phase voltage setpoint (v _a , Vb, v _c );

a step of transforming said three-phase voltage setpoint (v _a , Vb, v _c ) into a reference voltage vector (Vref), defined by two components ( ^{v "} , ^V? ) in a first two-dimensional reference (?) ; a plurality of sectors (S1-S6) being defined in said first two-dimensional reference (^), each sector (S1-S6) corresponding to a space portion of the two-dimensional reference (αβ) formed between two instantaneous voltages (V1 - V6) of the inverter (20) angularly adjacent;

a step of correcting (40) the vector reference voltage (Vref); and a step of controlling said inverter (20) as a function of a corrected reference three-phase voltage obtained at said correction step (40),

characterized in that the correction step (40) comprises:

a step of determining three variables (sgn _a , sgn _p , sgn _aP ) with three possible values (-1, 0, 1) as a function of the component signs ( ^{v "} , ^V? ) of the vector reference voltage (Vref) ) in the first two-dimensional landmark

and a comparison of the absolute values of said components ( ^{v "} , ^V? ),

a step of calculating (44) trigonometric values (Sk, Ck) of an angle made by the bisector of a sector (S1-S6) in which said vector reference voltage (Vref) is located, with the abscissa (a) of the first two-dimensional mark (αβ) as a function of the variables (sgn _a , sgn _p , sgn _aP ) determined previously,

a calculation step (47) of components of the corrected three-phase reference voltage based on said previously calculated trigonometric values (Sk, Ck).

2. Control method according to claim 1, characterized in that said step of calculating (47) the components of the corrected reference three-phase voltage comprises a step of determining (45) a transformed vector in a second two-dimensional reference frame according to said trigonometric values (Sk, Ck).

3. Control method according to claim 2, characterized in that said calculating step (47) of the components of the corrected reference three-phase voltage comprises the saturation (46) of the transformed vector.

4. Control method according to any one of claims 1 to 3, characterized in that said step of determining three variables (sgn _a , sgn _p , sgn _aP ) comprises the comparison (43) of the absolute value of a component ( ^V? ) of the reference voltage vector (Vref) in the first two-dimensional coordinate system (?) with another component ( ^v? ) of the reference voltage vector (Vref) in the first two-dimensional coordinate system (?) multiplied by V3 .

5. Control method according to any one of claims 1 to 4, characterized in that said step of calculating (44) trigonometric values (Sk,

Ck) comprises calculating a cosine value (Ck) and a sine value (Sk) of said angle.

6. Control method according to claim 5 when it depends on the 4, characterized in that said calculation (44) of said sinus value (Sk) is performed regardless of the sector (S1 -S6), by a single equation function of the component ( ^V? ) of the reference voltage vector (Vref) in the first two-dimensional reference (?) and said comparison.

7. A control method according to claim 6 or 5 when it depends on the 4, characterized in that said calculation (44) of said cosine value (Ck) is performed regardless of the sector (S1 -S6), by a single equation function of said other component ( ^{v "} ) of the vector reference voltage (Vref) in the first two-dimensional reference (° ^) and said comparison.

An inverter control device (20) comprising:

means for receiving a three-phase voltage setpoint (v _a , Vb, v _c );

transforming means (29) of said three-phase voltage setpoint (v _a , Vb, v _c ) into a reference voltage vector (Vref), defined by two components ( ^{v '} , ^V? ) in a first two-dimensional reference ( ° ^); a plurality of ectors (S1-S6) being defined in said first two-dimensional mark

each sector (S1 -S6) corresponding to a portion of space of the two-dimensional coordinate system (αβ) formed between two instantaneous voltages (V1-V6) of the inverter (20) angularly adjacent;

correction means (40) for the reference voltage vector (Vref); and control means of said inverter (20) as a function of a corrected reference three-phase voltage supplied by said correction means (40); said correction means (40) of said reference voltage vector (Vref) being adapted to:

- to determine three variables (sgn _a , sgn _p , sgn _aP ) with three possible values (- 1, 0, 1) according to the signs of the components ( ^v ", ^V? ) of the vector reference voltage (Vref) in the first two-dimensional landmark (° ^)

, and a comparison of the absolute values of the said components ( ^v

calculating trigonometric values (Sk, Ck) of an angle made by the bisector of a sector (S1-S6) in which said vector reference voltage (Vref) is located, with the abscissa (a) of the first reference mark two-dimensional (αβ) according to said variables (sgn _a , sgn _p , sgn _aP ),

calculating components of the corrected reference three-phase voltage as a function of said calculated trigonometric values (Sk, Ck).

Electrical assembly comprising a three-phase voltage inverter and a control device according to claim 8.

10. Motor vehicle comprising an electrical assembly according to claim 9.