WO2001045242A1

WO2001045242A1 - Method for controlling an ac-ac type electric power converter

Info

Publication number: WO2001045242A1
Application number: PCT/FR2000/003560
Authority: WO
Inventors: Luc Raffestin
Original assignee: Crouzet Automatismes
Priority date: 1999-12-17
Filing date: 2000-12-15
Publication date: 2001-06-21
Also published as: FR2802730A1

Abstract

The invention concerns a method and a device (4) for controlling a chopping AC-AC type electric power converter with feeding a load impedance capable of having an inductive component through a first double-control two-way switch, a second switch being provided at the terminals of the load in free wheel . Each switch comprises a pair of switch controls arranged head-to-tail. The converter switches between conductive and off state. The method consists in continuously determining the direction of the current (3) flowing in the load. Sequences of signals selectively controlling the switch controls of the switches (VC0, VC1, VD0 VD1) are generated from said determination of the current direction and of the chopping control signals (VPWM) of the alternating supply voltage. The generation of control signals (VC0, VC1, VD0, VD1) is ensured by a microcontroller or a microprocessor (40).

Description

METHOD FOR CONTROLLING AN ALTERNATIVE-ALTERNATIVE ELECTRICAL ENERGY CONVERTER

The invention relates to a method for controlling an electrical energy converter of the alternating-alternating type.

It also relates to a device for controlling such a converter for implementing the method.

It relates more particularly to the field of strong currents and / or high voltages.

To fix the ideas, we will place ourselves in the following in the case of the preferred application of the invention, namely the high power converters.

In the context of the invention, the term "converter" must be understood in its most general sense. It generally relates to devices for supplying energy to a load of the inverter, chopper or so-called switching type.

The conversion is carried out by means of switches, generally of the semiconductor type, actuated sequentially by control signals, so as to "cut off" a primary supply voltage at the rate of the control signals. The average energy supplied to a load supplied by such a converter depends in particular on the duty cycle, which will be called η, between the conduction periods and the blocking periods of the converter. We can therefore adjust this energy, and what is more, regulate it, precisely by playing on this duty cycle η.

To regulate the output voltage, it is sufficient to measure it and act on the switching control signals at the converter input by a conventional feedback process. One of the commonly used regulation methods uses so-called pulse width modulation, also known by the acronym "PWM" (for "Puse Width Modulation"). When we consider the domain of high currents and / or high voltages, even if it is, a priori, a load deemed to be resistive, the latter is not entirely free of an inductive component (if only the inductance of the wiring connecting the load to the converter) It follows that one cannot suddenly interrupt the current in the load because it would develop at the terminals violent overvoltages thereof These overvoltages risk damaging the converter switches, which are designed to support a maximum blocking voltage These sudden overvoltages can also induce parasitic phenomena on the primary voltage supply network or electromagnetic disturbances Or there are electromagnetic compatibility ("EMC") standards to be observed Also, it has been proposed to use devices making it possible to avoid this parasitic phenomenon, in particular to devices known under the name "freewheeling diode" ( or "free wheeling diode" according to the English terminology)

FIGS. 1A to 1 C, appended to this description, schematically illustrate an example of such a device and explain its operation, in the case of a DC-DC converter

It is assumed that an impedance Z consisting of a resistor R and of an inductive part, of inductance /, is supplied by a DC voltage source via a switch K The latter is controlled by a rectangular signal V _c of period T , in state "1" for periods of time and in state "0" for periods of time t ₂ The ratio between these two values defines the above-mentioned duty cycle η FIG. 1 C illustrates the evolution of the signal V _c in relation to an arbitrary time axis t

During the time interval t, the switch K is closed and the impedance load 2 is subjected to the voltage \ = A current is established / which does not instantly reach a maximum value which would be equal to VeJR but increases progressively due to the presence of the inductance / according to an exponential law, as it is well known This growth takes place during the time interval t: To avoid any phenomenon of parasitic overvoltage, place at the terminals of the impedance Z a diode D (called "freewheeling") in the non-direction conductor During the interval ti, this diode D does not play an active role since it does not conduct

On the other hand, when the control signal V _c goes to the zero state during the time interval t ₂ , the switch K goes to the blocked state Always due to the presence of the inductance /, by effect of inertia, the current in the load cannot suddenly stop II develops a voltage across the inductance / reverse polarity of the voltage \ = It follows that the diode D is now polarized in the right direction and becomes conductive The current will continue to flow in the impedance load Z and decrease towards a value close to zero, if t ₂ is much greater than ti, until extinction of the inductive energy accumulated in the inductance / The presence diode D therefore avoids any overvoltage Figure 1 C illustrates the variations of the current / as a function of time t

When it is desired to carry out, no longer a conversion of the continuous type of the type which has just been mentioned briefly, but of the alternating-alternating type, it is also possible to use switches and "freewheeling" circuits. However, the problem becomes more complex, because the input supply voltage, for example the voltage supplied by the sector, is now, by definition, alternative The switch or switches used must therefore be bidirectional, if we want to use both alternations It is the same for so-called "freewheeling" circuits. It is therefore no longer possible to use simple passive elements, such as diodes, but active controllable elements.

In the known art, these problems have been solved by using power semiconductor switches, which will be referred to hereinafter as bi-controllable bidirectional switches or switches, or more simply "IBB"

We first used power bipolar transistors, thynstors, tnacs or similar components. Since the 1980s, hybrid components such as, for example, "IGBT" ("Insulated") have been used.

Bipolar Transistor "), MOS (" Metal Oxide Semiconductor ") transistors,

"GTO"("Gum Tum-Off Transistor") or "IGCT"("Insulated Gâte Commutated Thynstor ") The characteristics of such components are described, by way of nonlimiting example, in" THE TECHNIQUES OF THE ENGINEER "volume D3I (see in particular the article by Philippe LETURCQ" Power semiconductor components eigen characters ") When the selected components cannot themselves block a reverse voltage greater than or equal to the peak operating voltage of the system, the addition of additional diodes is then made necessary

The "IGBT" component, in particular is very advantageous for high voltage power switching II combines the advantages of bipolar transistors, in particular low voltage drop in saturated state, and those of thyπstors and "MOS" components, in particular very good. voltage withstand and low consumption for the control signals It is therefore possible to control it in voltage To fix the ideas, in what follows, it will be considered, without this in any way limiting the scope of the invention, that the switches and "freewheeling" circuits are made from such components

More precisely, each switch is in reality produced on the basis of a pair of bi-controllable bidirectional switches. FIG. 2, appended to this description, very schematically illustrates a simplified example of an ACV-AC converter according to the prior art , made from "IGBTs" components More precisely, it is an alternative bridge arm

The power source at the input Ve _~ of the alternative type, supplies the impedance load Z, consisting of a resistor R and an inductive portion of impedance /, through a switch 1 in series with it. ci The latter consists of a pair of "IGBT" components, 10 and 11, arranged head to tail, that is to say in anti-parallel association, which we will call IBB

Similarly, the so-called "freewheeling" circuit must also be able to drive in both directions. It follows that it also consists of an IBB, referenced 2 consisting of two "IGBTs", 20 and 21. It is arranged as previously to load terminals Z, so that the inductive current can be looped back when it is conductive

The two IBBs are therefore arranged in cascade between terminals A and B of the alternating input voltage source V _e _ with a common point, a terminal B, which is assumed to be connected to ground.

In FIG. 2, the directions of the current l _{z have been shown} in the load Z "Direction 1" and Direction 0 "flowing in the load Z during the two half-waves of the input voltage Ve- More generally, l index "1" is associated with "Direction 1" and index "0" with "direction 0" In operation, the CONV converter will alternate between two distinct states the conducting state, during which the switching IBB 1 is passing and the "freewheeling" IBB 2 is blocked during this state, the voltage of the power source Vβ ~ is applied to the load Z, and the effective current in the load increases in a similar manner to what has been described with reference to FIG. 1C, the blocked state, or "freewheel", during which the switching IBB 1 is blocked, but the "freewheeling" IBB 2 is on during this state, the voltage of the power source V _e ~ is not applied to the load Z, and the current through the load loops on itself and decreases until comp disappears inductive energy in the load and its wiring (inductance I)

More precisely, in the on state, during a first alternation of the supply voltage V _* -, positive, it is the switching "IGBT" 10 which leads to the on state during the second alternation it is the "IGBT" switching, negative, the "IGBT" switching 11 which takes over We observe a similar chronology for the "IGBTs" of "freewheeling,

20 and 21, in the blocked state

In FIG. 2, the control signals applied to the control terminals ("dropouts") are referenced V _C o and V _C ι, for the switching IBB 1, and V _DO and V _D ι, for the switching IBB "freewheel" 2 These signals must follow a precise chronology, as will be shown below The alternation between the two states "conductor" and "blocked" thus makes it possible to regulate the average voltage applied to the cnarge Z and therefore the current which crosses it, for example by using the aforementioned pulse width modulation technique. ("PWM") The operation of the converter in FIG. 2 is therefore similar to that in FIGS. 1A to 1 C, but takes account of the two alternations of the supply voltage which result in two directions of current in the load Z, on the one hand, and in the two IBBs 1 and 2 on the other hand

In theory, the implementation of the techniques which have just been recalled from a particular embodiment gives the expected results However, in practice, the devices of the known art are not free from problems In particular , the main problem that arises is related to the transitions between the two conductor and block states mentioned above.

Indeed, if for a moment, the IBBs, 1 and 2 are both controlled (therefore conductive), the power source V _* - is short-circuited, since these IBBs are arranged in cascade and connected to the terminals of the voltage source V _e - The current will increase violently through the two IBBs, 1 and 2, which can lead to their destruction or at least limit their lifespan This will also create significant disturbances on the network food

If, on the contrary, the two IBBs, 1 and 2, are simultaneously blocked, the inductive current through the load Z is suddenly interrupted This will cause a violent overvoltage at the terminals of the IBBs, 1 and 2 which can also lead to their destruction and disturb the network In this case the "freewheeling" effect no longer plays its role

It is easy to understand that the synchronization of the switching 1 and "freewheel" IBBs 2 is critical in this system and can pose serious implementation and adjustment problems if the reaction times of the constituent components of the IBBs, 1 and 2, and the inevitable dispersion of the physical characteristics of these components

The invention therefore sets itself the aim of overcoming the problems encountered in the devices and systems of the known art and proposes a method of control of an alternating-alternating electrical energy converter using bi-controllable bidirectional switches

The method according to the invention allows transitions between the two operating states, conductive and blocked, of these bi-controllable bidirectional switches, eliminating the need for critical synchronization.

To do this, according to an important characteristic of the invention, the direction of the current in the load supplied by the alternating-alternating electrical energy converter is determined and, from this determination, control signals are generated, so to selectively control bi-controllable bidirectional switches

The main object of the invention is therefore a method of controlling an electrical energy converter of the alternating-alternating type supplying a load impedance which may have an inductive component by an alternating voltage source through at least a first switch. bidirectional bi-controllable, said converter comprising a second bidirectional bi-controllable switch disposed at the terminals of said load, in cascade with said first switch, called "freewheel", so as to allow the looping of an inductive current in said impedance charging, said switches each comprising two one-way switches arranged head to tail, said method comprising the generation of sequences of control signals transmitted to said one-way switches of said first and second switches, so as to bring them alternately into a conductive state, during which said first switch is p assant and said second switch is blocked, and in a blocked state during which said first switch is blocked and said second switch is on, characterized in that it comprises the continuous determination of the direction of the current flowing in said load impedance, and, as a function of said direction of current, in that it comprises the following sequences, by selective sending of one of said control signals a / during the transition from said conductive state to said blocked state ai) priming of at least said switch turns in the direction of current of said second switch a2) then blocking of at least said switch turned in the direction of current of said first switch, and b / during the transition from said blocked state to said state conductor b1) ignition of at least said switch turned in the direction of current of said first switch, b2) then blocking of at least said switch turned in the direction of current of said second switch It is possible preferably to also provide the following steps a / during the transition from said conductive state to said blocked state

- before step ai, blocking of said switch turned in the direction opposite to the current of said first switch, b / during the transition from said blocked state to said conductive state - before step b1, blocking of at least said switch turned in the opposite direction to the current of said second switch

Also, optionally the following steps

- after step a2, also striking of said switch (20-21) turned in the opposite direction to the current (l _z ) of said second switch (2), and - after step b2, also striking of at least said switch (10-

11) turned in the opposite direction to the current (l _z ) of said first switch

(1)

The subject of the invention is also an alternating-alternating electrical energy converter for implementing the method. The invention will now be described in more detail with reference to the accompanying drawings, among which FIGS. 1A to 1C illustrate schematically the principle and operation of a DC-DC converter provided with a so-called "freewheel" device, according to the prior art, - Figure 2 schematically illustrates a simplified example of an AC-AC converter implementing switches bidirectional bi-controllable, also incorporating a device called "freewheeling", according to known art; FIG. 3 illustrates a block diagram of a concrete embodiment of the control circuit of an AC-AC converter for implementing the method according to the invention; FIG. 4 illustrates an example of a circuit for determining the direction of current in the load supplied by the AC / AC converter, usable for the control circuit of FIG. 3; and FIG. 5 is a diagram illustrating the main signals used in the control circuit of FIG. 3.

To describe in more detail the control method according to the invention, reference is again made to FIG. 2. In fact, the method of the invention does not require modifying the architecture of the circuits constituting the AC converter - alternative CONV itself, which constitutes an additional advantage. Only the control circuits of the controlled switching components or switches, in this case the "IGBTs" 10, 11, 20 and 21, must be specific in order to incorporate the provisions specific to the invention.

We will now examine the two main transitions: transition "conductive state to blocked state" and transition "blocked state to conductive state".

1. Transition from "driver state to blocked state":

In the case of a transition from the conducting state to the blocked state, it is possible to block the switching IBB 1 in the opposite direction to the current in advance, since this direction is not used. According to the alternation considered, in the case of a current in the "Direction 0", one thus blocks I "IGBT" 11, and in the opposite case, one blocks I "IGBT" 10.

At this instant, the priming of the "freewheel" IBB 2 in the direction of the current no longer risks creating a short circuit of the supply V _e -, and can therefore be carried out in complete safety. In the case of a current in "Direction 0", the "IGBT" 20 of the "freewheel" IBB 2 is started. Then, it is possible to deactivate the switching IBB 1 in the current direction, since the priming of the "freewheeling" IBB 2 in the current direction is already carried out, which prevents any risk of interruption. of current in the load, therefore the appearance of overvoltage. In the case of a current in "Direction 0", I "IGBT" 10 is blocked.

Finally, it is possible to initiate the "freewheel" IBB 2 in the opposite direction to the current, since any risk of short-circuiting of the supply is eliminated. In the case of a current in "Direction 0", the "IGBT" 21 of the "freewheel" IBB 2 is started. 2. 2. Transition "blocked state towards conductive state"

In the case of a transition from the blocked state to the conducting state, the following sequence is carried out in a similar manner:

- blocking of the "freewheel" IBB 2 in the opposite direction to the current ("IGBT" 21 in "Direction 0"); - initiation of switching IBB 1 in the current direction

("IGBT" 10 in the 0 direction);

- blocking of the freewheeling IBB 2 in the current direction ("IGBT" 20 in the "Direction 0")

- Ignition of switching IBB 1 in the opposite direction to the current ("IGBT" 11 in "Direction 0").

In the case of a current in "Direction 1", the sequence is identical by reversing the "free wheel" IGBTs 21 and 20, and the switching "IGBTs" 10 and 11.

As described above, the system balances between two states characterized in that one of the BWIs is completely primed, while the other is totally blocked.

A variant of the method according to the invention, as it will be described, is not to initiate the "IGBTs" opposite to the direction of the current. Indeed, these will not drive anyway. The disadvantage of this variant of the method according to the invention is a delay when the direction of the current changes.

However, in return, the advantage is divided by a factor of two. the use of the control circuits of the components making up the IBBs and therefore, a priori, of extending their service life

We will now describe an exemplary embodiment of the control circuit 4 of a CONV AC-type converter of the type of that of FIG. 2 or any other similar AC-to-AC converter

In the example described, the control circuit 4 is produced on the basis of a microprocessor or a microcontroller 40 or any similar information processing system In reality, the reference circuit 40 comprises a microprocessor properly said and various circuits forming an interface, to carry out any necessary conversion of signals, in input and in output as well as circuits necessary for its good functioning clock etc, which have not been represented

This includes conventional voltage supply input terminals, "GND" connected to ground and "VCC" connected to a positive voltage, for example + 5V for the standard "TTL" The voltage + 5V can be supplied by supply circuits 41, themselves supplied, via a rectifier bridge consisting of two diodes, D _A and D _c , the anodes of which are connected to the above-mentioned terminals A and C. The operating mode of the supply circuits 41 can be various static regulator, switching power supply, etc. The only constraint is to provide a voltage meeting the standards of the standard used for microprocessor 40, the power to be supplied being a priori low

Also conventionally, the microprocessor 40 has an βι input called "ON / OFF" by turning it on and off, and the circuits it controls

For the control of an AC-to-AC converter CONV (FIG. 2) according to the known art, and also according to the invention as will be shown, the microprocessor 40 can be used to generate the control signals of the "IGBTs" , 20, 21, 10 and 11, ie Vco, Va, V∞ and V _m , respectively Indeed, in themselves, these signals are similar to those used in the known art, at least as far as characteristics such as for example their amplitudes or their shapes According to the main characteristic of the invention, the difference main relates to the pilot switching instants by a signal coming from the detection of the direction of the current IZ in the load Z

An additional input e _{2 is provided} receiving a sampling control signal, which will be called V _PW M IF, for example, the aforementioned pulse width modulation method or "PWM" is used, the signal is developed, in a conventional manner in itself from the voltage which develops at the terminals of the charges Z (between the terminals A and B) and which is reinjected in feedback The circuits allowing the detection of the amplitude of the voltage of output and development of a feedback signal are well known per se, and there is no need to re-decode them To fix the ideas, it is assumed that the control signal "PWM" proper has a chopping frequency of 20 kHz and an average duty cycle η of 50% The switching frequency is generally much higher than the frequency of the alternating voltage of the load supply source, typically 50 or 60 Hz, depending on the country, or at most some hundreds of ertz for some apps ications (avionics for example) The basic signal, with two logical states "0" and "1", defining the intervals of passing and blocked states, respectively This signal is produced from an internal or external clock not shown The first parameter (frequency) is fixed It suffices that the feedback signal V _PWM acts on the duty cycle η to obtain a desired mean output voltage value and its regulation with respect to a predetermined reference

From this control signal injected at the input e ₂ and from the sign of the input voltage V _e ~ the microprocessor 40 produces the control signals Vco, Va, V _DO and V _D -ι, according to a pre-established chronology in four times In reality, for reasons of voltage compatibility in particular, the microprocessor 40 does not directly produce these signals II is provided with four outputs Sco, Sa, SDO and S _D ι generating four signals according to the above-mentioned chronology, varying between 0V and + 5V in the example described There are, between these outputs and the control inputs of the "IGBTs" 10 to 21 corresponding, electrical isolation interface circuits, 42 to 44 To fix the ideas, we can realize such circuits based on optoelectronic couplers with high galvanic isolation. At outputs, the interfaces, 42 to 45, generate the control signals properly. said, Vco, Yι _D O and V _D respectively These signals are the temporal copy of the signals present on the four outputs S _C o Sa Sco and S _D ι of the microprocessor 40

Until this stage of the description, the control circuits 40 to 45 do not differ, a priori, from control circuits according to known art.

According to a first important characteristic of the invention, there is added to these control circuits a circuit 3 for continuous detection of the direction of the current l _z in the load Z II can be a current sensor proper or a circuit making it possible to derive this direction from other available variables, as will be shown with reference to FIG. 4

If it is a current sensor, it can be based on different physical principles magnetic coupling of the type "amperemetπque clamp" etc. In this case, the circuit 3 delivers on its output a signal V _s representing the sign of the current / _z , signal transmitted on an additional input e _sgn of the microprocessor 40 and directly usable by the latter to generate the control signals, Vco, Vd, D _O and lY, according to a chronology respecting the conditions which have been set forth below -above

Whatever the variant of the method according to the invention, there is a need to know the direction of the current l _z in the load Z This necessity can constitute an economic disadvantage, a current sensor is actually used, as according to the first embodiment which has just been described If the measurement of the current l _z is desired moreover, for purposes which are outside the scope of the invention, a current sensor allowing this measurement can also provide information which can be used to determine the direction of the current l _z in the load Z On the other hand, if such a current measurement is not necessary, the sensor in question is used only for the needs specific to the invention

Rather than having recourse to a costly current sensor, it is possible to deduce the direction of the current from the direction of the voltage across the terminals of the switching IBB 1 when the latter is conductive When the latter is blocked, the voltage at its terminals corresponds to the supply voltage V _e - and must not be taken into account A voltage detector circuit can be produced in a very simple manner using very inexpensive electronic components. A possible realization of this voltage detection function can be carried out by the electronic circuit 3 illustrated in FIG. 4, or by any other circuit. similar This solution has the advantage of being particularly economical and yet very efficient

In the example described, the electronic circuit 3 is produced based on an NPN transistor TR. A protection diode Di is arranged between the base and the emitter of the transistor TR,., Polarized in the opposite direction to the emitter-base junction. The collector of transistor 7T? Ι is charged by a resistor R ₃ , connected at its free end to the terminal "+" of a continuous supply for example of 5V (standard "TTL"), the terminal "-" of the power supply connected to the transmitter The base is connected, on the one hand, to the + 5V terminal of the power supply, via a high excursion resistor R ₂ (resistance called "pull-up" according to English terminology Saxon) and, on the other hand, to one of the terminals of the switching IBB 1, via another resistor R ,, in this case the common point B between the IBB 1 and 2. The transmitter is connected at the other terminal. C, of the switching IBB 1. The supply of electrical energy can be supplied by the supply circuits 41 (FIG. 3). The ohmic values of the different resistances and the characteristics of the semiconductor components, diode and transistor, naturally depend on the precise application envisaged, in particular on the amplitude of the voltage to be measured and the maximum voltage developed across the terminals of the BWI 1. The determination of these different characteristics and values is within the reach of the skilled person and does not need to be developed.

The detection circuit is based on the positive or negative polarization of the base-emitter junction of transistor Y via the above-mentioned "pull-up" resistor, so that the switching point of transistor T is located close to the area where the voltage across the switching IBB is zero The output voltage V _s on the collector of transistor TR, then varies between 0 and + 5V (in the example described), depending on the direction of the current in l 'IBB switching 1. In fact, saturation voltages "+" or "-" are measured IBB switching terminals 1, depending on whether I '"IGBT" 10 or!'"IGBT" 1 1 is in the conducting state, that is to say according to the alternation in progress of the supply voltage V _{e ~}

During the conductive state of the switching IBB 1 I information of circuit 3 (output V _s ) can be directly exploited On the other hand in the blocked state, the direction of the current before blocking must be memorized Indeed during blocking, the current in the load Z is looped in the "freewheeling" IBB 2, decreases, is canceled, but does not change direction The management of the information received by the circuit 3 is carried out by the microprocessor 40 which has circuits of internal or external storage volatile memory of the "RAM" type, registers, etc. In the two above cases, the output signal V _s is transmitted to the input e _sgπ of the microprocessor 40 The latter by comparison with the logic states " 0 "or" 1 "of the control signal, develops an internal signal stored in memory which will be called hereinafter S _c The flow diagram of FIG. 5 illustrates the variations of the main signals as a function of time represented by a horizontal axis 0 -, of arbitrary origin t = 0 synchronized with a zero of the current l _z these are naturally ideal forms, these functions being in reality disturbed by the "hash" We can moreover attenuate the disturbances by filtering the harmonics of high ranks by a capacitor (not shown) of low capacitive value placed at the terminals (A and C) of the primary power source V _* .

The curves on the upper part of Figure 5 represent the input voltage V _* . and the current l _z in the load Z, respectively The corresponding signals are both functions with sinusoidal variation, of the same period T, but shifted in time by a value ΔT, due to a phase shift due to the inductance / of the load causing a phase shift amplitude delay φ It follows that these two functions do not go through zero at the same times We can therefore consider four different cases, which we illustrated in Figure 5 in the time intervals , T- _\ to T ₄ , bounded respectively by the times r = 0, t = t, t = t ₂ , t = fe, and t = t ₄ , it being understood that this sequence is repeats indefinitely, as long as the CONV converter (Figure 2) is running. These time intervals correspond to the input voltage \ γ and current l _{z conditions} given in the table below.

In the example of FIG. 5, the current i _z is arbitrarily positive during the time interval T and negative during the time interval 7 ₃ . During these time intervals, the sign of the voltage at the terminals (A and B) of the switching IBB 1 remains constant: curve S _ιgn .

On the other hand, during the intervals T ₂ and T, the sign of the voltage across the terminals of the IBB 1 changes at the rate of the signal V _PWM : direction of the voltage V _* . for the "freewheeling" state (the IBB of "freewheel 1 being passing) and direction of the current l _z for the passing state. The curve S _iZ , representing the internal variable" direction of l _z ", samples the sign of the voltage across the switching IBB 2 at the time of the transition between the on state and the "freewheel" state. By convention, the on state is controlled by the logic "1" state of the V _FWM signals and the "freewheeling" state by the logic "0" state of these signals.

FIG. 3 also shows an auxiliary circuit, referenced 46, allowing the detection of a short circuit in the switching IBB 1 and its protection in the event of a short circuit actually detected. To do this, this detection and protection circuit 46 is placed between the aforementioned terminals C and B, the terminal C being connected to a short circuit detection input e _dcc . The power supply (+5 V) can be supplied by circuits 41.

The operation of the circuit 46 is based on the measurement of the saturation voltage of the "IGBTs" 20 and 21 in the on state. Indeed, the saturation voltage of such a component (as well as many other similar components) will exceed a predetermined threshold, depending on the component precisely used, when the current passing through it increases very strongly This is naturally the case during the appearance of a short circuit To fix ideas, the order of magnitude of the threshold whose overshoot must be detected is typically 4 V An output signal, in all or nothing , for example varying between 0 and +5 V, is transmitted to a so-called "reset" input of the microprocessor 40 The short-circuit detection causes the generation of control signals which block the two switching "IGBTs", 20 and 21 , and pass by the two "IGBTs", 10 and 11, of "freewheeling"

The operations which the microprocessor 40 must carry out, in particular generation of the various control signals, Vco, Va, V _DO and \ Z _D1 , as a function in particular of comparisons carried out on the logic states of the input signals V _PWM and e _sgrh so to respect the chronology which has been recalled, can be controlled by a program or a recorded firmware. By way of example, as is well known, a firmware can be resident in a fixed memory, of “ROM” type, or programmable, of the "PROM", "EPROM", or similar type, etc., memory (not shown) internal or external to the microprocessor 40

It is easy to see, on reading the above, that the invention achieves the goals it has set itself. It notably prevents any possibility of short-circuiting, due to the simultaneous conduction of the switching and connection IBBs. "freewheeling", as well as the appearance of undesirable phenomena of overvoltage, due, in a dual way, to the simultaneous blocking of these two IBBs. This characteristic can be obtained without having to respect a strict synchronization between the different control signals of the individual switches ("IGBTs"), but simply by detecting the direction of the current in the load and by deriving a control signal from this detection. The implementation of the method does not result in a significant increase in the complexity of the circuits used and / or in the cost of production. As has been shown, most of the control circuits may moreover be common to similar circuits used. in the known art In any event, the control circuits, specific or not to the invention, call upon technologies commonly available on the market (for example microprocessors, etc.) and inexpensive Finally, the AC converter alternative ACTUALLY does not require any adaptation to be associated with a control circuit implementing the method according to the invention the converter can have a more complex structure and include a greater number of IBBs The invention is however not limited solely to the embodiments explicitly described, in particular with regard to FIGS. 3 to 5 In particular, although the switches were supposed to be of the "IGBT" type as indicated, many other types of components can be used in the context of the invention Numerical values have been specified only to fix ideas

They essentially depend on the precise application targeted

Claims

1. Method for controlling an alternating-alternating electrical energy converter supplying a load impedance which may have an inductive component by an alternating voltage source through at least one first bi-controllable bidirectional switch, said converter comprising a second bi-controllable bidirectional switch arranged at the terminals of said load, in cascade with said first switch, called "freewheel", so as to allow the looping of an inductive current into said load impedance, said switches each comprising two one-way switches arranged head to tail, said method comprising the generation of sequences of control signals transmitted to said one-way switches of said first and second switches, so as to bring them alternately into a conductive state, during which said first switch is on and said second switch e st blocked, and in a blocked state during which said first switch is blocked and said second switch is on, characterized in that it comprises the continuous determination (3) of the direction of the current (l _z ) flowing in said load impedance ( Z), and, as a function of said direction of current (/ _z ), in that it comprises the following sequences, by selective sending of one of said control signals: a / during the transition from said conductive state to said blocked state: ai) blocking of said switch (10-1 1) turned in the opposite direction to the current (l _z ) of said first switch (1); a2) then ignition of at least said switch (20-21) turned in the direction of current (l _z ) of said second switch (2); a3) then blocking at least said switch (10-11) turned in the current direction (l _z ) of said first switch (1); and b / during the transition from said blocked state to said conductive state b1) blocking of at least said switch (20-21) turns in the opposite direction to the current (l _z ) of said second switch (2), b2) then initiation of at least said switch (10-11) turned in the direction of current (l _z ) of said first switch (1), b3) then blocking of at least said switch (20-21) turned in the direction of current (l _z ) of said second switch (2)

2. Method according to claim 1, characterized in that it also comprises the following steps

- after step a3, also striking said switch (20-21) turned in the opposite direction to the current (l _z ) of said second switch (2), and - after step b3, also striking at least said switch (10-

11) turned in the opposite direction to the current (l _z ) of said first switch (1)

3. Method according to one of claims 1 to 2, characterized in that said direction of the current (l _z ) in said load impedance (Z) is determined from the sign of the voltage developed across the terminals (CB) of said first switch (1) when in said conductive state

4. Control device for an alternating-alternating electrical energy converter for implementing the method according to any one of claims 1 to 3, characterized in that it comprises a system for processing the program information recorded (40) for the generation of said sequences of control signals (Vco, V _C ι, V _DO , V _D1 ) from an alternating chopping control signal (V _PWM ) of said alternating voltage source (\ and a signal (S _/ ) representing said determination of the direction of the current in said load impedance (Z)

5. Device according to Claim 4, characterized in that the said information processing system is carried out on the basis of a microprocessor or a micro-controller (40)

6. Device according to one of claims 4 or 5, characterized in that it comprises galvanic isolation interfaces (42, 43 44, 45) disposed between outputs (Sco, S _C ι, S _DO , S _DI ) of said information processing system (40) generating said control signals (Vco, Vcι, _D , ) of said sequences and of command inputs α of said one-way switches (10, 1 1, 20, 21)

7. Device according to any one of claims 4 to 6 characterized in that said alternating chopping control signal (V _PWM ) is a pulse signal modulated in pulse width, so as to regulate the amplitude of the developed voltage across said load impedance (Z)

8. Device according to any one of claims 4 to 7, characterized in that it comprises a circuit (3) for determining said direction of the circulating current (l _z ) in said load impedance (Z) made on the basis of a bipolar transistor switch (TRi), the base-emitter junction of which is biased using a high pull-up resistor known as a "pull-up" (R ₂ ), so that the switching point of said transistor (TRi) between conductive and blocked states is located close to the voltage developed across said first switch (1) when it is at said conductive state, and in that an output signal (V _s ), taken from the collector of said transistor (TR,) represents a first direction of said current (l _z ) in a first switching state and a second direction of said current in a second communication state, said output signal (V _s ) being transmitted on an input (e _sgn ) for detecting the direction of current (l _z ) of said tra information processing (40)

9. Device according to claim 8, characterized in that said information processing system (40) comprises storage means for storing the direction of said current (l _z ) before said transition from said state conductor of said state blocks and keeps this memorized state during time intervals for which the sign of said alternating supply voltage (\ Y) and of said current (l _z ) flowing in said load impedance (Z) are opposite, so generating a variable (Sι) taking first and second logic states, a first logic state being associated with a first direction of said current (I _Z ) and a second logic state with a second direction of said current (l _z ), said variable ( S _z ) constituting said signal representing said determination of the direction of the current (l _z ) flowing in said load impedance (Z)

10. Device according to any one of claims 4 to 6 characterized in that said one-way switches (10 1 1 20 21) of said switches (1, 2) are semiconductor components of the so-called 'IGBT' type

11. D. device according to claim 10, characterized in that said semiconductor components (10, 1 1, 20, 21) having a so-called saturation voltage when they are conductive, it further comprises a circuit (46) arranged at the terminals (CB) of said first switch (1), for measuring said saturation voltage and detecting that it exceeds a preset threshold characteristic of a short-circuit current through said load impedance (Z), and in that said circuit (46) delivers an output signal (V _dcc ) transmitted to said information processing system (40) so as to generate control signals blocking the switches (10, 11 ) of said first switch (1) and making the switches (20, 21) of said second switch (2) conductive