WO2012089966A1

WO2012089966A1 - Reactive power compensation circuit and a method of using such a circuit

Info

Publication number: WO2012089966A1
Application number: PCT/FR2011/053129
Authority: WO
Inventors: Ignace Rasoanarivo
Original assignee: Institut National Polytechnique De Lorraine
Priority date: 2010-12-28
Filing date: 2011-12-21
Publication date: 2012-07-05
Also published as: FR2969847A1; FR2969848A1

Abstract

The invention relates to a reactive power compensation circuit and to a method of using such a circuit. The circuit (3) compensates the reactive power absorbed by a variable power load (1), this compensation circuit (3) being connected in parallel with the variable power load (1) relative to an AC voltage source (2). This circuit comprises: at least one compensation capacitor (C) for delivering reactive power to the load (1); an absorbing coil (7) for absorbing surplus reactive power; a static converter (6) formed by switch-mode-controlled power transistors (T1, T2, T3, T4), this static converter (6) powering the absorbing coil (7); a harmonic suppression filter (4) for removing harmonics from the current delivered to the load (1); and, a regulating module (8, 9, 10) for controlling the power transistors (T1, T2, T3, T4) using a duty cycle (α) that is an image of the phase shift between the voltage and the current of said voltage source (2), so as to ensure that at any moment the voltage and the current of said voltage source (2) are in phase.

Description

- -

"Reactive energy compensation circuit and method implemented in such a circuit."

The present invention relates to a compensation circuit of a reactive energy absorbed by a variable power load. It finds a particularly interesting application in domestic consumption or more generally "general public". The consumption of each household home with household appliances equipped with small engines (refrigerator, freezer, dishwasher, washing machine, ...) or inductance coils (wave oven or induction) includes a often a large part of the reactive energy inherent in the operation of these devices. In general, any electrical apparatus using an alternating current involves two forms of energy: the active energy and the reactive energy, the whole constituting the apparent energy supplied by the electrical network. Active energy provides useful work in mechanical or thermal form. Reactive energy makes it possible to create magnetic fields, it supplies electrical and magnetic circuits such as capacitors and coils included notably in asynchronous motors, transformers, converters, etc.

To account for the consumption efficiency of the reactive power of an electrical appliance, a power factor commonly known as cosine phi (coscp) is used. This factor is the ratio of the active power consumed to the apparent power supplied.

A power factor close to 1 indicates a low reactive power consumption and optimizes the operation of an installation.

In the small power range, the power factor correction principle is very widely used. The general synoptic is composed of cascaded converters: a single-phase Graëtz bridge type rectifier and a storage switching converter. The charge is almost equivalent to pure resistance. The depollution of the network is obtained by the insertion of special filters. The switching frequency is well above 20kHz. The power of the device is less than lOkW. These converters generally use Mosfet transistors - - ultra fast. For example, document US5751138 discloses a conditioner for compensating the reactive energy and the harmonics of the mains current. The reactive energy is compensated by means of a "stepped" wave inverter consisting of GTO thyristors.

Also known is US5187427 which describes a static compensator of reactive energy. It notably describes an automatic regulation by capacitor bank. This compensator comprises an inverter which integrates IGBT transistors connected in pairs between the collector of one and the transmitter of the other.

Document US Pat. No. 7,335,297 describes a system with two circuits. A first thyristor-based circuit for performing the reactive energy compensation function. A second circuit for compensating harmonics is realized by means of an active filter associated with a passive filter, the assembly being parallel to the first circuit.

The systems of the prior art have many components that penalize the power consumption and make these systems cumbersome.

The present invention aims to overcome the aforementioned drawbacks by proposing a new solid compensation circuit, compact, compact and inexpensive. Another object of the invention is a compensation circuit capable of compensating reactive energy for small electrical equipment and household appliances. At least one of the aforementioned objects is achieved with a compensation circuit of a reactive energy absorbed by a load with variable active and reactive powers, this compensation circuit being arranged in parallel with the variable power load with respect to a source. alternating voltage. According to the invention, this compensation circuit, also called a conditioner, comprises:

at least one compensation capacitor for supplying reactive energy to the load,

an absorption coil for absorbing a surplus of reactive energy, - -

a static converter controlled according to a switching operation, this static converter consisting of power transistors and supplying the absorption coil;

an anti-harmonic filter for eliminating harmonic components of the current supplying the load, and

a regulation module for controlling the power transistors from a duty cycle which is an image of the phase shift between the voltage and the current of said voltage source, and so as to impose at any instant a zero phase shift between the voltage and the current of said voltage source. The static converter according to the invention is of the chopper type for converting an AC signal to an AC signal.

The compensation capacitor is a capacitor C sized to compensate for at least the maximum reactive energy required by the load. The absorption coil may be an inductance coil for example activated only to absorb any excess reactive energy within the system. The static converter may comprise two arms with power transistors for essentially supplying the absorption coil. The harmonic filter preferably attenuates any harmonic components generated by the static converter. This filter may be low-pass type and include the largest part of the compensation capacitor. With the compensation circuit according to the invention, the capacitor and the coil are designed to have a unit coscp at any time for a variable power load. This reduces energy consumption in residential and industrial homes.

The compensation circuit thus developed having only one converter, is of simplified structure and topology. Compared to the systems of the prior art, the compensation circuit according to the invention is simple to design, economical, compact, complies with electromagnetic compatibility standards and can therefore be placed inside the variable power load which is advantageously a low power appliance. This compensation circuit according to the invention can also be placed outside the load with variable active and reactive powers. In the latter case, it can then compensate a group of devices whose number is limited by the maximum compensation capacity.

According to a first advantageous variant of the invention, the antiharmonic filter is arranged at the input of the compensation circuit according to the invention, and this filter comprises a circuit RLC for Resistance-Inductance-Capacitor, the capacitor of this circuit RLC being said at least one a capacitor used for compensation. This reduces the number of components used for the design of the compensation circuit: the same capacitor is used for compensation and filtering.

With such a compensation circuit, it is not necessary to add sector filters to depollute the network current because the filter according to the invention is an integral part of the structure of the compensation circuit, resulting in increased compactness. The realization of the control-command part is totally digital thanks to a digital signal processor (DSP), thus reliable and robust.

According to a second variant of the invention, the absorption coil and the compensation capacitor are arranged at the output of the static converter. Preferably, the static converter supplies the capacitor and the coil directly. The absorption coil may be arranged in parallel or in series with the compensation capacitor.

According to an advantageous characteristic of the invention, the static converter comprises two arms whose common junction point constitutes an output for supplying the absorption coil; each arm comprising two IGBT transistors connected in series by their emitter, and each IGBT transistor being directly in parallel with a diode disposed in opposite direction. Such a structure allows for efficient conversion.

According to the invention, pulse width modulation (PWM) regulation can be carried out, the regulation module can then comprise:

an angle measuring module which determines the phase difference between the voltage and the current of the power source; - -

a module for transforming a DC voltage level differential from the previously measured angle;

a duty cycle generator which converts the previously obtained differential DC voltage level into a duty cycle; and

a control circuit for the gates of the power transistors.

And the angle measurement module can include:

a circuit for measuring a phase shift angle between the voltage and the power source current,

a circuit for transforming the phase shift angle into phase shift voltage,

a comparison circuit between this phase shift voltage and a preset reference voltage, and

a proportional-integral type corrector fed by the comparison circuit and generating a continuous signal for the duty cycle generator.

Advantageously, regulation is carried out by modulated hysteresis. In this case, the duty cycle generator may include:

an adder for adding a triangular carrier with the DC signal output as a proportional-integral corrector output, a shaping circuit which generates the control triggers of the gates of the power transistors, and

a switching circuit which appropriately distributes the control signals of the gates of the power transistors.

This makes it possible to improve the shape of the current wave delivered by the network. In this case, the harmonic distortion rate of the current is further improved and the power factor of the compensation circuit easily tends to unity.

According to an advantageous characteristic of the invention, the regulation module is configured to work with a switching frequency f _{sw of} less than five kHz, preferably between 1 and 2 kHz. In the case of modulated hysteresis control, it is the frequency of the triangular carrier that is set below 5 kHz, preferably between 1 and 2 kHz.

With such a low switching frequency and hard commutation, the IGBT transistor is highly efficient. - -

When using IGBT transistors, they are chosen such that gate resistors of these transistors have values in the range of 5 to 33 ohms so as to reduce common mode currents.

The IGBT transistors may preferably be those of the trade already having internally an integrated diode and placed in the inverted position. So the diodes used in the converter are not independent chips reported to the converter.

The compensation capacitor can be subdivided into two elements: a large part of the anti-harmonic filter (at least 95%) and a smaller part (at most 5%) placed in parallel with the absorption coil.

The regulation module can be fully supported by a DSP using three waves to be able to fully perform its function: the voltage and current wave of the network, and the current wave in the absorption coil. This DSP can:

•. regulate to have zero phase shift at all times

between the voltage and current of the power source; and

•. provide the control triggers for the drivers of the transistors; ·. ensure the zero crossing of the current without impairing

the integrity of the power switches.

It is expected a close control by generating the commands of the triggers of the drivers of the transistors, principle of the dead time at the zero crossing of the current in the absorption coil is completely ensured by the DSP. The close control is intended to prevent simultaneous opening or simultaneous closing of the power transistors.

According to another aspect of the invention, there is provided a method implemented in a compensation circuit as described above. According to the invention, the power transistors are controlled so that the voltage across the absorption coil is kept substantially equal to the product between the duty cycle and the voltage of said voltage source.

The converter then behaves like an autotransformer, whose transformation ratio is given directly by the ratio - - cyclic continuously variable. It is thus possible to continuously regulate the voltage delivered to the absorption coil, so that an automated process is carried out for a continuous adjustment of the energy absorbed by the absorption coil.

An automated process is thus performed for the continuous adjustment of the reactive energy involved.

An all-digital DSP can be used to fully emulate the control module digitally, a close control, a management of the control triggers of the gates of the power transistors, and a control of the zero crossing of the current of the absorption coil. To this end, the management of the transistor gate control triggers avoids any risk of short-circuit switches or any risk discontinuity of the current of the absorption coil. The DSP can be used to perform control functions from three waves: the voltage and mains waves, and the current wave in the absorption coil.

The static converter may comprise two arms whose common junction point constitutes an output for supplying the absorption coil; each arm comprising two IGBT transistors, connected in series by their emitter, and each IGBT transistor being associated with a preferably internal diode, head to tail. The regulation module can be configured by means of a DSP-based module so that the currents of the two arms are in perfect encroachment, that is to say that there is simultaneous conduction for a very short instant of two common anode or cathode switches.

Advantageously, the regulation module is configured so that the currents of the two arms are in perfect encroachment.

Other advantages and characteristics of the invention will appear on examining the detailed description of a non-limiting embodiment, and the appended drawings, in which:

Figure 1 is a schematic view of a compensation system according to the invention; - -

FIG. 2 is a graphical representation of the physical quantities, voltage and current, during a reactive energy compensation;

Fig. 3 is a schematic diagram of the static converter used in the compensation system according to the present invention;

Figures 4 and 5 show configurations and operating waves obtained by the qualitative analysis. The quantities indicated in these figures are those defined in FIG.

FIG. 6 is a view of the dynamic switching tracking of the IGBT transistors of the converter of FIG. 3;

Fig. 7 is a view illustrating the waves of the control voltages of the IGBT transistors and currents in the converter of Fig. 3;

FIG. 8 is a view illustrating the voltage and current waves at the terminals of an IGBT switch of the converter of FIG. 3;

FIG. 9 gives the general control diagram, analog or digital, of the compensation;

Fig. 10 is a view illustrating curves of mains voltage and mains and load currents for a coscp of 0.8;

FIG. 11 is a view illustrating curves of the network voltage and of the mains and load currents for a coscp of 0.3,

Fig. 12 is a diagram showing the equivalence between ^ ^P ' ^∞s ^ and the equivalent load ^R ' ^L

FIG. 13 gives an electronic circuit diagram representing the two main operating sequences of the converter associated with the whole of the anti-harmonic filter (L _F , C), the switches of which one is controlled in commutation during the interval of time - - and the other during the time interval ^ and the absorption coil (R _{a /} L _a ), and another scheme of this sinusoidal model set in which the converter behaves as a ratio autotransformer of transformation equal to the duty ratio a,

FIG. 14 is a diagram of an electrical circuit for the harmonic current of switching frequency f _sw , taking into account all the real elements of the assembly, that is: the compensation capacitor C, the elements of the coil of the filter harmonic (internal resistance R _F , inductance L _F ), the load (R, L), the internal impedance of the network (p, λ), _ _

FIG. 15 is the equivalent diagram of the electrical circuit of the system, referred to the frequency of the feeder network f ₀ ,

FIG. 16 is a simplified modeling diagram for calculating the output voltage of the converter, FIG. 17 is a Bode curve for ^Ih , and h is the harmonic rank associated with the switching frequency f _sw , Ih is the current absorbed by the filter, I _hr is the residual current of the harmonic of rank r able to circulate in the network (according to Figure 14)

absorbs ur

Fig. 18 is a Bode curve of ^E for the low load, the absorber representing the voltage across the absorption coil.

Although the invention is not limited thereto, FIG. 1 shows a compensation circuit 3 according to the invention applicable to a domestic electrical appliance. This device constitutes a load 1 supplied from the electrical network 2 in voltage v and current i reciprocating. Load 1 absorbs sinusoidal current i with more or less reactive energy. This load 1 has a variable equivalent load which can be characterized by an absorbed active power P and a cosine of the load argument φ.

The compensation circuit 3 according to the invention is arranged in parallel with the load 1 with respect to the electrical network 2. The input of the compensation circuit is represented by the connections E1 and E2. This means that the voltage v across the load is the same as the voltage between the connections E1 and E2. On the other hand, the current i of the electrical network 2 is divided into a current I _C harge feeding the load 1 and a current (not shown) to the compensation circuit via El.

The compensation circuit 3 comprises at the input a filter 4 illustrated schematically and comprising a coil 5 receiving the input current via El and at least one capacitor C disposed between the output of the coil 5 and the input E2. The coil 5 is characterized by an inductance L _f and a resistor R _f . The filter 4 supplies a static converter 6 which generates a voltage equal to av across an absorption coil 7; a being a duty cycle obtained from the voltage v and the current i of the electrical network 2. The absorption coil 7 is characterized by an inductance L _a and a resistor R _a . According to the invention, capacitor C represents both - - a compensation capacitor for transmitting reactive energy to the load 1, and at the same time a filter capacitor.

Advantageously, these two functions are grouped into one and the same capacitor, but they can be separated with:

a filtering capacitor C of a capacitance adapted to produce the filter 4 and arranged as shown in FIG. 1, and

at least one compensation capacitor C1 or C2 of a capacity adapted to compensate for the reactive energy of the load and disposed in a first variant C1 in series with the absorption coil 7, and in a second variant C2 in parallel with the absorption coil 7, as shown in dashed lines in FIG.

The converter 6 is controlled by a control circuit 8 which receives the duty cycle a from a duty cycle generator 9. A comparator system 10 measures in real time the voltage v and the current i of the electrical network 2 so as to transmit a voltage setpoint to the duty cycle generator 9. By convention the comparator 10, the duty cycle generator 9 and the control circuit 8 together form a control module. An appliance usually operates between two load modes: at full load identified by (I _Ma x, φΜίη), and at idle or at rest with (I _M in, cpMax) - In these two load variation intervals, the compensation circuit 3 reacts in such a way that the mains voltage v and the current output i always remain in phase. In this case, the current delivered by the network may be lower than that before the compensation of the reactive energy: at ¹ , it would be ¹ * ⁰⁰ ^.

For the full load operating point (I _Ma x, φΜίη), the compensation circuit 3 is practically turned off, so the duty cycle a is close to zero. In this situation, the filter capacitor C compensates for the reactive power absorbed by the load by providing reactive power Q _c, such that the total reactive power of the installation is zero. According to the compensation vector diagram given in FIG. 2, the current delivered by the network i is less than the charging current I _C ha _rg e - At full load, the active power absorbed is maximum and the reactive power absorbed is - - also maximum but fully compensated by the reactive energy Q _c of capacitor C.

For the point of no-load operation (I _M m, q), the excess reactive energy Q _a is dissipated in the absorption coil 7. The excess reactive energy Q _a is the difference between the reactive energy at most Q _Max and reactive energy at least Q _Min . This is done for a given value at _Max of the duty cycle. For this duty cycle, the current discharged by the network remains close

I p ² I p ² + o ²

of the one absorbed by the load i = J- instead of i = J- ^ τ ²² -, thus serves to produce only the active power of the load and remains close to that absorbed by the load,

The static converter then behaves like a variable transformation ratio transformer, represented by the duty ratio a. The structure of the converter according to the invention is illustrated in FIG.

These are two arms, series 11 and parallel 12, joined at a common point A representing the output of the converter 6. On the series arm 11, there is a switch T1 constituted by an IGBT transistor in parallel with a diode. The emitter of the IGBT transistor is connected to the anode of the diode, the collector of the IGBT transistor being connected to the cathode of the diode. Subsequently, such an assembly consisting of an IGBT transistor and a diode arranged in parallel in the manner indicated above, will be called "IGBT switch". Thus, each arm has two IGBT switches arranged in series and connected to each other by their transmitter. The arms 11 and 12 comprise two IGBT transistors T1 and T2, T3 and T4 respectively, with their internal diodes back-to-back, these transistors are placed in series and connected by their emitter. The output signal at point A is a modulated current signal as a function of the conduction and interruption of the various switches of the inverter. These IGBT switches are controlled by control signals C1, C2, C3 and C4 coming from the control circuit 8.

Figures 4, 5, 6 and 7 show configurations and operating waves obtained by the qualitative analysis. The quantities indicated in these figures are those defined in FIG.

Figure 4 shows the configuration sequences to illustrate the failover with high viability, that is to say without impairing the integrities of the power switches in the presence of the configuration. - - represented by the transistor T1 and the diode d2 to the configuration represented by the transistor T4 and the diode d3, and vice versa with E> 0 and Io> 0.

Thus, Figure 4 (a) indicates the initial sequence with Tl passing: the current then passes through T1 and d2. In FIG. 4 (b), the control of transistor T4 is set to "1", transistor T1 remains on: this leads to FIG. 4 (c). And after an interval of time At, the blocking control of T1 leads to the phenomenon of encroachment between the pair T1 and D2 and the other pair T4 and d3. This phenomenon being completed, in FIG. 4 (d): the two switches T4 and d3 conduct the current Io, and behave as a free wheel diode with respect to the load. The command at 1 of the switch T1 puts the phenomenon of encroachment between T4 and d3, and Tl and d2 (FIG.4 (e)). When Tl is conducting completely, the diode d3 is blocked because its cathode anode voltage is practically equal to -E. Setting the T4 command to 0 does not change the previous configuration which is identical to that of Figure 4 (a). The waveforms associated with this analysis are given in Figure 5.

FIG. 6 shows the sequences of configurations to illustrate the switching with high viability, that is to say without risk of destruction of the solicited switches, of the configuration represented by the transistor T2 and the diode d1 towards the configuration represented by the transistor T3 and diode d4, and vice versa with E> 0 and Io <0.

Thus, Figure 6 (a) indicates the initial sequence with passing T2: the negative sign current then passes through T2 and dl. While keeping for a time interval At the command c2 = 1, in FIG. 6 (b), as soon as the control of the transistor T3 goes to "1", the phenomenon of encroachment occurs between the switches T2 and d1, and T3 and d4. When the current in T2 and d1 vanishes, the anode-cathode voltage of d1 is approximately equal to -E, thus d1 is blocked. In FIG. 4 (c), the two switches T3 and d4 behave as a wheel diode with respect to the load, conducting the current Io and setting the control of T2 to "0", do not change the previous configuration. In FIG. 6 (d), setting the control of the switch T2 does not change the configuration of FIG. 6 (c) either. In FIG. 6 (e), the blocking control of T3 restores the phenomenon of encroachment between T3 and d4, and T2 and d1. - -

When T2 drives completely, the diode d3 is blocked because its cathode anode voltage is practically equal to -E. The configuration then returns to that of Figure 6 (a). The waveforms associated with this analysis are given in Figure 7.

For E <0 and Io> 0 and E <0 and Io <0, the same analyzes are carried out.

FIG. 8 illustrates voltage and current waves at the terminals of a switch during the sign change of the charging current, according to the value of At. On the left panel, no dead time is applied, ie At = 0, between the switching commands C1 of the transistor T1 and C3 of the transistor T3, and causing a sudden cut in the charging current. There then appears a peak voltage, for example about 800V, during a period of time of 33.82ps across Tl. On the right panel, a dead time At = l / (2f _sw ) has been set: the peak of voltage has completely disappeared and the change of sign of the current is carried out without problem. This avoids any abrupt opening of the converter during the change of sign of the charging current.

The converter according to the invention is of direct type. It can operate in pulse width modulation (PWM), according to the principle of comparison of a sinusoid and a triangular frequency signal.

Due to the commutations of the transistors of the converter, harmonic components of current are generated in the network. The main function of filter 4 is to eliminate these harmonic components. In addition, in order to reduce the converter losses, essentially switching, the switching frequency is chosen relatively low, between one and two kHz. In addition, such frequency values make it possible to select gate resistors that reduce common mode currents.

The input filter 4 of the converter, the low switching frequency and the values of the gate resistor give the compensation circuit according to the invention a very good behavior in comparison with the Electromagnetic Compatibility (EMC) standards relating to household electrical appliances. differential mode currents and common mode currents.

Laminar wiring with grounding is also provided to reduce common mode currents in the converter. - -

By way of non-limiting example, it is possible to use copper foils (thickness 0.5 mm) and insulating sheets (thickness 0.035 mm) for producing planar conductors. These drivers, sometimes called "bus bar", are very widely used on the converter arms.

FIG. 9 shows a synoptic of the compensation control making it possible to impose at any instant a zero phase shift between the waves of the mains voltage and the current output. We find the comparator system 10 and the duty cycle generator 9. There is a measurement module 13 for measuring the phase difference between the voltage v and the current i of the electrical network. The phase shift φ generated is then transformed into a voltage Vcp by a transformation module 14. This voltage Vcp is compared with a reference voltage for a zero phase shift by means of a comparator 15. The output of the comparator 15 supplies a proportional regulator. integral 16 which in turn feeds a module 17 to generate the duty cycle a.

Results obtained by simulation are given on:

- Figure 10 for a full load operating point, with P = 2000W and coscp = 0.8, the duty cycle being 0.05, and

- Figure 11 for a point of no-load operation, with P = 200W and coscp = 0.3, the duty cycle being 0.37.

In both figures, it can be seen that the final phase differences between the mains voltage and the mains current are practically zero, and the currents are sinusoidal.

For the point of operation under load, the mains current is lower than that of the load, either in peak value of 12.5 A instead of 15.4 A, or an apparent power of 2030 VA, instead of 2500 VA. The compensation circuit according to the invention is practically disconnected because the duty ratio is 0.05. This is why the network current has practically no harmonic components.

For the point of no-load operation, the grid current has small harmonic components: however, their value remains below the limits allowed by the EMC standards. The value of the grid current is slightly higher than that of the load because of joules losses in the resistive components of the coils of the filter and the absorber (5 and 7). In this case the power supplied is 778 VA instead of 667 VA. - -

FIG. 12 is a diagram illustrating the equivalence between the torque (P, coscp) and a passive element circuit (RL) of the load. The variable equivalent load is characterized by the absorbed power? and the cosine of the load argument ^.

The equations that link these two couples are:

co _r is the pulsation of the V-value supply network.

Non-limiting calculation methods of characteristic values of the circuit according to the invention will now be described.

1) - Calculation of the compensation capacitor C.

The value of the reactive power to compensate should be the greater of ^(VI mx ^sm ^ min) _e ^(VI min max ^sin ^) _ y _e i are respectively the voltage and current across the load 1. For taking into account the other inductive elements of the assembly, as in the case of active filters, a setting coefficient ^Q of the global reactive energy is given. In this case, in a purely sinusoidal mode, the value of the compensation capacitor C can be calculated from the full load, according to:

IO * V ²

2) - Current harmonic filter and absorption coil.

2-1): The elements of the filter

The switching operation of the converter generates currents harmonics whose fundamental frequency is the frequency of

f

^sw cutting of the converter. A low-order Butterworth type 2 low-pass filter is placed between the network and the converter. feeds an impedance ^Za . In Figure 13, it is noted that the capacitor C defined above is part of this filter.

For the second-order Butterworth filter, we have the following relations:

0707

VS

2 * π * f *

a ² )

(3)

In these relationships, f _cr denotes the cutoff frequency of the Butterworth filter.

The value of the absorption coil ^La will be defined for a duty ratio ^ttMax , verifying the characteristic impedance of the filter, such that:

C ^ Ma

Is:

The

C (4) The cut-off frequency of the filter f _cr will be chosen rather close to or even slightly greater than 50 Hz in view of a strong attenuation of the current harmonics for the ^switching frequency ^fsw , of the filtering of the low frequency components especially 150 Hz and 300 Hz HZ. 2-2): The higher harmonics of the network current.

Figure 14 shows the system by reporting all the passive components of the elements. In this figure, by posing:

• ω: the variable "pulsation" in the establishment of the Bode diagram · ^J : the impedance of the compensation capacitor;

• ^Zres = Ρ ⁺ ί ^Λω the internal impedance of the network;

• ^z ch = ^R + J ^LiW _: the impedance of the load; - -

• ^{Zlf = rp} : the impedance of the filter coil; he comes

-ch

¾ ^Z r * ^Z ch _{+ ¾} Z _r + Z _ch

^Z r ^{+ Z} ch (5)

/ _hr is the current of the harmonic of rank r, / ^' _h is the total harmonic current.

For the ^switching frequency ^sw , the calculation of the different elements

¼ (j2, rf _sw ) → 0

must lead us to the relationship: ^h . Note that because of the non-linear behavior of the converter, the effects of low frequency switching are not necessarily completely eliminated in the network current.

2-3): Calculation of the reactive energy.

As a first approximation, it is considered that the converter has no losses. By neglecting the effects of cutting, the equivalent diagram of the absorption coil brought back to the source side is given in Figure 15.

The choice of the absorption coil must cancel the reactive energy for given duty cycles for both limit loads (high load and low load). The equivalent impedance of the system vis-à-vis the network is written:

z _a * z _r

(- z _n + z - _r + ^Z i ^ v) * ^Z c ^c h ^h

Total ^

^{+ Z} LF) ^{+ Z} ch

(6)

The resolution of the equation ^ ^ otaie) ⁰ makes it possible to find

· A duty cycle ^ f for the high load; • ^Asup duty cycle for low load.

2-4). Calculation of the output voltage of the converter. - -

Due to the choice of the critical frequency of the filter rather close to 50 Hz, for the choice of the power transistor sizes, the antiresonance phenomenon must be perfectly controlled.

So, according to Figure 16, it comes

z _n * z _c

^Z ch * ( ^Z LF + ^~ z _a + z _c

Equations 1, 2, 3 and 4 make it possible to calculate the elements of the circuit according to the invention for the load points considered.

The harmonic distortion rate (abbreviated THD)

According to an advantageous embodiment of the invention, the preferred dimensioning of the circuit according to the invention is obtained with harmonic distortion rates (TH D) currents of less than 10%, and a peak voltage across the absorption coil less than 500V. In this case, we choose ideally:

> f _cr = 55Hz;

> k _Q = 1.4;

By way of non-limiting example, the actual components of the circuit may be the following:

> IGBT calibrations: 600V; 30A

> For the filter: C = 140 p F; L _F = 58mH; L _A = 45mH.

Which give :

• a cut-off frequency f _c = 55.85Hz;

• a maximum duty cycle: ^a Ma _X = 0.8335

Thus, Figures 17 and 18 illustrate performance response curves of the circuit according to the invention. In Figure 17, we see the curve - - Bode report - and we see that the component at 150Hz is at - i _h

15dB, 250Hz at -25dB, 350Hz at -32dB, 1000Hz at -50dB.

_ _^ Vabsorbeur [dB] In Figure 18, we see the Bode plot of the ratio - -

E

and there is a surge output converter, around 75Hz.

Of course, the invention is not limited to the examples that have just been described and many adjustments can be made to these examples without departing from the scope of the invention.

Claims

Compensation circuit (3) for a reactive energy absorbed by a load (1) with variable active and reactive powers, this compensation circuit being arranged in parallel with the variable-power load with respect to a source (2) of AC voltage ; characterized in that it comprises:

at least one compensation capacitor (C) for supplying reactive energy to the load (1),

an absorption coil (7) for absorbing a surplus of reactive energy,

- a static converter (6) controlled according to a switching operation, the static converter (6) consisting of power transistors (T1, T2, T3, T4) and supplying the absorption coil (7);

an anti-harmonic filter (4) for eliminating harmonic components of the current supplying the load (1), and

a regulation module (8, 9, 10) for controlling the power transistors (T1, T2, T3, T4) from a duty cycle (a) which is an image of the phase shift between the voltage and the current of said source (2) of voltage, and so as to impose at any time a zero phase shift between the voltage and the current of said voltage source (2).

2. Circuit (3) according to claim 1, characterized in that the anti-harmonic filter (4) is arranged at the input of this compensation circuit

(3), and in that this filter (4) comprises an RLC circuit for resistance-inductance-capacitor, the capacitor (C) of this circuit RLC being said at least one compensation capacitor.

3. Circuit (3) according to claim 1, characterized in that the absorption coil (7) and the compensation capacitor (C1, C2) are arranged at the output of the static converter (6).

4. Circuit (3) according to claim 3, characterized in that the absorption coil (7) is arranged in parallel with the compensation capacitor (C2) or in series with this compensation capacitor (Cl).

5. Circuit (3) according to any one of the preceding claims, characterized in that the static converter (5) comprises two arms (11, 12) whose common junction point (A) constitutes an output for the power supply of the absorption coil (7); each arm comprising two IGBT transistors (T1, T2, T3, T4) connected in series by their emitter, and each IGBT transistor being directly in parallel with a diode (d 1, d 2, d 3, d 4) arranged in opposite directions.

Circuit according to one of the preceding claims, characterized in that the regulating module (8, 9, 10) comprises:

- An angle measuring module (10, 13) which determines the phase difference between the voltage and the current of the power source (2);

a module for transforming a differential DC voltage level (10, 14) of the previously measured angle;

- a duty cycle generator (9, 17) which converts the previously obtained differential DC voltage level into a duty cycle; and

a control circuit (8) for the gates of the power transistors (T1, T2, T3, T4).

7. Circuit (3) according to claim 6, characterized in that the angle measuring module (10) comprises:

a circuit for measuring a phase shift angle (13) between the voltage and the power source current (2),

a circuit (14) for transforming the phase shift angle into phase shift voltage,

a comparison circuit (15) between this phase shift voltage and a predetermined reference voltage, and

a proportional-integral corrector (16) fed by the comparison circuit (15) and generating a continuous signal for the duty cycle generator (17).

Circuit according to one of Claims 6 or 7, characterized in that the duty cycle generator (17) comprises:

an adder for adding a triangular carrier with the DC signal output as a proportional-integral corrector output (16),

a shaping circuit which generates the control triggers of the gates of the power transistors (T1, T2, T3, T4), and

a switching circuit which appropriately distributes the control signals of the gates of the power transistors (T1, T2, T3, T4).

9. Circuit (3) according to any one of the preceding claims, characterized in that the regulation module (8, 9, 10) is configured to work with a switching frequency f _S w less than five kHz.

10. Circuit (3) according to claim 9, characterized in that the control module (8, 9, 10) is configured to work with a switching frequency f _sw between 1 and 2 kHz.

11. Circuit (3) according to claim 9 or 10, characterized in that the power transistors (T1, T2, T3, T4) are IGBT transistors, gate resistors of these transistors have values in a range of from 5 to 33 ohms to reduce common mode currents.

12. Circuit (3) according to one of the preceding claims characterized in that the static converter (6) is reciprocating-alternative chopper type.

13. Method implemented in a compensation circuit (3) for a reactive energy absorbed by a load (1) with variable active and reactive powers, this compensation circuit (3) being arranged in parallel between an alternating voltage source (2) and the variable power load (1) and comprising:

at least one compensation capacitor (C) for supplying reactive energy to the load (1), an absorption coil (7) for absorbing a surplus of reactive energy,

a regulation module (8, 9, 10) for controlling the power transistors (T1, T2, T3, T4) from a duty cycle (a) which is an image of the phase shift between the voltage and the current of said source (2) of voltage, and so as to impose at any time a zero phase shift between the voltage and the current of said voltage source (2);

process in which the power transistors (Tl,

T2; T3, T4) so that the effective voltage across the absorption coil (7) is maintained substantially equal to the product between the duty cycle (a) and the voltage of said voltage source (2).

14. The method as claimed in claim 13, characterized in that an all-digital DSP is used to fully digital emulate the regulation module (8, 9, 10), and the close-up command for the generation and management of the control triggers. grids of the power transistors and the control of the zero crossing of the current of the absorption coil, so that the power switches are never short-circuited or fully open simultaneously.

15. The method as claimed in claim 14, wherein the DSP is used to carry out the control command function from two waves: the voltage and current waves of the network, and the internal management function of the triggers of the transistors. the current wave in the absorption coil, so as not to impair the power switch integrals.

16. Method according to one of claims 13 to 15, characterized in that the static converter (6) comprises two arms (11, 12) whose common junction point (A) constitutes an output for feeding the coil absorption (7); each arm comprising two IGBT transistors (T1, T2, T3, T4) connected in series by their emitter, and each transistor IGBT (T1, T2, T3, Τ4) being associated with an inner diode (d1, d2, d3, d4) head to tail;

and in that, according to claims 14 and 15, the control module is configured by means of a DSP-based module so that the currents of the two arms (11, 12) are in perfect encroachment.