FR2573256A1 - Method and device for dealing with earthing faults, in particular in a medium-voltage three-phase network - Google Patents

Abstract

The neutral NMT is normally connected to earth through a low impedance ZNT. In response to an earthing fault in one of the medium-voltage phases, the said impedance ZNT is increased abruptly in order to make it subsequently change gradually towards a Petersen coil. If the fault disappears, the impedance ZNT is returned to a low value. In the contrary case, the medium-voltage network NMT is isolated.

Description

Method and device for dealing with earth faults, in particular in a three-phase medium voltage network.

The invention relates to the operation of medium-voltage three-phase electrical networks. It is more particularly concerned with the elimination of earth faults which may occur in such a network, while avoiding, as far as possible, cutting the load on the network by the series circuit breaker provided for this purpose.

There are several categories of ground faults: - so-called self-extinguishing faults, which do not trigger the medium-voltage series circuit breaker, and constitute 5 to 15% of cases; - transient faults, eliminated by a rapid triggering, of a few tenths of seconds which constitute 70 to 80% of cases, 50% of which concern a single phase; - semi-permanent faults, eliminated by slow triggering of a few seconds, which constitute 5 to 15% of cases, including 5% of purely single-phase faults; - permanent faults, which require intervention on the network and numerous interruptions; these latter defects constitute 3 to 8% of cases.

In addition, an earth fault can constitute the starting point for an evolving fault, which is in principle neither self-extinguishing nor permanent, and concerns 20 to 25% of cases.

Apart from self-extinguishing faults, the present invention relates to single-phase fugitive and semi-permanent faults, and where appropriate certain of the progressive faults.

In addition, three cases can be distinguished, depending on the structure of the medium-voltage network and its operating mode - certain networks are isolated neutral, that is to say that the impedance between the neutral terminal and the earth is practically infinite. The fault current is then closed by the parasitic capacities which the network presents with respect to the earth, and which are equivalent to a single so-called zero sequence capacity.

- in other networks, the neutral is earthed by an impedance, not very reactive, and weak compared to the zero sequence parasitic capacity.

- finally, we also know how to put the neutral to earth by an inductance, called a Petersen coil, which is tuned with the zero sequence parasitic capacitance to form a plug circuit at the network frequency.

In the latter case, the Petersen coil can often cancel ground fault currents on its own. It therefore eliminates semi-permanent faults, without the need to open the circuit breaker mounted in series on the medium voltage network.

In the first two cases, auxiliary circuit breakers mounted in parallel on each phase are used. Their conduction makes it possible to temporarily derive the current conducted by the parasitic zero sequence capacitance, and if necessary by the aforementioned low impedance. Thus, the current which passes through the fault - and maintains it - is fleetingly canceled, without it being necessary to open the series circuit breaker of the medium voltage network.

None of the three techniques which have just been described is entirely satisfactory.

Indeed, the development of medium-voltage networks having both aerial and underground sections leads to zero sequence parasitic capacities which are large and scalable, and therefore difficult to compatible with both isolated neutral operation and that with coil operation. from Petersen. In addition, these two techniques are ill-suited to detecting and locating faults.

There remains the grounding of the neutral by a low impedance, which is the only one to allow detection and localization of the fault. On the other hand, it requires the use of a "shunt" circuit breaker in parallel, to eliminate semi-permanent faults. This shunt circuit breaker in turn has a drawback, which is to transform a possible second earth fault into a high-value two-phase fault.

The Applicant therefore posed the problem of finding means making it possible to eliminate earth faults, in particular fugitive and semi-permanent faults, without cutting the load on the medium voltage network, whatever the value. or the evolution of the homopolar parasitic capacity, while preventing a second earth defect from having serious consequences.

The invention also aims to allow in this same context the detection and localization of the initial fault.

For this, according to one of its aspects, the invention offers a method for treating earth faults, in particular in a medium voltage electrical network, the neutral of which is normally connected to earth by a low impedance.

According to the proposed method, in response to an earth fault, the following steps are implemented: a / said impedance is suddenly increased, so that it then progressively evolves towards a Petersen coil tuned with the parasitic capacitances presented by the network. with respect to earth, b / the said impedance is reduced to a low value if the fault disappears, while the medium voltage network is cut off from its supply only if the fault does not disappear after a predetermined time.

Preferably, before the disjunction, the method comprises at least one attempt to temporarily force said impedance to a low value in order to clear the fault, this attempt being followed by a return to the vicinity of the granted impedance of the Petersen coil. , if the fault does not disappear.

According to another aspect of the invention, in operation a /, the impedance is controlled to remain close to the tuning value of the Petersen coil for a selected time, subject to the extinction of the fault. Impedance control is based in particular on measurements of at least one of the following electrical parameters: current passing through the impedance, voltage across the impedance, as well as their changes since the start of the operation at/.

In known networks with low neutral-earth impedance, it is known to measure zero sequence currents in each Medium Voltage feeder and / or line portion. Very advantageously, the method of the invention provides for comparing each of these zero sequence currents with a reference value defined according to the inverse curve representing, as a function of the value of the fault current, the response time separating said abrupt increase d 'impedance of the start of the fault; the faulty M.T. (and / or portion of line) feeder is considered to be that whose measured current is equal to, or closest to, the reference value.

The invention also offers a device allowing the implementation of the above method. This device comprises means with variable impedance and abrupt switching, capable of being connected between neutral and earth, and comprising a circuit with variable impedance, the reactive component of which is mainly inductive, and having terminals of neutral and earth for said connection, and a switching circuit, interacting with the variable impedance circuit, to pass it from a normal state of low impedance to said variable impedance state.

Preferably, the variable impedance and abrupt switching means comprise a medium voltage / low voltage transformer, the primary circuit of which defines the variable impedance circuit, the latter being obtained at least in part by cross saturation of the magnetic circuit of this transformer , while the secondary circuit of the same transformer contains the switching circuit.

 Very advantageously, the magnetic circuit of medium voltage / low voltage transformer comprises an open circuit in laminated iron, which closes, through two air gaps, on a return core carrying an auxiliary winding suitable for controlling the saturation of the magnetic circuit , by applying to this winding an adjustable direct current; while the primary medium voltage winding is mounted between the neutral terminal and the earth terminal, the secondary low voltage winding is closed on a limiting circuit breaker, which allows the sudden switching of said impedance, d 'A low normal value to a higher value, while the direct current applied to the auxiliary bearing allows to vary the inductance presented by the primary winding of the transformer.

Very advantageously, the return core of the magnetic circuit is of generally cylindrical shape.

In a first embodiment, this return core is provided at its two ends with corresponding notches allowing the housing of at least part of the auxiliary winding, at the same time as the delimitation of the saturation zones created by said auxiliary winding. These notches can be distributed at different heights from the return core.

As a variant, or in addition, the turns of the auxiliary winding pass into different radii of the cylinder that constitutes the return core.

On the other hand, the return core can consist of a plurality of stacked toroids, and provided with respective windings.

The return core may also consist of a plurality of toroids, provided with respective coils, and arranged in a configuration having an axis of symmetry parallel to the axis of the alternating magnetic flux produced by the open circuit in laminated iron mentioned above. .

Other characteristics and advantages of the invention: will appear on examining the detailed description below, as well as the appended drawings, in which - FIG. 1 very schematically illustrates an example of the general structure of a medium network voltage; - Figure 2 is a perspective view schematically illustrating a device for the implementation of the invention; - Figures 2A to 2C are simplified diagrams to better understand the operation of the device of Figure 2; - Figure 3 is the equivalent electrical diagram of the device of Figure 2; - Figures 4A and 4B illustrate a first embodiment of the return core of the device of Figure 2; - Figures 4D to 4E illustrate certain electrical characteristics of the return core of Figures 4A and 4B ;; - Figures 5A and 5B illustrate an alternative embodiment of the return core; - Figures 6A and 6B illustrate another alternative embodiment of the return core; - Figures 7 and 8 illustrate two other alternative embodiments of the return core; - Figure 9 is a simplified single-line electrical diagram of a medium-voltage network illustrating the role of zero sequence current detectors; FIG. 10 is a diagram illustrating an inverse curve representing, as a function of the value of the fault current, the response time of the device of the invention; FIGS. 11 and 12 are diagrams illustrating the implementation of the method according to the invention, respectively depending on whether one ends up or does not lead to the elimination of the earth fault current.

The appended drawings are, at least in part, bearers of information which they alone can fully define. Consequently, they are incorporated into the description to contribute not only to a good understanding of it, but also to the definition of the invention, if applicable.

In FIG. 1, a three-phase medium voltage source SMT is defined by three individual sources SP1, SP2 and SP3, connected in star form from the neutral NMT. It supplies respectively three medium voltage phases denoted MP1, MP2 and MP3. These phases, which constitute the medium voltage network
RMT, supply user loads, denoted CU. A DS series circuit breaker is defined by switches DS1, DS2 and DS3, respectively interposed on phases MP1, MP2 and
MP3, near their respective sources.

Neutral NMT is connected to earth by a homopolar parasitic equivalent capacity CNT which, as already indicated, represents all the parasitic capacities of the three phases of the network with respect to earth.

The invention is concerned with networks of the type in which the neutral NMT is further connected to earth by an impedance
ZNT, which is according to the prior art a fixed impedance of low value.

When a DMT earth fault appears on one of the phases, here the MP3 phase, the prior art provides for parallel circuit breakers or shunts such as DP3, which allow the extinction of the fault, at least when it is semi-permanent type.

More specifically, in France, single-phase fault currents on overhead networks are limited to 300 amps by a ZNT impedance which is a resistance of 40 ohms. This limitation prevents dangerous earth surges, while allowing fault detection. When present, the shunt circuit breaker makes it possible to improve the quality of service, by almost canceling the fault current, hence the possibility of self-extinction of many arcs.

However, during the operation of this shunt circuit breaker such as DP3, the healthy phases are brought to the compound voltage. This increases the probability that a second earth fault will appear, which will immediately degenerate into a two-phase fault. this results in a risk of potential rise in medium-voltage and low-voltage earths during the second fault, since the current, not unlimited at the medium-voltage substation, can then be much greater than 300 amps.

In addition, during the closing of the DP3 shunt circuit breaker, a power of approximately 3.6 MW is dissipated in the 40 ohm resistor.

The risks associated with a possible second fault, as well as the importance of the power dissipated in ZNT = 40 ohms, lead to limiting the "shunt blow" to a few tenths of seconds. This shows many of the drawbacks of the technique used up to now in the networks with which the invention is concerned, it being further observed that certain faults could be eliminated by extinction, (that is to say a blow of shunt ) of longer duration.

 In networks other than those to which the invention relates, the techniques already mentioned of the Petersen coil and of the isolated neutral are known.

In both cases, fault detection is very difficult, and cancellation of the fault current is imperfect, due to the residual capacitive current (through the CNT capacity), which is of the order of 3 amperes per kilometer for an underground network, and 5 amps per hundred kilometers for an overhead network. This leads to at least 10 amps, approximately.

In the case of the isolated neutral, the residual current maintains the faults, because the self-extinction threshold of a fault is of the order of 10 amperes.

In the other case, where the ZNT impedance is a fixed impedance tuned in a Petersen coil, the residual current is in principle zero. it is otherwise in practice, because the development of the underground parts makes it very difficult to maintain the agreement between the coil and the zero sequence parasitic capacitances of the medium voltage network.

As already indicated, the Applicant has therefore turned to a medium voltage network of the type in which the neutral is normally connected to earth by an impedance of low value.

According to the invention, this impedance is a variable impedance. Before describing in detail the process associated with it, it is necessary to consider the device that r-6n will interpose in place of the impedance ZNT between the neutral and the earth.

After this interposition, it is naturally no longer necessary to provide a shunt circuit breaker such as DP3.

The device of FIG. 2, which is capable of constituting a means with variable impedance and of abrupt switching, is presented as a medium voltage / low voltage transformer.

it can be broken down into a variable impedance circuit, defined by the primary circuit of the transformer, in cooperation with its magnetic circuit, and into a switching circuit, defined by the secondary of the transformer.

The magnetic circuit of the transformer, generally designated by 10, is illustrated in the form of a laminated iron frame and in a general U shape. The bottom of the U is designated by 11, while its branches consist of yokes 12 and 13 , which extend substantially parallel to each other. Of course, any transformer equivalent to that described may be used, the person skilled in the art knowing the technologies applicable in the matter.

On the part 11 of the magnetic circuit, there is a secondary winding 22, low voltage, wound near this part 11, as well as a primary winding 21 of high voltage, wound more outside. The winding 21 has a terminal intended to be connected to the medium-voltage neutral
NMT, and another terminal intended to be earthed.

The low voltage winding has two output terminals 23 and 24 connected to the terminals of a "circuit breaker-limiter" 25, suitable for operating at low voltage, and which can be combined with a remote control 26. As is known to man art, a "circuit breaker-limiter" is a low-voltage circuit breaker performance.

Between the yokes 12 and 13, the magnetic circuit closes, through at least two air gaps 14-1 and 14-2, on a return core 15, carrying an auxiliary winding 16 allowing the control of its saturation. For this control, a control member 17 applies to the auxiliary winding 16 a direct current adjustable in intensity and / or in voltage.

In short, the function of the limiting circuit breaker 25 is to allow sudden switching of the reduced impedance between the neutral and the earth, from a low normal value to a higher value. For its part, the control member 17 makes it possible to produce in a controlled manner cross-saturation of the magnetic circuit of the transformer. FIG. 2C thus shows that the alternating flow fa which follows the general U shape of the part 10 crosses with the flow <Pfc due to the direct current applied to the winding 16
The illustration given in FIG. 2C takes account of the fact that the core 15 is here of generally cylindrical shape. Materially, it consists of rolled sheets, suitably oriented to limit the eddy currents.

FIG. 2A very schematically illustrates the structure of the magnetic circuit, as a whole. It shows that the U-shaped part 10 has a magnetic impedance Lm. We can likewise associate an impedance Le with the air gaps 14-1 and 14-2. Finally, the return core 15 has a magnetizing impedance Lrr which is variable as a function of the direct control current. FIG. 2B illustrates the same magnetic circuit in top view, where the part 10 is illustrated in dashed lines. It allows the length 1 of the continuous magnetic path, shown by a circular arrow, to appear in the core 15.

The equivalent electrical diagram of the device in FIG. 2 can then be defined as in FIG. 3. it comprises: - the resistance (series) R21 of the primary winding 21, followed - in parallel with three inductances, which correspond to the magnetizing impedances Let Lr and Lm already mentioned. (FIG. 3 also shows the effect of cross saturation between the impedance Lr and another impedance Lc, which is the inductance proper to the control winding 16); and - finally, still in parallel, we note the resistance R10, which corresponds to the losses in the iron of the magnetic circuit
On the secondary of the medium-voltage / low-voltage transformer, there appears, in series - the inductance L20, which is the leakage inductance between the windings 21 and 22; - resistance R22 of. the secondary winding 22; - and the series switch of the low tensinn circuit breaker 25.

In practice, we make sure that - the magnetizing impedance Le of the air gaps is much lower than the magnetizing impedance of the U-shaped part 10 of the magnetic circuit. The minimum value 'of the inductance Lr of the return core is chosen so that it is lower than the impedance Le of the air gaps, or of the same order as this. The adjustment is here obtained by varying the w / h ratio. At the other extreme, the maximum impedance Lr is chosen to be close to the magnetizing impedance Lm
When the circuit breaker 25 is closed, the impedance brought back by the device of figure 2 between the neutral and the earth is of low value, and mainly resistive, because the resistance R22 is then put in parallel on all the inductors located at the transformer primary.

On the contrary, when the circuit breaker 25 will open, the impedance increases suddenly, since the elements L20 and R22 then have no effect. It is already quite strongly inductive, and, by varying the inductance Lrt we can change the overall impedance presented by the device, in order to obtain the Petersen coil tuning.

 In order to make the control of this inductance more precise and / or easier, it is possible to envisage different embodiments of the return core 15.

FIG. 4A shows a first embodiment, where the core 15 retains its generally cylindrical shape, constructed from rolled sheets. But we observe that now the ends of the cylinder are provided with notches such as 150. FIG. 4B shows how it is possible, in these notches 150, to implant turns of the winding 16, which can be put in series, or ordered separately. FIG. 4B also shows, at the level of the two notches on the left, the shape which the magnetic flux lines can take on passing the notches. Correlatively, at the two right notches in FIG. 4B, a saturation zone ZS has been illustrated which can be obtained by sufficiently increasing the current in the turns passing through the notch located opposite.

It is observed that the saturation of the return core by a direct current, and following a path perpendicular to the alternating flux, makes it possible to reduce the magnetizing inductance Lr.

in a substantially linear fashion. Only the length l (FIG. 2B) of the continuous flux path determines the number of ampere-turns necessary for the saturation of the return core.

It is also observed that the transformer is advantageously high and narrow, in order to control significant variations of the magnetizing impedance Lr with a low direct current. However, in return, the own inductance Lc of this control winding increases with the height.

On its side, the circuit breaker-limiter 25 can be defined as follows: in normal operation, it trips for a hundred low voltage amps, which correspond substantially to a medium voltage imbalance current of the order of 1 amp. In the event of a frank single-phase fault, the circuit breaker-limiter can open in a few milliseconds, and limit the medium-voltage current to a peak of 100 amps, for example.

In the event of a more resistant fault, the reverse curve of the circuit breaker causes the opening in a few tenths of a second, or more if necessary. The network is then substantially in the condition of isolated neutral.

The controlled saturation of the return core then makes it possible to change the impedance brought in parallel between neutral and earth, towards the value corresponding to a
Petersen.

The characteristic "alternating inductance Lr / continuous control current" is defined by the distribution of ampere-turns in the volume of the magnetic circuit.

In the case of the control mode illustrated in FIGS. 4A and 4B, with dimensions 1 and h of the same order, the results illustrated in FIGS. 4C to 4E have been obtained. FIG. 4C recalls the notations, 1c defining the control current passing through the winding L r while 1a designating the alternating current passing through the winding Lr
FIG. 4D shows the evolution of 1a as a function of time, for a control current 1c of the form illustrated opposite in FIG. 4E. it appears first of all that the response time is very short, the value of the own inductance Lc being low. We also observe that the variation of the impedance Lr is of the order of 3.

Figure 4D also shows that there are practically no harmonics in the alternating current. This fact is important because it is desirable to cancel not only the components at 50 hertz of the fault, but also those which it can present at harmonic frequencies.

Again considering FIGS. 4A and 4B, it appears that the notches can be of different shapes from each other, in order to progressively pinch the alternating flow with the continuous flow.

FIGS. 5A and 5B illustrate a first alternative embodiment of the return core 15. In FIG. 5A, it can be seen that the notches are produced in the form of housings defined in the intermediate part of the cylinder 15, and provided with access slots towards the end. FIG. 5B also shows that these housings 151-1, 151-2 and 151-3 can be located at different heights with respect to the Ecuadorian axis of the cylinder 15. This same figure shows the flow lines in the vicinity of the housings 151-1 and 151-2, as well as a saturation zone ZS, produced mainly due to the current passing through the winding located in the housing 151-3.

This figure also shows that part of the winding 16 can remain wound on the external surfaces of the cylinder 15. This aspect of the invention is applicable to the other variants described.

FIGS. 6A and 6B illustrate a second variant embodiment of the return core. here it retains its general cylindrical shape, but the turns of the control winding 16, which pass on the one hand through the central hole of the cylinder, on the other hand return according to different radii thereof, passing between the sheets.

Note for example that the turns 16-1 go to the outer surface of the cylinder, while the turn 16-2 covers a small radial part of the cylinder, and the turn 16-3 a larger radial part of the latter , These coils can be ordered separately or together.

When, for example, the current flowing through the turns such as 16-2, which go up to a radius r, is sufficient to produce saturation, we then obtain a saturation zone ZS which is an annular cylinder starting from the interior of the core 15, and extending to radius r, as illustrated in FIG. 6B.

This achieves, in another way, the controlled "nipping" of the alternating flow.

The reluctance of the return core can also be made variable, and this in different ways. In this case, the core consists of separate toroids provided with control windings which can be common or separate.

FIG. 7 shows a stack over a height h, of several tori 15-1, 15-2 and 15-3. These tori are coaxial.

FIG. 8 illustrates another embodiment, where instead of being placed in series, as in FIG. 7, the toroids are now placed in parallel. More specifically, a central toroid 15-10-and a series of tori regularly distributed around it, noted 15-11 to 15-14, are provided.

These peripheral tori are externally tangent to a circle of radius r constructed from the center of the torus 15-10.

Here again, by an individual or global control of the toroids, one can act on the reluctance offered by the core back to the passage of the alternating flux
Of course, other arrangements of the toroids are possible, in particular. series / parallel fittings.

When the toroids are ordered separately, or separately, saturation is applied gradually, by switching and / or adjusting the current passing through each torus.

 Another solution is to simultaneously order all the toroids in the same way. This command nevertheless provides a sequential saturation of the individual toroids, due to the slight dispersion of their magnetic or electrical characteristics which will be controlled in practice (magnetic section or number of turns of each torus).

Those skilled in the art will understand that the function of the return core is to create "throttles" of the alternating flow, starting from saturated zones generated by the continuous flow.

The extent of these saturated zones depends on the control current. Any combination of the embodiments which have just been described, and / or any other means of varying the inductance Lr without generating disturbing harmonics, is of course conceivable for the implementation of the invention.

Figure 9 illustrates with a single line diagram the different departures of a medium voltage substation. We find the source SMT, with the neutral NMT and the impedance ZNT, which is now produced according to the invention. After the SMT source, we recognize the DS series circuit breaker, which supplies the JB busbar, as well as all the medium voltage line feeders, such as MP10 to MP14, each fitted with an individual series circuit breaker DS10 to DS14. In known manner, it is also possible to provide each line, downstream of the circuit breaker, with a zero sequence current detector such as DH10 to DH14.

Besides this, some of the lines, like MP14 here, can once again be subdivided into ramifications such as MP140, with its zero sequence detector DH140, which in turn is subdivided into MP141 and MP142. The vertical line in Figure 9 illustrates the medium voltage JB busbar of the substation, which provides. The supply of individual feeders.

When a single-phase earth fault affects a feeder, all the other feeders will be the seat of a homogeneous current due to the capacities of the lines with respect to the earth.

In certain cases, these parasitic currents can be considerably higher than the ampere, especially if the departure has a long underground section.

At present, a zero sequence current detector is not able to differentiate a very resistant earth fault current from a capacitive current due to a fault on another feeder.

it is otherwise with the material which has just been described.

Indeed, the characteristics of the device of FIG. 2, and in particular of its low voltage circuit breaker being determined in advance, the earth fault current will be cut by the low voltage circuit breaker according to a well defined inverse curve, which depends on these characteristics.

This curve is illustrated diagrammatically in FIG. 10. It will make it possible to characterize and locate the fault in the following manner - we note td the time it takes for the low voltage circuit breaker to open, measured from the start of the fault. During this time, the different detectors, as illustrated in figure 9, will have measured different respective values of the zero sequence current, which we note - if a detector sees a zero sequence current 1d whose duration td corresponds to what the curve gives reverse, then it is on the faulty start ;; - if, on the contrary, a detector sees a zero sequence current Id for a time much lower than the corresponding time td on the curve in FIG. 10, then the current which is the source of the fault is greater than the current measured by the detector, and that the fault affects another start.

Those skilled in the art will understand that the device which has just been described allows the implementation, in all its variants, of the method according to the invention as defined above.

Figures 11 and 12 will now allow a better perception of the implementation of the method, on the basis of two particular examples.

In figure II, figure 11A represents the effective value Ieff of the fault current, figure 11B represents the effective voltage Veff of the neutral, figure 11C represents the open or closed state of the low voltage circuit breaker and figure 11D represents the current 1c control applied to the winding 16, these three figures having the same time scale. Figure 11 shows a sequence effectively leading to the elimination of the earth fault current.

Figure 12 does the same, with the same notations, but in the case of a sequence which does not lead to the elimination of the fault, and will therefore require a cut of the medium voltage by the series circuit breaker.

In FIG. 11, the fault appears at time to, marked by a sudden change in the fault current Ieff from a value substantially zero up to the value Id. At the same time, the voltage of the neutral Veff rises to a value V1, which is relatively low, because the impedance ZNT then has a low value. The level thus defined has the duration td already mentioned. At the end of this level, the low voltage circuit breaker opens (Figure 11C), which has the effect of increasing the voltage of the neutral. up to V2, at the same time as lowering the effective intensity of the fault current to a residual current value Idr.

In other words, the fault current is derived, without pulling out, towards the zero sequence capacity of the network, whose higher impedance brings up the neutral voltage.

A priori, no transient overvoltage can occur.

The increase in impedance also has the effect of decreasing the current in the fault, but the fault may still remain above the auto-extinction threshold.

From this moment, a DC voltage is applied to the control winding 16. The control current 1c increases regularly due to the high inductance of the winding. it will gradually tend, in a few hundred milliseconds for example, towards point A which represents the instant when the ZNT impedance defines a chord circuit agreement with the homopolar parasitic capacitances of the network (Petersen coil). The DAE time interval (Figure 11B) represents the time during which the fault current remains below the self-extinguishing threshold. This duration can be of the same order of magnitude as that of the current shunt circuit breaker (0.2 to 0.3 seconds).

If the extinction occurs, the impedance of the fault becomes negligible compared to that of the plug circuit, which is no longer tuned. The elimination of the fault results in a drop in the neutral voltage, inside the zone hatched in Figures 11A and 11B.

During this time, the control current Ic has exceeded the value which corresponds to the tuning of the plug circuit. The neutral is therefore found connected to earth via an inductor whose impedance is lower than the zero sequence capacitance of the network.

The slight imbalance in the phase-to-earth capacitors results in a residual voltage on the inductor (Vr, Figure 11B).

If this residual voltage is normal, the low-voltage circuit breaker can be closed (figure liC). Elimination of the fault is then complete.

 On the contrary, as will now be seen with reference to FIG. 12, it may be that the voltage of the neutral does not decrease sufficiently.

It will be observed that the time scale is, in FIG. 12, more condensed than in FIG. 11. With this reservation, we find again the same beginning, with an intermediate voltage
VI (FIG. 12B) which coexists with the nitial value Id of the fault current, then a higher voltage V2, which coexists with a lower value of the fault current, noted Id2.

On the other hand, during the first passage A1 at the tuning value of the Petersen plug circuit in coil, it was not possible to obtain the extinction of the fault. The control current Ic (FIG. 12D) continues to rise up to the neighborhood of its maximum value ICmax. The impedance ZNT then decreases again, which results in an increase in the effective fault current in Id3 (figure 12A, after the dot at the same time as by a slight decrease in the voltage of the neutral in V3 (figure 12B).

After reaching its maximum value, the current Ic decreases again, which causes it to go back to A2 by the tuning value of the plug circuit, hence a further reduction in the fault current, at the same time as a return of the voltage from neutral to its previous value of single voltage (figure res 12A and 12B). It is then planned to try to clear the fault. This is done by one or more closings of the low voltage circuit breaker. FIG. 12C illustrates two of them, at the end of which the voltage of the neutral still does not decrease. It is then possible to go directly to a medium voltage cut by the series circuit breakers.

However, an interesting variant of the present inven -'- tionprovoit a servo of the current Ic, crossing the control coil to the value here which corresponds to the tuning dn circuit plug. Indeed, the agreement will then result in a high impedance compared to that of the default, (except in the case of self-extinction where the agreement becomes useless). At the time of tuning, the voltage is therefore maximum, and practically equal to the single voltage Vs of the medium voltage source.

Points A1 and A2 can easily be detected, which correspond to the minimum of the effective fault current, or even to the maximum of the voltage. These points having been detected, the value of the corresponding current is measured ica
It is then possible, in the event of failure of the preceding attempts, to carry out an "forced" self-extinguishing attempt, by injecting a current of value ica into the control winding. The duration of this forced self-extinction can reach several seconds, which was practically impossible with the means of the prior art.

If this is not successful (as shown in Figure 12), a medium voltage cut must be made by the series circuit breakers.

It will be noted that the sequences described in connection with FIGS. 11 and 12 lead to the extinction of the fault current (if its characteristics allow it) whatever the value of the zero sequence capacitance, that is to say whatever the l status of the medium voltage network at the time of the fault (out of service, troubleshooting configuration for example).

More generally, the invention is remarkable by the following aspects - at the time of the fault, successively carrying out the functions of neutral earthed, then of isolated neutral, and finally of Petersen coil; - limitation of the energy dissipated in the fault, and characterization of the fault current, which facilitate its detection and location; - particular realization of these functions using a low-voltage circuit breaker-limiter, and a magnetic circuit whose structure allows its adjustment in Petersen coil; - supply of unique, flexible and scalable equipment, which it is sufficient to connect between neutral and earth, and whose control and control are easy, and very particularly possible with digital means.

In addition, the medium-voltage neutral is earthed during normal operation, which ensures that the three phases remain at single voltage.

By switching to isolated neutral operation by a breaking device - (the low-voltage circuit breaker) - the current and energy of single-phase faults are limited, which facilitates the extinction of non-permanent faults, and moreover limits medium voltage and low voltage earth potential increases, hence, in the second degree, the possibility of avoiding certain evolutionary faults.

By choosing the operating characteristic of the cut-off device, a powerful transient fault regime can be maintained, which makes it possible to detect and locate it, as well as to "burn" certain types of fault.

By operating in an adaptable and / or slave transient Petersen coil, we manage to cancel the fault current to cause self-extinction, while avoiding aggravating the consequences of a possible second single-phase fault.

Many modes of operation of the device of the invention can be envisaged. This is due, on the one hand, to the variety of settings, in particular the following - size of the low-voltage circuit breaker-limiter, - forced and "natural" self-extinguishing procedure, - fault clearing procedure, - reclosing criteria after fault, - medium voltage cut-off criteria.

On the other hand, it is possible to adopt different control and command modes for implementing the method and the device according to the invention - sequence according to FIG. 11, triggered locally by the opening of the low voltage circuit breaker; - any sequence, remotely controlled by the high voltage / medium voltage substation automatons, - sequence programmed by a local microprocessor automaton, which also manages the search for the Petersen coil tuning and the corresponding current servo-control.

 These numerous modes of operation and control of the device constitute an advantage: those skilled in the art benefit from these flexibilities to adapt and develop the device in response to needs.

Finally, those skilled in the art will understand that the present invention is not limited to the embodiment described, in particular as regards the device. It is in fact possible for it to conceive of numerous variants thereof without departing from the scope of the claims below.

In particular, to ffiieux control the evolution of voltages during transient regimes and thus remain compatible with the isolation levels, it may be advantageous to use MV and / or LV surge arresters within the device.

Such transient overvoltage protection elements can be fitted to the primary and / or secondary of the transformer.

Claims (15)

Claims. 1. Method for treating earth faults, in particular in a three-phase medium voltage network, the neutral of which is normally connected to earth by a low impedance, characterized in that in response to an earth fault a /. said impedance (ZNT) is suddenly increased so as to gradually evolve it towards a coil of Petersen tuned with the parasitic capacities (CNT) that the network presents with respect to the earth, and b /. said impedance is reduced to a low value if the fault disappears, while the medium voltage network (RMT) is cut off from its supply if the fault does not disappear after a predetermined time. 2. Method according to claim 1, characterized in that before the separation., The operation a /. includes at least one attempt to temporarily force said impedance to a low value in order to clear the fault, this attempt being followed by a return to the vicinity of the granted impedance of the Peters? n coil, if the fault does not disappear . 3. Method according to one of claims 1 and 2, characterized in that the operation a /. the impedance (ZNT) is controlled to remain close to the Petersen coil tuning value for a selected time, subject to the extinction of the fault. 4. Method according to claim 3, characterized in that the control of the impedance (ZNT) is based in particular on measurements of at least one of the following electrical parameters: current passing through the impedance (ZNT), voltage across the impedance (ZNTt, as well as their changes since the start of the operation a /. 5. Method according to one of claims 1 to 4, in which homopolar currents are measured in each start or portion of line, characterized in that each of these homopolar currents is compared to a reference value defined according to the inverse curve representing, as a function of the value of the fault current, the response time separating said abrupt increase in impedance from the start of the fault, the start or the portion of line of fault being considered as that whose measured current is equal to the reference value, or closest to it. 6. Device allowing the implementation of the method according to one of the preceding claims, characterized in that it comprises variable impedance and abrupt switching means capable of being connected between neutral and earth, and comprising - a circuit at :; variable impedance (10-17, 21), the reactive component of which is mainly inductive, and having neutral and earth terminals for said connection, and - a switching circuit (22, 26) interacting with the impedance circuit variable, to change it from a normal state of low impedance, to said state of variable impedance. 7. Device according to claim 6, characterized in that the variable impedance and abrupt switching means comprise a medium voltage / low voltage transformer, the primary circuit (10-17, 21) defines the variable impedance circuit, that- ci being obtained at least in part by cross saturation of the magnetic circuit of this transformer, while the secondary circuit (22-26) of the same transformer defines the switching circuit. 8. Device according to claim 7, characterized - in that the magnetic circuit (10) of the medium voltage / low voltage transformer comprises an open circuit (11-1-3) in laminated iron, which closes, through at least two air gaps (14-1, 14-2) on a return core (15) carrying an auxiliary winding (16) suitable for controlling the saturation of the magnetic circuit, by applying to this winding (16) a direct current adjustable, - in that the primary medium voltage winding (21) is mounted between a medium voltage neutral terminal and an earth terminal, and - in that the secondary low voltage winding (22) is closed on a circuit-breaker-limiter (25), this circuit-breaker-limiter (25) allowing the abrupt switching of said impedance, from a low normal value to a higher value, while the direct current applied to the auxiliary winding (16) makes it possible to vary the inductance presented by the primary winding of the transformer. 9. Device according to claim 8, characterized in that the return core (15) of the magnetic circuit is of generally cylindrical shape. 10. Device according to claim 9, characterized in that this return core is provided at its two ends with corresponding notches (150) allowing the housing of at least part of the auxiliary winding at the same time as the delimitation of the saturation zones created by said auxiliary winding. 11. Device according to claim 10, characterized in that the notches (151-1, 151-2, 151-3) are distributed at different heights of the return core. 12. Device according to one of claims 9 to 11, charac terized in that the turns (16-1, 16-2, 16-3) of the auxiliary winding (16) pass in different rays from the cylinder that constitutes the core back (15).  13. Device according to one of claims 8 and 9, characterized in that the return core consists of a plurality of stacked toroids (15-1, 15-2, 15-3), and provided with respective windings. 14. Device according to claim 8, characterized in that the return core consists of a plurality of toroids (15-10 to 15-14) provided with respective coils, and arranged in a configuration having an axis of symmetry parallel to the axis of the alternating magnetic flux produced by said open circuit in laminated iron. 15. Device according to claim:) 8, characterized in that the primary and / or secondary of the transformer is equipped with elements for protection against transient overvoltages.