FR2999294A1 - Detection device for directional detection of ground fault in e.g. insulated neutral balanced three-phase network, has location module including evaluation unit of phase shift between homopolar voltage and homopolar current derivative - Google Patents

Abstract

The device has a first module (20) to provide a signal representative of homopolar voltage (V0) of a set of phases. A second module (40) provides a signal representative of derivative with respect to a time of homopolar current (I0). An activation unit activates the second module according to a detection signal (D) of occurrence of a ground fault in a multiphase medium voltage network. A location module (50) includes an evaluation unit to evaluate a phase shift between the homopolar voltage and the derivative of the homopolar current provided by the second module with respect to pi. Independent claims are also included for the following: (1) an indicator for indicating passage of ground fault (2) a relay for protecting a current line against ground fault (3) a method for directional detection of a ground fault in a multiphase medium voltage network (4) a method for protecting a current line during occurrence of a ground fault.

Description

The invention relates to directional detection of a fault in the earth, in particular in a compensated neutral network. In particular, the invention relates to a method for detecting a ground fault in a network that furthermore makes it possible to determine whether the fault is located upstream or downstream of the detection point. The method according to the invention is based on the analysis of the signal in steady state after defect. In another aspect, the invention relates to a detection device adapted to implement the above method. In particular, the directional earth fault detection device comprises means for interpreting the current and voltage signals after the transient disturbances generated by a fault in order to give it a relative location. The invention finally relates to a fault signaling device and a tripping relay comprising sensors for the homopolar current and voltage of the network and supplying the preceding detection device with the signals enabling the signaling, for example by light, or the triggering of a device for breaking the network.

STATE OF THE ART Fault detection devices in three-phase networks make it possible to trigger the protection of the loads and / or to assist in locating said faults. For example, FIG. 1 represents a diagram of a medium voltage electrical distribution network 1 which comprises a three-phase transformer 2, the secondary of which may comprise a common neutral conductor generally connected to the ground by an impedance 3; alternatively, the neutral may be isolated, that is, there is no physical connection between the neutral point of the network and the earth. The secondary is also connected to a main distribution line which feeds the start lines 4, 4 ', 4 ", some of which may have a circuit breaker or other cut-off device 5 protecting them at the beginning. ', 4', composed of overhead lines and / or underground cables, may be subject to various defects, which it is important to detect and locate in order to overcome the problems generated: power failure, degradation of insulation materials, not to mention the safety of people. A fault detection device 6 installed on the start lines 4 ', or line sections 4, can serve as a fault passing indicator, for example lighting a warning light 7; a device 6 may also be associated or integrated with a protection relay 8 adapted to control the opening of the contacts of the circuit breaker 5. Among these defects, the most common are the single-phase faults, located outside the source station, in which a phase is in contact with the ground, or the rupture of an aerial cable in case of bad weather in particular. However, between the line conductors 4 and the earth, in particular if the starting lines 4 are buried cables, can appear high capacity values 9, causing the circulation of large homopolar currents Io in case of Ground Fault 10. To avoid false detection of a ground fault detecting device 6 caused by a failure of a capacitive link adjacent feeder 9, devices and methods have been developed to distinguish whether a fault 10 is present. is located downstream of the detection device 6 'or upstream of the device 6,, 1. In addition, the ground 3 of the secondary of the transformer 2 may itself be of such a nature as to limit the amplitude of a ground fault 10: on the MV distribution networks, it is called ground fault (or sensitive earth) ) when the level of the fault current is low (and therefore hardly detectable), either because the fault resistance is high, or because the ground 3 of the neutral at the transformer of the source station comes to limit this current of fault (for example, a compensation coil or isolated neutral), or because the ground itself is resistive in nature. More sensitive measurements of the current ("sensitive earth fault") must be made for these very strong faults. The documents EP 1 475 874, EP 1 890 165, FR 2 936 319 or WO 2006/136520 propose devices and methods for detecting and locating the direction of earth faults, which, however, rely on an important sampling during the transient period of time. default. The document FR 2 936 378 proposes for its part a low sampling determination which, however, is not suitable for three-pole cables, since the current of each of the phases is not available there. Moreover, in all these cases, the identification of the presence of a fault 10 is based on the detection of overcurrents in at least one phase of the network 1, and this over a minimum duration of defect, which may be inappropriate for sensitive earth faults, or in the case of networks with neutral compensated or isolated. It thus appears that the fault detection devices can be optimized.

DISCLOSURE OF THE INVENTION Among other advantages, the invention aims to overcome the disadvantages of devices and methods for directional detection of existing earth fault. In particular, the directionality principle implemented is based on the analysis of the homopolar current parameters in the steady-state after the fault, which nevertheless remains suitable for the networks in compensated neutral mode, and which does not require an important sampling of the signals. representative (typically, a sampling frequency less than 1 kHz, for example of the order of 500 or 600 Hz, is sufficient), which can be implemented in equipment that does not have significant memory at the software level and / or hardware. In one of its aspects, the invention relates to a method of directionally detecting a ground fault in a multiphase network, in particular with compensated neutral, using the signals representative of the voltage and the derivative of the zero sequence current of the network. , and a comparison of their relative angle, after the process has been triggered following obtaining a signal indicative of the presence of a ground fault. The method according to the invention preferably comprises a preliminary step for generating said signal indicating the presence of a fault, comprising obtaining a signal representative of the homopolar voltage of the network and its comparison with a threshold; if this step is provided, then the signal representative of the homopolar voltage can be used also for directional detection.

The method according to the invention thus comprises obtaining signals representative of the homopolar voltage, respectively of the time derivative of the homopolar current, which may be derived from the direct processing of representative signals supplied by dedicated sensors, or from a step of calculating said signals from the voltage, respectively the current, obtained for each phase of the network. Preferably, these signals are filtered, and / or sampled at a frequency below 1 kHz, in particular of the order of 600 Hz. The method furthermore comprises determining the angle between the signals representative of the voltage and the voltage. derived from the homopolar current, and preferably from its complementary 180 °; the angle thus determined is compared to a fixed parameter, in particular an angular range centered on 180 °, which amounts to comparing its complementary to a threshold. According to the result of the comparison, the method according to the invention indicates whether the defect is upstream or downstream of the point of obtaining the signals, and in particular: if the angle is close to 180 °, the defect is considered as upstream . According to a preferred embodiment of the method according to the invention, said directional detection method is associated with an actuation of a cut-off device for isolating the section from the point downstream of which a fault has been detected.

The invention also relates to a directional detection device for a ground fault, suitable for a multiphase medium voltage network, and in particular when the network is neutral compensated. The directional detection device according to the invention can be associated with current and voltage sensors which provide it with the corresponding representative signals. The directional detection device may furthermore be part of a fault-passing indicator, for example by activating light-type warning means if a defect downstream to the sensors is detected. In a particularly preferred embodiment, the directional detection device according to the invention is associated with a protection relay of the line, the warning means causing the actuation of a cutoff device of the line to isolate the section on which a fault has been detected.

The device according to the invention comprises a first module capable of providing a signal representative of the homopolar voltage of the network; preferably, the first module comprises means for receiving the signal representative of the voltage of each of the phases and summing means for supplying said homopolar voltage. The first module may comprise sampling means, in particular at a frequency below 1 kHz, and / or filtering means such as an analog filter. The device according to the invention comprises a second module capable of supplying a signal representative of the derivative with respect to the time of the homopolar current. Preferably, the second module comprises means for receiving a signal representative of the homopolar current and a shunt block; the means for receiving a signal representative of the homopolar current may comprise means for receiving the signal representative of the current of each of the phases and summing means for supplying said homopolar current. The second module may comprise sampling means, in particular at a frequency below 1 kHz, and / or filtering means such as an analog filter. The device according to the invention further comprises means for activating the second module as a function of a signal for detecting the occurrence of the fault. Preferably, the device comprises means for providing said detection signal connected to the first module and comprising a comparator of the signal representative of the homopolar voltage at a detection threshold. The device according to the invention finally comprises a locating module connected to the first and second modules, comprising means for determining the angle between the homopolar voltage and the derivative of the homopolar current, and means for comparing said angle with a fixed parameter. which is in particular a range of values centered on 180 °. Preferably, the device comprises means for determining the 180 ° complement of said angle, and the comparison means are adapted to compare said complementary to a threshold. BRIEF DESCRIPTION OF THE FIGURES Other advantages and features will become more clearly apparent from the following description of particular embodiments of the invention, given by way of illustration and in no way limiting, represented in the appended figures.

Figure 1, already described, shows an electrical network in which earth fault detection devices can be used. Figure 2 shows schematically and filtered representative signals of the currents, derived from the current and zero sequence voltages, as well as the relative angle, at the occurrence of a fault to the earth respectively downstream and to the upstream of the detection device. Figure 3 illustrates the detection method according to a preferred embodiment of the invention.

Figure 4 shows a block diagram of a ground fault detection device according to a preferred embodiment of the invention. DETAILED DESCRIPTION OF A PREFERRED EMBODIMENT The invention will be described for a compensated neutral balanced three-phase network 1, in which each line 4, 4 ', 4 "comprises three underground phase conductors 4A, 4B, 4c, the secondary Transformer 2 is connected to earth by a Petersen coil 3, and the homopolar current Io is zero in the absence of a fault By homopolar current Io (or "zero sequence current" according to the English terminology), is meant, to a possible factor of three, the vector summation of the different phase currents, or the current corresponding to the instantaneous resultant of the phase currents, sometimes called the residual current, which possibly corresponds to the ground fault current. According to the English terminology) or the leakage current It should be noted that it is possible to deviate from this situation, in particular with a non-zero homopolar current / voltage, e the network may comprise another number of phases. In addition, the neutral regime may not be compensated, and for example neutral neutral.

When an earth fault occurs on one of the phases, the current of said phase ideally becomes zero downstream of the fault and sees its amplitude increase upstream: there is therefore a current and a a homopolar voltage Io, Vo downstream of the fault 10.

However, as illustrated in FIG. 2, the frequent occurrence of a current also upstream of the capacitive connection fault 9 is noted; moreover, after a very short transient period of the order of 20 to 40 ms, the compensation coil 3 of the neutral point quickly reduces the fault current I o to less than 30 at peak, i.e. 20 A rms, value that can be difficult to detect. In particular, if the network compensation is fine enough, the transient period is too short for reliable detection, and the homopolar current Io remains very low in steady state, upstream or downstream: detection of the occurrence of said fault 10 would require a comparison with a very low threshold, leading to false positives, or the fault 10 is simply not identified by the protection 6, 8.

It should be noted that the compensation generated by the coil 3 is not perfect, and that the voltage V after defect 10 remains significant: in practice, after the transient phase, and before triggering the protection 8 upstream of the departure 4, this voltage remains of the order of several kilovolts (at least 20% of the line voltage), which is sufficient for detection in the steady state. The invention therefore proposes to use the measurement of the homopolar voltage Vo for the detection D of a ground fault, by comparing the value with a threshold So. The homopolar voltage Vo can be measured directly by an appropriate sensor 12, or calculated from the measurement of the phase voltages VA, VB, Vc also produced by a protection equipment type VPIS ("Voltage Presence Indicator System") located on the transformer 2. Moreover, although the zero sequence current Io is very low in steady state after defect 10, the invention proposes to use this stage for the relative location of the defect 10. In fact, FIG. that, depending on whether the fault is upstream or downstream of the detection point, after the transient disturbances, the phase difference between the voltage and the current is different: the phase shift of Io is in advance of n / 2 on Vo downstream of the defect 10, while the phase shift is different from n / 2 upstream. A processing of the data relating to the homopolar voltage Vo and to the current Io detected by adapted sensors 14 in steady state is used according to the invention. In particular, the invention uses the signal corresponding to the derivative of the homopolar current dIo / dt in the post-transient mode, which makes it possible: "to enlarge" the signal Io of small amplitude due to the compensation by the coil 3 of neutral point, multiplying it by the angular pulsation w of the network 1, for a better perception; to directly position this signal dIo / dt with respect to the axis of the homopolar voltage V0: in the case of an upstream fault only, the two values are on the same axis, to the phase shift errors introduced by the sensors 12, 14 near. Due to the steady-state detection, it is possible to use a lower sampling, in particular less than 1 kHz, for example 12 points per period (corresponding to a sampling frequency f -ech = 600 Hz for a network 1 of frequency F = 50 Hz), which alleviates the calculations and reduces the cost of the equipment compared to the high frequency sampling, typically from 2500 to 3000 Hz, necessary for a transient measurement at the beginning of the fault 10.

Thus, in a method according to a preferred embodiment according to the invention shown schematically in FIG. 3, in a first step, a signal representative of the homopolar voltage Vo is acquired, preferably filtered. The acquisition of the signal can be carried out directly, via a dedicated sensor 12, or through the summation of three values of the phase voltages VA, VB, Vc, which signals are also preferably filtered and sampled VAf *, VBf * , Vcf *. The zero sequence voltage signal is compared with a threshold S0: if the threshold is not exceeded, no fault 10 is considered to be present on line 4 and the location method is not engaged; otherwise, a fault 10 is considered to be detected D. In addition to any suitable means can give an indication as to the detection D, once the fault has been detected, the derivative with respect to time of the signals representative of the currents IA, IB, Ic of phase is determined, for example by software or hardware derivation of signals provided by current sensors 14A, 14B, 14C. Preferably, the signals representative of the derivatives of the currents are filtered and sampled, the order of these steps possibly being different from that indicated (for example, the filtering may take place before the derivative). The three signals are summed to obtain a signal representative of the derivative of the sampled sampled homopolar current of Iof * / dt. Alternatively, the homopolar current Io could be measured directly via a dedicated torus 14, then sampled, filtered and derived: such a direct measurement of the homopolar current is more accurate and avoids cumulative sensor errors when summation of the measurements of each line current. To obtain a reliable measurement, it is preferred to measure the signal for two periods.

Preferably, the same duration is used for the acquisition of the homopolar voltage, but it is not essential. Due to the shifting introduced by the shunt of the homopolar current Io, the signal dIo / dt is in phase opposition (180 °) with the voltage Vo in the case of a fault located upstream of the detection, since in this case the current is of capacitive origin 9. The method thus comprises a step of determining the angle p between the derivative of the homopolar current dIo / dt and the homopolar voltage Vo, or its complement a = 180 - p. This determination can be carried out by any known method: graphical or software or computationally measurement of the passages of the signals by zero, respectively at minimum or maximum, or elliptical representation, or even calculation of phaser, etc., or any combination. The complement a of the angle p between the derivative of the homopolar voltage dIo / dt and the homopolar voltage Vo is then compared to a threshold of angle Sa: under the threshold S, the angle a can be considered as zero to the inaccuracies due to the sensors, and the fault is located L as upstream of the measuring point; if the angle a is greater than the threshold, the location L will indicate the fault as downstream of the measuring point. The threshold Sa can be determined according to the precision I ', the sensors 12, 14 used, and the sampling frequency f - ech selected. In fact, in the case of sampling at x points per period, each sample covers an angular sector of 360 ° / x, corresponding to an accuracy of ± 360 ° / 2x: thus, the amplitude of the angular sector around 180 ° between the considered signals can be considered as (0 max Pr, 360 x F '= Sa, with F frequency of the network 1, that is to say that an upstream fault -' 2 x fech) corresponds to f3 e [180-Sa, 180 + Sa], or 1 al <max. It is possible for the current and voltage sensors 12 to be capable of introducing themselves phase shifts (which are in this case typically of ± 90 ° in the case of inductive or capacitive bridges), or that the I and V are obtained via other equipment also introducing a phase shift: the described principle can be easily adapted by adjusting the threshold Sa appropriately.

The method according to the invention can be implemented in a protection relay 8, in a fault indicator 6 with warning system 7, by implementation in a suitable directional earth fault detection device 100. A device 100 according to a preferred embodiment of the invention is shown diagrammatically in FIG. 4. It comprises a first module 20 making it possible to obtain a signal representative of the homopolar voltage Vo from information supplied by voltage sensors 12A. 12B, 12C adapted, for example those of the VPIS. The first module 20 preferably comprises, in undifferentiated order, means 24 for filtering signals VA, VB, Vc as an analog filter and sampling means 26, advantageously operating at less than 1 kHz; the filtered sampled signals VAf *, VBf *, Vcf * are processed by summation means 28 making it possible to determine the homopolar voltage Vof *. Alternatively, the summing means 28 can be adapted to directly process the signals received via the sensors 12A, 12B, 12C, filtering and sampling succeeding it.

Alternatively, if the network is provided with a sensor 12 adapted, the signal representative of the homopolar voltage Vo can be received directly by the first module 20 of the device 100 according to the invention, including the modules 24, 26 for filtering and sampling are then adapted to process the signal.

The first module 20 is thus able to supply a value Vo of the homopolar voltage, and the device according to the invention comprises means 30 for detecting a fault, in particular a module for comparing the homopolar voltage Vo obtained by the first module 20 with a threshold So. If the threshold is exceeded, then a fault D is detected: the device 100 may be provided with means for indicating the fault, for example a light or audible indicator, contact relays or a communication link.

In addition, the fault detection means 30 are adapted to activate a second module 40 of the device 100 according to the invention making it possible to obtain a signal representative of the derivative with respect to the time of the homopolar current Io. The second module 40 comprises a shunt block 42 for determining the derivative with respect to time of the signals representative of the phase currents IA, IB, IC provided by suitable sensors 14A, 14B, 14C, for example detection cores. In the illustrated embodiment, the derivatives dIA / dt, dIB / dt, dIc / dt are filtered by suitable means 44 as an analog filter, and then the filtered signals of IAddt, dIBf / dt, dIcf / dt are conditioned by sampling via sampling means 46, advantageously operating at less than 1 kHz. The filtered sampled signals dIAf * / dt, dIBf * / dt, dIcf * / dt are processed by summation means 48 of the second module 40 to determine the derivative of the homopolar current of Iof * / dt. It should be noted that the different processing means of the second module 40 can be organized differently: in particular, the summing means 48 can be adapted to directly process the signals received from the sensors 14A, 14C, 14B, the shunt block 42 being able to then succeed him; similarly, filtering and sampling can be arranged differently.

Alternatively, if the network is provided with a sensor 14 adapted, the signal representative of the homopolar current Io can be received directly by the second module 40 of the device 100 according to the invention, the means 42, 44, 46 of derivation, of filtering and sampling are then adapted to process the signal representative of homopolar current Io directly.

The results of the processing of the first and second modules 20, 40 are communicated to a relative positioning module 50 of the defect 10. The locating module 50 comprises first means 52 for calculating the angle p between the two values provided, namely the homopolar voltage Vo and the derivative of the homopolar current dIo / dt; advantageously, the first means are adapted to calculate the complement a, difference between n / 2 and (3, of said angle, the module 50 then comprises comparison means 54 for determining the relative location, for example means making it possible to compare with a The value of the complementary angle a has the value Sa, and the device 100 of FIG. 4 can be advantageously associated with a protection relay 8 for electrical networks, or with a fault-passing indicator for lines 4 average. underground or aerial voltages connected in network 1, the output of the locating module 50 triggering the breaking of a circuit breaker 5, the lighting of a light or any other means of security and / or warning. according to the invention thus allow: - a directional earth fault detection on a medium voltage network 1, including compensated neutral, on the basis of the homopolar voltage information and homopolar current: a directional earth fault detection without the need for high frequency sampling due to the use of the characteristics of the network 1 in steady state after a fault, and not in a transient regime; a directional earth fault detection that overcomes the problems of adjustment threshold, of capacitive current level according to the departures 4, and avoids "missing" defects in the event of precise compensation of the network 1, because of its initiation on a homopolar voltage threshold, and not the zero sequence current; a directional earth fault detection which allows an amplification of the data conveyed by the current by exploiting the derivative of the homopolar current and not the actual current; a directional earth fault detection which allows the simplification of the calculations and signal processing, because of the use of the homopolar current derivative which allows to align this magnitude directly in the axis of the homopolar voltage in the case of a fault upstream of the device, avoiding the recourse to the calculation of the phasors of currents and voltages or the polarization angle to define an upstream / downstream trip plan.

Although the invention has been described with reference to a three-phase distribution network in which the neutral is grounded by compensated impedance, it is not limited thereto: other types of multiphase networks may be concerned by the invention; in particular, any neutral scheme is appropriate. Moreover, although described with determination and treatment of the instantaneous homopolar voltage Vo for the detection of the fault, the method according to the invention can use the variation of said voltage Vo with respect to its value determined over a prior period: this variant s is particularly interesting in the case of networks having a slight imbalance between phases, whose homopolar voltage Vo is non-zero in non-default situation.

In fact, the various circuits, modules and functions presented in the context of the preferred embodiment of the invention can be made of analog, digital or programmable components operating with microcontrollers or microprocessors, and the representative signals described can having forms of electrical or electronic signals, data or information values in memories or registers, optical signals that can be displayed in particular on LEDs or screens, or mechanical signals acting with actuators. Likewise, the current sensors may be different from the described toroids, such as Hall effect or Faraday effect sensors, or magnetoresistors; the voltage can be detected by potential transformers, capacitive divider, resistive divider, again by Pockels effect sensors.

Claims (15)

REVENDICATIONS1. Directional detection device (100) for a ground fault (10) in a multi-phase medium voltage network (1) comprising: a first module (20) capable of supplying a signal representative of the zero sequence voltage (Vo) of the all phases; a second module (40) capable of providing a signal representative of the time derivative of the zero sequence current (Io); means for activating said second module (40) as a function of a signal (D) for detecting the occurrence of a ground fault in the network (1); a locator module (50) including means (52) for determining the angle (f3) between the zero sequence voltage (Vo) provided by the first module (20) and the derivative of the zero sequence current (dIo / dt) provided by the second module (40), and means for comparing said angle (f3) with a fixed parameter to determine whether the defect is upstream or downstream of the device (100). Directional detection device (100) according to claim 1 wherein the first module (20) comprises means for receiving the signal representative of the voltage (VA, VB, Vc) of each of the phases of the network (1) and the summing means (28) for providing said homopolar voltage (Vo). 3. directional detection device (100) according to one of claims 1 or 2 wherein the second module (40) comprises means for receiving a signal representative of a current (I) of the network (1) and a block of derivation (42) adapted to provide the derivative with respect to the time (dI / dt) of a signal representative of a current (I). Directional detection device (100) according to claim 3, in which the means for receiving a signal representative of a current are adapted to receive the signal representative of the current (IA, IB, Tc) of each of the phases of the network (1). and wherein the second module (40) includes summing means (48) for providing a signal representative of the zero sequence current (Io). 5. directional detection device (100) according to one of claims 1 to 4 wherein the first and / or the second module (20, 40) comprise sampling means (26, 46) and / or an analog filter (24, 44). 6. Directional detection device (100) according to one of claims 1 to 5 wherein the locating module (50) comprises means (52) for determining the complementary (a) 180 ° angle (f3). between the homopolar voltage (Vo) and the derivative of the homopolar current (dIo / dt), and the comparison means (54) are adapted to compare said complementary (a) with a threshold (Se). 7. directional detection device (100) according to one of claims 1 to 6 further comprising means (30) for detecting the occurrence of a ground fault in the network connected to the first module (20) and the activation means of the second module (40), said detection means (30) comprising a comparator of the signal representative of the homopolar voltage (Vo) at a detection threshold (S0). 8. Ground fault indication indicator (6) comprising current and voltage sensors (12, 12A, 12B, 12C, 14, 14A, 14B, 14C) arranged on a line (4, 4A, 4B, 4C) ) an electrical network (1) to be monitored and comprising a directional fault detection device (100) according to one of claims 1 to 7 connected to said sensors (12, 12A, 12B, 12C, 14, 14A, 14B, 14C) for receiving signals representative of current and voltage. 9. Earth protection relay (9) comprising at least one fault indicator according to claim 8 and means for actuating a cutoff device (5) according to the results of the location module (50) of the device directional detection device (100) of the indicator. 10. A method for directionally detecting (D, L) a ground fault (10) in a multiphase network (1) comprising triggering, after obtaining a signal (D) indicating the presence of said defect to the earth (10), the directional determination (L) of the defect (10), said directional detection comprising the successive steps of: - obtaining signals representative of the homopolar voltage (Vo) and the derivative with respect to the time of the current homopolar (dIo / dt); - determining the angle (f3) between the signals representative of the zero sequence voltage (Vo) and the time derivative of the zero sequence current (dIo / dt); comparing said angle (3) with a fixed parameter to indicate whether the detected fault (D) is downstream or upstream of the place where the representative signals were obtained. The directional detection method according to claim 10 wherein the signal (D) indicative of the presence of a ground fault (10) is obtained by comparing the zero sequence voltage signal (Vo) with a detection threshold of default (S0). 12. Directional detection method according to one of claims 10 to 11 wherein obtaining a signal representative of the homopolar voltage (Vo), respectively of a signal representative of the derivative with respect to the time of the zero sequence current ( dIo / dt), consists in calculating said voltage, respectively said current, as a function of the signals representative of each of the voltages, respectively each of the phase currents (IA, Is, Tc). 13. Directional detection method according to one of claims 10 to 12 wherein the comparison of the angle (3) between the signals representative of the homopolar voltage (Vo) and the derivative with respect to the time of the homopolar current (dIo / dt) includes the comparison of its complement (a) at 180 ° with a threshold (Se). 14. Directional detection method according to one of claims 10 to 13 wherein the obtaining of signals representative of the zero sequence voltage (Vo) and the time derivative of the zero sequence current (dIo / dt) comprises a sampling at frequency below 1 kHz. A method of protecting a current line (5) upon occurrence of a ground fault (10) comprising actuating a cut-off device (6) of said line (5) if a earth fault (10) has been detected by a method according to one of claims 10 to 14 downstream of said cut-off device (6).