EP2399140A1

EP2399140A1 - A sensor and process for detecting an electric pulse caused by a partial discharge

Info

Publication number: EP2399140A1
Application number: EP10708805A
Authority: EP
Inventors: Luigi Testa; Gian Carlo Montanari
Original assignee: TECHIMP Tech SRL
Current assignee: TECHIMP Tech SRL
Priority date: 2009-02-17
Filing date: 2010-02-15
Publication date: 2011-12-28
Also published as: WO2010095086A1; IT1392776B1; ITBO20090082A1

Abstract

A sensor (1) for detecting an electric pulse caused by a partial discharge and propagating along an axial direction in an electric apparatus (2) with cylindrical geometry comprises a conducting wire (10) wound on a support (9) to form a winding (11) with at least one turn, so that a component of the magnetic field generated outside the apparatus (2) by said pulse, concentric with the axial direction, concatenates with the winding (11), when the sensor (1) is operatively connected to the apparatus (2), for producing at terminals (12) of the winding (11) a voltage correlated with the pulse. The support (9) comprises a thin layer, having a thickness which is much less than its extension in a longitudinal direction, said at least one turn of the winding (11) being extended in the longitudinal direction, in such a way that, when the sensor (1) is operatively connected to an outer surface of the apparatus (2), said at least one turn forms a flux concatenation section positioned in a radial plane and extended in the axial direction.

Description

A sensor and process for detecting an electric pulse caused by a partial discharge

Technical Field

This invention relates to a sensor and process for detecting an electric pulse caused by a partial discharge.

In general terms, the invention applies to the sector of diagnostic systems for medium- and high-voltage apparatuses, based on the measurement of partial discharges.

More specifically, this invention relates to a sensor and process for detecting partial discharges in an electrical apparatus with cylindrical geometry, for example a coaxial cable. A partial discharge is an electric discharge limited to a portion of the insulation of the apparatus, and does not therefore cause failure of the apparatus (since it does not short-circuit its electrodes) but constitutes a cause and an effect of gradual degradation of the apparatus.

A partial discharge is a transfer of charge to a limited region, that is, to the discharge site (generally located at a point where there is a defect in the insulation). The charge transfer generates a pulse of current that propagates along the apparatus.

In this specific case, where the apparatus has a cylindrical geometry and thus extends axially, the pulse of current related to the partial discharge propagates axially in the apparatus.

Background Art

There are numerous technical solutions regarding partial discharge sensors.

This invention relates in particular to sensors consisting of magnetic probes and, more specifically, broadband magnetic probes.

In this context, the sensors commonly used can sense an axial component of the magnetic field generated by the discharge pulse. Patent document WO2006/122415, for example, describes a sensor of this type. Patent document US5530364 also illustrates a sensor of this type (in Figures 6 and 7). These sensors comprise a winding turn designed to encircle the apparatus in such a way that the axial component of the magnetic field flux produced by the discharge pulse concatenates with the winding turn. The axial component of the magnetic field is generated by the discharge pulse when the pulse propagates in the apparatus at a speed which has a non-axial component, that is to say, which has a tangential component that is not zero. This is the case of coaxial cables whose shield consists of wires wound in a spiral. These sensors therefore have the disadvantage of not allowing the discharge pulses to be detected in all situations but only in special cases; in effect, generally speaking, the axial component of the magnetic field generated by the discharge pulse is zero. For example, such is the situation in a coaxial cable where the shield consists of a uniform conductor. Patent document US5530364 also illustrates (Figures 9 and 10) a different type of sensor, which senses a tangential component of the magnetic field generated by the discharge pulse which propagates along the axial direction, that is to say, a component of the magnetic field concentric with the axial direction defined by the apparatus with cylindrical geometry. This component of the magnetic field is also known as residual magnetic field since, in a coaxial cable, this component should be zero under conditions of perfectly uniform distribution of the currents corresponding to the discharge pulse travelling the high voltage conductor (core) of the cable and the low voltage one (shield) of the cable. Experimental evidence shows, however, that there is usually a residual tangential component in the magnetic field generated by the discharge pulse propagating axially along the apparatus.

The advantage of sensors of this type is that they do not have the limitations mentioned above. Furthermore, these sensors make it possible to detect the discharge pulse even at considerable distances from the source of the discharge, exploiting the fact that the cable acts as a waveguide for the signal.

In prior art solutions for sensors of this type (for example, the one described in US5530364) there is a cylindrical ferromagnetic core round which a conductor is wound to form a winding. In use, this cylinder is placed transversally to the axis of the apparatus in such a way that the tangential magnetic field lines, that is, lines concentric about the axial direction, travel along the core and concatenate with the winding to produce, at the ends of the winding, a voltage that is correlated with the discharge pulse. These prior art partial discharge sensors sensitive to the tangential component of the magnetic field generated by the discharge pulse have disadvantages which are such that nowadays these sensors are hardly ever used in practice.

First of all, these sensors have very low sensitivity.

The serious risk of this is that apparatus may be mistakenly considered free of partial discharges simply because the intensity of the discharges cannot be detected by the sensor.

Moreover, these sensors are greatly affected by disturbances and noise extraneous to the apparatus.

This further increases the risk of mistakenly interpreting the signals detected for diagnostic purposes. Furthermore, these sensors are in no way adaptable to the geometry of the apparatus they are required to measure.

Patent document WO2008/008105 discloses a sensor which can be coupled to an electrical cable in order to detect the partial discharges along the cable.

The sensor comprises a ferromagnetic core having the shape of closed ring. The ring forms an aperture in which the cable is placed. Around the ferromagnetic core there is a winding which acts as the secondary winding of a current transformer and which carries the signal to a measuring instrument. Further, concentric conductor wires of the cable shield are gathered together to form a conductor and the conductor is routed through the closed ring formed by the ferromagnetic core and connected to earth.

The conductor for connecting the cable shield to earth and passing through the ferromagnetic ring of the sensor constitutes a bypass essential to the operation of the sensor.

One disadvantage of this sensor is that it is troublesome to install (because making the bypass takes time). Another shortcoming is that it cannot be coupled to the cable at any section but only in the proximity of the breaks in the cable shield.

A sensor which can be coupled to an electrical cable in order to detect the partial discharges along the cable is also disclosed in patent document WO97/10515.

This sensor forms a so-called Rogowski coil. In that coil, the winding is configured in such a way that when the sensor is coupled to the sensor, it fully encircles the cable, extending over an arc of approximately 360 degrees.

Therefore, a sensor of this kind does not work if it is placed above (that is, outside) the cable shield. In light of this, WO97/10515 teaches applying the sensor directly to the insulating surface of the cable.

This constitutes a shortcoming because the sensor can be used only on unshielded cables or, alternatively, must be installed on cables whose shield has been prepared beforehand (that is, when the cable is not in use) by a complicated and awkward operation of removing a portion of the shield to allow installation of the Rogowski coil.

Aim of the Invention

This invention has for an aim to provide a sensor and a process that overcome the above mentioned disadvantages of the prior art.

More specifically, the aim of the invention is to provide a particularly efficient and reliable sensor and process for detecting an electric pulse caused by a partial discharge and propagating along an axial direction in an electric apparatus with cylindrical geometry, based on the detection of a component, concentric with said direction, of the magnetic field generated by said pulse.

Another aim of the invention is to provide a partial discharge pulse detection sensor and process which are particularly sturdy and immune to noise.

Another aim of the invention is to provide a partial discharge pulse detection sensor and process which are particularly versatile and simple to implement, with the possibility of connecting the sensor at any section of the cable, outside the cable (even if the cable is shielded). Another aim of the invention is to provide an instrument and a process for detecting partial discharge pulses and capable of locating the source of the partial discharges.

These aims are fully achieved by the sensor, process and apparatus according to the invention as characterized in the appended claims. In general terms, this invention is based on the detection of a residual magnetic field, consisting of a non-zero tangential magnetic field (generated by a current travelling the cable axially, that is, longitudinally) detectable outside a shielded cable.

In effect, the Applicant has found that the residual field is correlated with a non-uniformity in the distribution of the current corresponding to the partial discharge pulse which propagates in the cable shield.

Indeed, the current corresponding to the discharge pulse is concentrated in a lateral portion of the shield corresponding to a certain angular zone. Hence, this phenomenon (that is, the imbalance in the angular distribution of the discharge current between the core and the shield of the cable is at the basis of the above mentioned residual magnetic field.

More specifically, the sensor according to the invention is a sensor for detecting an electric pulse caused by a partial discharge and propagating along an axial direction in an electric apparatus with cylindrical geometry, comprising a conducting wire wound on a support to form a winding with at least one turn, so that a component of the magnetic field generated outside the apparatus by said pulse, concentric with the axial line, concatenates with the winding, when the sensor is operatively connected to the apparatus for producing at the terminals of the winding a voltage correlated with the pulse.

According to the invention, the support comprises a thin layer having a thickness which is less than its extension along a longitudinal direction, said at least one turn of the winding being extended in said longitudinal direction; the sensor being able to connect to the apparatus in such a way that, when the sensor is operatively connected to an outer surface of the apparatus, said at least one turn forms a flux concatenation section positioned in a radial plane and extended in the axial direction. Also according to the invention, the support has a transversal extension of predetermined length and is shaped in such a way as to encircle at least one portion of the lateral surface of the apparatus. Further, the winding is configured in such a way that it forms a surface corresponding to a predetermined angle of less than 360 degrees, without completely surrounding the apparatus when the sensor is operatively connected to the outer surface of the apparatus.

It should be noted that according to the invention the support is shaped in such a way as to form a high-reluctance circuit for a flux of said magnetic field generated outside the apparatus concentric with the axial direction.

It should also be noted that according to the invention an assembly composed of the support and the winding wrapped around it is shaped in such a way as to form active magnetic components distributed on a surface corresponding to a predetermined angle of less than 360 degrees, without completely surrounding the apparatus, when the sensor is operatively connected to the outer surface of the apparatus. The expression "active magnetic components" means turns and/or materials designed to concentrate the magnetic field, that is to say, is used to denote any component designed to produce at least one of the following effects: i) altering a magnetic field distribution; ii) inducing an electric voltage when it interacts with a magnetic field. Therefore, the problem tackled by this invention is that of providing a sensor that can be connected to the outside of a shielded cable in order to detect partial discharge signals by detecting said residual field without having to break or access the cable shield.

More specifically, this invention provides two alternative solutions to the problem.

In a first solution, the support is made of ferromagnetic material. In that case, the support forms a broken ring, or ring portion, that is to say, it is shaped in such a way as to encircle only a portion of the lateral surface of the apparatus, without encircling it completely.

In a second solution, the support is made of non-ferromagnetic material. In that case, the support forms a ring portion or complete ring (that is to say, if it is made of non-ferromagnetic material, the support might if necessary be shaped in such a way as to completely encircle the cable, when the sensor is connected to it).

In both the solutions, the winding is configured in such a way that it forms a surface corresponding to a predetermined angle of less than 360 degrees, without completely surrounding the apparatus, when the sensor is operatively connected to the outer surface of the apparatus.

Described below are other technical features, in all the possible combinations, regardless of the above mentioned alternative solutions, unless expressly stated that a particular feature depends on whether the support is made of ferromagnetic or non-ferromagnetic material.

Brief Description of the Drawings

These and other features of the invention will become more apparent from the following description of a preferred, non-limiting embodiment of it, with reference to the accompanying drawings, in which: - Figure 1 is a perspective view of a sensor according to this invention;

- Figure IA is a perspective view of the sensor of Figure 1, connected to a cable;

- Figure 2 illustrates the sensor of Figure 1, in a cross section according to a plane perpendicular to the axis of the cable; - Figure 3 illustrates the sensor of Figure 1, in a cross section according to a plane containing the axis of the cable;

- Figure 4 is a side view of the sensor of Figure 1;

- Figure 5 shows the sensor of Figure 4 according to another embodiment;

- Figure 6 schematically illustrates two sensors connected to a cable, according to another aspect of this invention.

Detailed Description of the Preferred Embodiments of the Invention The numeral 1 in the accompanying drawings denotes a sensor according to the invention.

The sensor 1 is a broadband sensor; more specifically, the sensor 1 is a broadband magnetic probe. The sensor 1 is a sensor for detecting an electric pulse caused by a partial discharge and propagating along an axial direction Z in an electric apparatus 2 with cylindrical geometry.

More specifically, the sensor 1 is applied to coaxial cables. It should be noted, however, that the sensor 1 can also be advantageously applied to other electrical apparatuses with cylindrical geometry, such as single-core and multi- core cables and related accessories, GIS, TA and TV high-voltage ducts.

In the drawings the numeral 2 denotes a coaxial cable comprising a high- voltage conductor 3 (also known as the core of the cable) surrounded by a semi- conductive layer 4 on the high-voltage side; the numeral 5 denotes an insulating layer surrounding the layer 4, in turn surrounded by a semi-conductive layer 6 on the low-voltage side, which is in turn surrounded by a conductive shield 7; around the shield 7 there is a sheath 8 for mechanically protecting the cable 2.

As regards the propagation of the pulses of current caused by partial discharges in coaxial cables, attention is drawn preliminarily to the following. The transfer of charge associated with the partial discharge produces two pulses of current of opposite polarity which propagate in the cable, one in the core 3 and the other in the shield 7. Since the two pulses have the same intensity, in principle, the contributions to the tangential magnetic field generated by the two pulses outside the cable should cancel each other out. hi practice, however, it is known that a non-zero tangential magnetic field, also known as residual magnetic field, can usually be measured on the outside of the cable.

Research conducted by the Applicant has demonstrated that the current corresponding to the discharge pulse which propagates in the cable core can be considered distributed uniformly in the core section. The current corresponding to the discharge pulse which propagates in the shield, on the other hand, cannot be considered distributed uniformly since it is concentrated in a lateral portion of the shield corresponding to a certain angular zone. The zone of highest discharge current density is particularly narrow in the proximity of the discharge site (that is, in the zone of the cable where the partial discharge occurred), while it tends to widen with increasing distance from the discharge site.

This phenomenon (that is, the imbalance in the angular distribution of the discharge current between the core and the shield of the cable ) is at the basis of the above mentioned residual magnetic field.

That also means that the value of the tangential magnetic field measurable on the outer surface of the coaxial cable varies with the variation of the angular position.

The sensor 1 comprises a support 9 round which a conducting wire 10 is wound to form a winding 11.

The winding 11 comprises at least one turn formed by the conducting wire 10 wound round the support 9. Preferably the winding 11 comprises a plurality of turns. According to the invention, however, the winding 11 might also have a single turn.

The winding 11 has a pair of terminals 12.

In use, the sensor 1 is connected to the apparatus 2 in a way that a component of the magnetic field (generated outside the apparatus 2 by the discharge pulse), concentric with the axial direction Z, concatenates with the winding 11 to produce at the terminals 12 of the winding 11 a voltage correlated with the discharge pulse.

The expression "discharge pulse" is used herein to mean the pulse of electric current which is caused by the partial discharge and which propagates axially along the apparatus 2.

According to the invention, the support 9 comprises a thin layer having a thickness which is less (or much less) than its extension along a longitudinal direction.

The conducting wire 10 is wound round the support 9 in such a way that the turns (or the at least one turn) of the winding 11 extend along said longitudinal direction.

It should be noted that each turn of the winding 11 (the winding 11 might comprise one turn only) forms a magnetic flux concatenation section.

The numeral 13 in the drawings denotes magnetic field lines (or magnetic flux lines) of the tangential component, that is to say lines which are concentric about an axial direction Z.

In use, the sensor 1 is operatively in contact with the outer surface of the apparatus 2; in the example illustrated, the sensor 1 is operatively in contact with the outer surface of the sheath 8. That way, when the sensor 1 is operatively connected to an outside surface of the apparatus 2, each turn of the winding 11 forms a flux concatenation section positioned in a radial plane end extended in the axial direction Z. Preferably, each turn of the winding 11 is shaped in such a way as to form, a substantially rectangular flux concatenation section, with the long side approximately extending for a greater length (for example at least ten times greater) than the short side; that way, when the sensor 1 is operatively connected to the outer surface of the apparatus 2, each turn is positioned with the long side in the axial direction Z and the short side in the radial direction.

Therefore, the sensor 1 is connected to the apparatus (more specifically, to the outer surface of the apparatus 2) in such a way that the at least one turn is positioned with the long side in the axial direction and the short side in the radial direction.

In novel manner, therefore, the winding 11 is extended in the longitudinal direction and flattened in the direction corresponding to the thickness of the support 9.

Preferably, the extension of each turn in the longitudinal direction is at least ten times greater than the extension of the turn in the direction of the thickness of the winding 11.

More preferably, the extension of each turn in the longitudinal direction is at least eighty times greater than the extension of the turn in the direction of the thickness of the winding 11. For example, the support 9 has a thickness of approximately 5 mm and a longitudinal extension of approximately 500 mm; in that case, each turn of the winding 11 has an extension in the longitudinal direction of approximately 500 mm and an extension in the direction of the thickness of the winding 11 (that is, of the thickness of the layer 9) of approximately 5 mm. As regards, more specifically, the value of the thickness of the support 9 and of the winding 11, this falls preferably within the range of 1 - 10 mm and, more preferably, is approximately 5 mm.

It should be noted that the preferred values of the winding 11 thickness were determined in a study of the physical phenomena involved in the detection of magnetic coupling discharge signals, following extensive in-depth research conducted by the Applicant.

In practice, it was found that for very low winding 11 thickness values (for example, less than 1 mm), the quantity of flux that concatenates with the winding is too small to guarantee adequate sensitivity (the value of sensor inductance becomes too low).

Very high winding 11 thickness values (for example, greater than 10 mm), on the other hand, lead to a worsening of the frequency response of the sensor 1, because the inductance of the sensor increases too much, and the sensor tends to detect more noise from outside the apparatus while the quantity of useful flux concatenated with the winding increases less than proportionally, which means that the overall sensitivity of the sensor diminishes. As regards, more specifically, the value of the longitudinal extension of the support 9 and of the winding 11, this is preferably greater than approximately 5 mm and, more preferably, greater than approximately 100 mm.

Preferably, the support 9 has a transversal extension of predetermined length and is shaped in such a way that it encircles at least one portion of the lateral surface of the apparatus 2, said portion corresponding to a predetermined angle.

The conducting wire 10 is wound on the support 9 with one or more turns positioned in such a way that the winding 11 covers a surface corresponding to said predetermined angle. Thus, in use (that is, when the sensor 1 is operatively connected to the outer surface of the apparatus 2), the winding 11 encircles a portion of the lateral surface of the apparatus 2 corresponding to a predetermined angle, without completely surrounding the apparatus 2.

It should be noted that the greater the transversal extension of the winding 11 (and hence the higher the number of turns of the winding 11) the greater the impedance of the sensor 1.

Preferably, the winding 11 is shaped in such a way that it encircles a portion of the lateral surface of the apparatus corresponding to an angle of between 20 and 310 degrees. More preferably, the winding 11 is shaped in such a way that it encircles a portion of the lateral surface of the apparatus corresponding to an angle of between 40 and 240 degrees.

Still more preferably, the winding 11 is shaped in such a way that it encircles a portion of the lateral surface of the apparatus corresponding to an angle of between 60 and 120 degrees.

Even more preferably, the winding 11 is shaped in such a way that it encircles a portion of the lateral surface of the apparatus corresponding to an angle of approximately 90 degrees.

In effect, the transversal extension of the support must be sufficiently large to guarantee coupling to a significant portion of the magnetic field to be detected

(since the component of the magnetic field the sensor detects is, at least in the proximity of the discharge source, distributed non-uniformly along the lateral surface of the cable with reference to the angular position).

At the same time, the transversal extension of the support 9 must be small enough to allow the sensor 1 to be easily mounted on the cable.

The support 9 may be made of a ferromagnetic material (for example, it may be a sheet of ferrite for high frequency radio applications) or of any non- ferromagnetic material such as, PVC o, more preferably, vulcanized rubber.

Use of ferromagnetic material for the support 9 means increasing the inductance of the sensor 1. Reducing the inductance by using a non-ferromagnetic material, however, may be compensated by using a winding 11 with a higher number of turns.

In light of this, attention is drawn to the following as regards the number of winding 11 turns.

If the support 9 is made of a ferromagnetic material, the winding 11 preferably has two or three turns. If the support 9 is made of a non-ferromagnetic material, the winding 11 preferably has between ten and twenty turns.

In any case, if the material used for the support 9 is ferromagnetic, the support 9 is shaped in such a way as to encircle a limited portion of the lateral surface of the apparatus 2, that is to say, a portion corresponding to a predetermined angle of less than 360 degrees, without encircling the apparatus 2 completely (that is, a transversal section of the apparatus).

If the material used for the support 9 is non-ferromagnetic, on the other hand, the support 9 is shaped in such a way as to encircle at least one portion of the apparatus 2 (when the sensor is connected to it), with the possibility of also encircling the apparatus 2 completely; that is, the support 9 is shaped in such a way as to encircle a portion of the lateral surface of the apparatus 2 corresponding to a predetermined angle less than or even equal to 360 degrees.

The sensor 1 also comprises anchoring means (not illustrated in the drawings) for anchoring the sensor 1 to the apparatus 2, so that it is operatively in contact with a desired portion of the outer surface of the apparatus 2.

For example, the anchoring means comprise a strip of Velcro applied to a surface of the support 9. Alternatively, the anchoring means comprise self-locking arms or film strap.

That way, according to the example illustrated, the sensor is applied to the sheath 8 of the cable 2, with the possibility of fastening it to the sheath itself at a desired angular position.

That way, the sensor, or more specifically, the winding 11 is operatively in contact with the apparatus 2.

This is important to optimize the sensitivity of the sensor. In effect, the Applicant has found that the tangential magnetic field that the sensor is designed to detect is mostly concentrated near the outer surface of the apparatus 2. It should be noted that the support 9 is made of a flexible material to allow winding 11 curvature around the longitudinal direction along which the turns extend.

For example, the support 9 comprises a sheet of ferrite (in which case the support is made of a ferromagnetic material). Alternatively, the support 9 comprises a sheet of vulcanized rubber.

The fact that the support 9 (that is, the sensor 1) is made of a flexible material makes the sensor particularly versatile. In effect, the sensor can be coupled to cables (or other equipment with cylindrical geometry) of any diameter. Thanks to the flexibility of the sensor and to the presence of the means for anchoring the sensor itself to the apparatus 2, it is very easy to connect the sensor to the apparatus 2 in such a way that it is operatively in contact with the outer surface of the apparatus by adapting the curvature of the sensor to the curvature radius of the apparatus.

In addition to or instead of making the sensor from flexible material, the support 9 according to the invention might also be shaped in such a way that the winding 11 is curved around the direction along which the turns are extended.

This makes it particularly suitable for encircling an apparatus 2 of predetermined geometry.

For example, the sensor is in the shape of a sheet approximately 0.5 mm thick, approximately 500 mm long and between 100 mm and 400 mm wide extending round a circular arc.

Advantageously, the sensor 1 is shaped in such a way that, in use, each turn is positioned in a radial plane, said planes being included in a sheaf of planes that intersect along an axis of the apparatus 2. According to another aspect of this invention, the sensor 1 comprises a conductive layer 14 applied to one face of the support 9 to cover a corresponding portion of the winding 11, in such a way that, in use, (when the sensor 1 is operatively connected to the outer surface of the apparatus 2), the winding 11 is interposed between the outer surface of the apparatus 2 and the conducting layer 14; the layer 14 therefore has a shielding action on the winding 11 against the electromagnetic waves from the outside of the face of the sensor 1 (that is, of the support 9) opposite the one the layer 14 is applied to (therefore, in use, the shielding action is against noise from outside the apparatus T).

This advantageously reduces the sensitivity of the sensor to disturbances and noise from outside the apparatus.

The conductive shield 13 comprises a thin sheet of aluminium (which is in turn flexible).

Preferably, the conductive shield 13 extends over a larger surface than the surface occupied by the support 9 (and hence by the winding 11), so that one edge of the conductive screen projects laterally relative to the support 9 and to the winding 11 ; this provides a particularly effective shielding effect. Moreover, the shield 13 has no ferromagnetic materials.

It should be noted that the conducting wire 10 that forms the winding 11 comprises an insulating layer (for example, an insulating enamel of known type and not illustrated) to insulate each turn from the other turns and from the insulating shield 14. The presence of the insulating shield layer 14 cooperates synergically with the flattened shape of the sensor 1 to optimize the latter' s sensitivity to the partial discharge signals propagating in the apparatus 2, when the sensor is operatively connected to the apparatus 2.

According to another aspect of this invention, the sensor 1 constitutes a system comprising a plurality of sensors of the type described above, each shaped in such a way as to encircle, in use, its own angular portion of the outer surface of the apparatus 2, preferably at a single section.

That way, the plurality of sensors substantially covers the entire outer surface of the apparatus 2 at a predetermined section (the sensor connecting section or detection section).

In the example illustrated in Figure 5, the sensor 1 comprises a system of two sensors (each shaped in such a way as to encircle a portion of the apparatus corresponding to an angle of approximately 120-180 degrees), that is a first sub- sensor 101 and a second sub-sensor 102; that is to say, the sensor 1 of Figure 5 is a composite sensor comprising a sensor 101 and a further sensor 102.

Therefore, the composite sensor 1 comprises, in addition to the first sensor 101:

- a second support comprising a thin layer shaped in such a way as to encircle at least one portion of the lateral surface of the apparatus 2 corresponding to a predetermined angle like the first support;

- a second winding extended in the longitudinal direction and wound on the second support in such a way as to encircle said portion of the lateral surface of the apparatus corresponding to said predetermined angle, without completely surrounding it, when the sensor 102 is operatively connected to the outer surface of the apparatus;

- means for anchoring the first and second supports to the apparatus 2 so that they encircle separate angular portions of the outer surface of the apparatus;

- an output module 15 connected to the terminals of the first and second windings, to make available at output the voltages produced at the ends of the windings.

These anchoring means are made as described above. The output module 15 comprises a connector designed to receive the signals detected by all the sub-sensors (of which there may be two or more) and transfer them to the output.

It should be noted that each sub-sensor has a winding 11 which encircles only a limited angular portion of the apparatus 2 without extending all the way round the circumference of the section of the apparatus (that is to say, the detection section) the sensor is coupled to.

This composite sensor is particularly advantageous for connection to a coaxial cable. In effect, as explained above, in coaxial cables, the intensity of the transversal magnetic associated with a partial discharge (that is, the residual field) is not uniform in the whole of the portion of space surrounding the cable but is highest at a spatial region corresponding to a limited angular portion. Upon installation of the sensor, however, (that is, when the sensor 1 is connected to the cable 2) it is not usually possible to know which angular portion of the outer surface of the cable will have the maximum residual magnetic field (the position also changes with the variation of the discharge site; in fact, upon installation, it is not even certain whether or not there will be partial discharges).

In light of this, the composite sensor has the advantage of avoiding the risk that the sensor 1 will be connected to an angular portion of the apparatus where the residual magnetic field to be detected is very moderate. In effect, the composite sensor guarantees that there is always at least oine sensor connected to the angular portion of the apparatus 2 where the tangential (that is, residual) magnetic field is largest.

The composite sensor also has the further advantage of allowing the signals detected by the sub-sensors connected to different angular portions of the same section of the apparatus 2 to be compared in order to derive information as to the position of the discharge site within the apparatus.

This invention also provides a process for detecting an electric pulse caused by a partial discharge and propagating along an axial direction in an electric apparatus 2 with cylindrical geometry.

The process comprises the following steps:

- preparing a sensor 1 as described above; - applying the sensor 1 to an outer surface of the apparatus 2 so that it encircles at least one portion of it in a detecting section, in such a way that a component of the magnetic field generated outside the apparatus 2 by said pulse, concentric with the axial direction (that is, the tangential, or residual, magnetic field) concatenates with the winding 11 to produce at winding 11 terminals 12 a voltage correlated with the pulse;

- detecting said voltage correlated with the pulse.

It should be noted that the winding 11 encircles a portion of the apparatus 2 corresponding to a predetermined angle (for example, between 120 and 180 degrees), and does not completely surround the section of the apparatus 2. According to the invention, during the application step, the entire section of the apparatus 2 is encircled by applying a plurality of windings 11 forming part of corresponding sensors 1 of a composite sensor (as described above).

According to another aspect of this invention, the process comprises, in addition to the steps described above, the following further steps: - preparing a second sensor IB (made as described above), the first sensor

IA and the second sensor IB having predetermined winding directions for the respective turns;

- applying the second sensor to an outer surface of the apparatus 2 in a second detecting section axially spaced from the first detecting section by a predetermined quantity, in such a way that a component of the magnetic field generated outside the apparatus by said pulse, concentric with the axial direction, concatenates with the winding of the second sensor IB to produce at the terminals of that winding a voltage correlated with the pulse;

- detecting a voltage correlated with a component of the magnetic field generated outside the apparatus 2 by the pulse, concentric with the axial direction, said component of the magnetic field being magnetically coupled to the second sensor IB;

- comparing the sign of the voltages detected by the two sensors IA, IB in the same predetermined time interval in order to detect the direction from which the pulse is propagating in the apparatus 2 so as to locate it.

In light of this, it should be noted that the sign of the voltage generated at the terminals 12 of the sensor 1 depends on the direction in which the conducting wire 10 is wound round the support 9 to form the winding 11 and on the direction from which the pulse of current associated with the partial discharge propagates axially in the apparatus 2.

The speed at which the pulse of current associated with the partial discharge propagates in the apparatus 2 is slightly less than the speed of light and can be estimated as a function of the physical properties of the apparatus.

Thus, if two signals are detected at the two sensors IA and IB substantially simultaneously (that is, the two detections are temporally spaced by an interval substantially corresponding to the time necessary for a pulse to travel the distance separating the two sensors at said estimated speed), it may be concluded that both these signals correspond to a single discharge pulse.

Therefore, by comparing the signals detected by the two sensors, it is possible to determine whether the source of the discharge pulse (that is, the source of the partial discharge that caused the pulse) is located between the two sensors or outside one or the other of the sensors.

It should be noted that the sensor 1 according to the invention is very simple and inexpensive and the application of the sensor to the apparatus is easy and involves no disruption in service.

The process described above is very effective in that it comprises applying to the apparatus 2 a very large number of sensors, positioned close to each other in order to provide very precise information for locating the discharge source.

This invention also provides a diagnostic instrument for detecting and locating partial discharges in an electric apparatus with cylindrical geometry, generating corresponding electric pulses propagating along an axial direction in the electric apparatus.

The instrument comprises at least a first sensor IA and preferably a second sensor IB (as described above), which can connect to the apparatus in a first and a second detecting section, respectively, being axially spaced from each other, for detecting corresponding voltage signals relating to the pulses caused by the partial discharges.

The instrument preferably also comprises a processing unit (not illustrated in the drawings) connected to the sensors 1 to receive the voltage signals, select pairs of voltage signals caused by the same partial discharge and compare the sign of said signals to obtain information about the location of the partial discharge relative to the detecting sections.

The processing unit comprises, for example, a computer, on which dedicated hardware is installed, according to substantially known technology. The processing unit defines a predetermined input impedance preferably configured to guarantee a very wide detection band (for example, a bandwidth of between 10 MHz and 100 MHz).

It should be noted that the main purpose of the sensor 1 and of the apparatus comprising the sensor is to maximize the sensitivity of the sensor.

For this purpose, it is important for the inductance of the sensor 1 to be optimized relative to the input impedance of the processing unit.

For example, the input impedance of the processing unit is preferably 50 Ohm and the impedance of the sensor 1 is preferably 50 Ohm. The inductance of the sensor increases with the increase in the number of winding 11 turns and of the cross section of each turn and also increases if a ferromagnetic material is used to make the support 9.

The value of sensor inductance must not be too small, otherwise the amount of flux that concatenates with the winding 11 is too small to allow an adequate level of sensitivity.

It is not, however, appropriate to increase sensor inductance too much because that would have the disadvantage of filtering out the components of the discharge pulse at higher frequencies, which would mean loss of sensitivity and fidelity (in terms of response linearity) for the sensor. Preferably, the sensor 1 is configured to have an inductance in the range of

2 to 5 micro H.

This invention therefore has the following advantages.

The sensor is applicable to any electrical apparatus with cylindrical geometry (to detect an electric pulse caused by a partial discharge and propagating along an axial direction in an electric apparatus) since it can sense, of the magnetic field generated by the pulse, the component that is concentric with the axial direction (that is, the tangential, or residual, component of the magnetic field).

In addition, the sensor is highly efficient and reliable because it maximizes the useful magnetic field flux (the one generated by the discharge pulse that propagates in the cable) in such a way as to optimize the sensitivity of the sensor.

Moreover, the sensor provided is particularly sturdy and immune to noise since it minimizes sensitivity to noise from outside the cable and to the components of the magnetic flux that are not useful. The instrument and process according to the invention, in addition to the advantages listed above, allow the partial discharge source to be located easily and effectively.

Claims

1. A sensor (1) for detecting an electric pulse caused by a partial discharge and propagating along an axial direction in an electric apparatus (2) with cylindrical geometry, comprising a conducting wire (10) wound on a support (9) to form a winding (11) with at least one turn, so that a component of the magnetic field generated outside the apparatus (2) by said pulse, concentric with the axial direction, concatenates with the winding (11), when the sensor (1) is operatively connected to the apparatus (2), for producing at terminals (12) of the winding (11) a voltage correlated with the pulse, the sensor being characterized in that, in combination:

- the support (9) comprises a thin layer of ferromagnetic material having a thickness which is less than its extension in a longitudinal direction, said at least one turn of the winding (11) being extended in the longitudinal direction, the sensor (1) being able to connect to the apparatus (2) in such a way that, when the sensor (1) is operatively connected to an outer surface of the apparatus (2), said at least one turn forms a flux concatenation section positioned in a radial plane and extended in the axial direction;

- the support (9) has a transversal extension of predetermined length and is shaped in such a way as to encircle only a portion of the lateral surface of the apparatus

(2), without encircling it completely;

- the winding (11) is configured in such a way as to form a surface corresponding to a predetermined angle of less than 360 degrees, without completely surrounding the apparatus (2), when the sensor (1) is operatively connected to the outer surface of the apparatus (2).

2. A sensor (1) for detecting an electric pulse caused by a partial discharge and propagating in an axial direction in an electric apparatus (2) with cylindrical geometry, comprising a conducting wire (10) wound on a support (9) to form a winding (11) with at least one turn, so that a component of the magnetic field generated outside the apparatus (2) by said pulse, concentric with the axial direction, concatenates with the winding (11), when the sensor (1) is operatively connected to the apparatus (2), for producing at terminals (12) of the winding (11) a voltage correlated with the pulse, the sensor being characterized in that, in combination:

- the support (9) comprises a thin layer of non-ferromagnetic material having a thickness which is less than its extension in a longitudinal direction, said at least one turn of the winding (11) being extended in the longitudinal direction, the sensor (1) being able to connect to the apparatus (2) in such a way that, when the sensor (1) is operatively connected to an outer surface of the apparatus (2), said at least one turn forms a flux concatenation section positioned in a radial plane and extended in the axial direction;

- the support (9) has a transversal extension of predetermined length and is shaped in such a way as to encircle at least one portion of the lateral surface of the apparatus (2);

3. The sensor according to claim 1 or 2, wherein said at least one turn of the winding (11) is shaped in such a way as to form, a substantially rectangular flux concatenation section, with the long side approximately ten times greater than the short side, the sensor being able to connect to the apparatus in such a way that said at least one turn is positioned with the long side in the axial direction and the short side in the radial direction.

4. The sensor according to any of the foregoing claims, wherein the conducting wire (10) is wound on the support (9) with a plurality of turns positioned in such a way that the winding (11) covers a transversal extension of the support (9) to form a surface corresponding to said predetermined angle, when the sensor (1) is operatively connected to the outer surface of the apparatus (2).

5. The sensor according to any of the foregoing claims, wherein the support (9) is shaped in such a way as to encircle a portion of the lateral surface of the apparatus (2) corresponding to an angle of between 60 and 120 degrees.

6. The sensor according to any of the foregoing claims, comprising anchoring means for anchoring the sensor (1) to the apparatus (2), so that it is operatively in contact with said portion of the outer surface of the apparatus (2).

7. The sensor according to any of the foregoing claims, wherein the support (9) is flexible, to allow winding (11) curvature around the longitudinal direction in which the turns extend.

8. The sensor according to any of the foregoing claims, wherein the support (9) is shaped in such a way as to form a winding (11) curvature around the direction in which the turns extend.

9. The sensor according to any of the foregoing claims, comprising:

- a second support (9) comprising a thin layer shaped in such a way that it can be connected to the apparatus (2) to encircle at least one portion of the outer surface of the apparatus (2) corresponding to a predetermined angle like the first support; - a second winding (11) extended in the longitudinal direction and wound on the second support (9) covering a transversal extension of the support (9), the sensor being able to connect to the apparatus (2) in such a way that the winding (11) encircles said portion of the outer surface of the apparatus (2) corresponding to said predetermined angle, without completely surrounding it; - means for anchoring the first and the second supports to the apparatus (2) so that they encircle separate angular portions of the outer surface of the apparatus (2);

- an output module connected to the terminals (12) of the first and the second windings, to make available at output the voltages produced at the ends of the windings (11).

10. The sensor according to any of the foregoing claims, comprising a conducting layer (13) applied to a face of the support (9) to cover a corresponding portion of the winding (11), in such a way that the winding (11) is interposed between the outer surface of the apparatus (2) and said conducting layer (13), when the sensor (1) is operatively connected to the outer surface of the apparatus (2).

11. The sensor according to any of the foregoing claims, wherein the support (9) is shaped in such a way as to form a high-reluctance circuit for a flux of said magnetic field generated outside the apparatus (2) concentric with the axial direction.

12. The sensor according to any of the foregoing claims, wherein the assembly composed of the support (9) and the winding (11) wound round it is shaped in such a way as to form active magnetic components distributed on a surface corresponding to a predetermined angle of less than 360 degrees, without completely surrounding the apparatus (2), when the sensor (1) is operatively connected to the outer surface of the apparatus (2).

13. A process for detecting an electric pulse caused by a partial discharge and propagating along an axial direction in an electric apparatus (2) with cylindrical geometry, characterized in that it comprises the following steps:

- preparing a sensor (1) according to any of the foregoing claims;

- applying the sensor (1) to an outer surface of the apparatus (2) so that it encircles at least one portion of it in a detecting section, in such a way that a component of the magnetic field generated outside the apparatus (2) by said pulse, concentric with the axial direction, concatenates with the winding (11) to produce at winding terminals (12) a voltage correlated with the pulse;

- detecting said voltage correlated with the pulse.

14. The method according to claim 13, further comprising the following steps: - preparing a second sensor (IB) according to any of the claims from 1 to 12, the first sensor (IA) and the second sensor (IB) having predetermined winding directions for the respective turns;

- applying the second sensor (IB) to an outer surface of the apparatus (2) in a second detecting section axially spaced from the first detecting section by a predetermined quantity, in such a way that a component of the magnetic field generated outside the apparatus (2) by said pulse, concentric with the axial direction, concatenates with the winding of the second sensor to produce at the terminals of that winding a voltage correlated with the pulse;

- detecting a voltage correlated with a component of the magnetic field generated outside the apparatus by the pulse, concentric with the axial direction, said component of the magnetic field being magnetically connected to the second sensor (IB);

- comparing the sign of the voltages detected by the two sensors (IA, IB) in the same predetermined time interval in order to detect the direction from which the pulse is propagating in the apparatus so as to locate it.

15. A diagnostic instrument for detecting and locating partial discharges in an electric apparatus (2) with cylindrical geometry, generating corresponding electric pulses propagating along an axial direction in the electric apparatus (2), comprising:

- a first and a second sensor (IA, IB) according to any of the claims from 1 to 12, which can connect to the apparatus in a first and a second detecting section, respectively, being axially spaced from each other, for detecting corresponding voltage signals relating to the pulses caused by the partial discharges; - a processing unit connected to the sensors (IA, IB) to receive the voltage signals, select pairs of voltage signals caused by the same partial discharge and compare the sign of said signals to obtain information about the location of the partial discharge relative to the detecting sections.