SPECIFICATION
Methods and apparatus for electric current measurement This invention relates to methods and apparatus for electric current measurement, and in particular to methods and apparatus for sensing the magnitude of current flowing in a conductor by the Faraday effect. The introduction of operating voltages for electric power transmission lines of the order of several hundred kilovolts has been accompanied by a significant problem relating to the monitoring and control of these lines. On the one hand, the high speed of operation required for example of protective systems to avoid severe damage in case of a fault can only be attained by the use of electronic equipment. On the other hand, such equipment is only suitable for relatively low voltage operation, and thus requires electrical isolation for use in conjunction with high voltage transmission lines.Although conventional transformers may be used for this purpose, they are very costly and inconvenient.
The possibility of using the Faraday effect to facilitate current measurement in these circumstances has been under investigation for many years. Such a technique relies upon the fact that when a beam of plane-polarised light is exposed to a magnetic field, the plane of polarisation is rotated through an angle proportional to the strength of the field (the Faraday effect). If the path traversed by the light is a closed loop around a conductor carrying an electric current, the rotation of the plane of polarisation is proportional to the magnitude of the current and the number of turns of the loop around the conductor.
Early proposals for implementing this technique involved projecting the beam from a plane-polarised laser positioned on the ground beneath a power transmission line at an assembly of glass blocks mounted on the line. The blocks were disposed on each side of the line and arranged to guide the laser beam back down towards an optical sensor/analyser on the ground. This sensor/analyser received the beam and measured the angle through which the plane of polarisation of the light was rotated in the blocks, from which the current strength could be deduced. In practice, this approach suffers serious problems from vibration and movement induced by wind or electromagnetic interaction between adjacent conductors.
The commercial production of optical fibres has revived interest in this use of the Faraday effect, for example by winding the fibre one or more times around the conductor and taking the ends of the fibre to the light source and sensor respectively. However, practical optical fibres have a significant intrinsic linear birefringence, which is subsequently augmented by linear birefingence caused by extrinsic factors such as bending and axial strain of the fibre. This birefringence reduces the change in polarisation due to the Faraday effect. It has been proposed (S.C. Rashleigh & R. Ulrich, AppLPhys.Lett. 34(11), 1st June 1979) to overcome the problem of birefringence by twisting the fibre along its own axis.However, this in turn introduces a serious sensitivity to changes in temperature - the cited authors suggest overcoming this by making the fibre coil in two sections, wound consecutively but twisted in opposite senses. As well as being difficult and complex to achieve in practice, this remedy would require an exact match in the number to twists of the fibre about its axis in the two sections.
According to a one aspect of this invention there is provided a method for sensing the magnitude of current flowing in a conductor by sensing the angle of rotation due to the Faraday effect of the plane of polarisation of a beam of linearly-polarised electromagnetic radiation which has traversed a path around the conductor, characterised in thatthe beam of radiation is caused to traverse the same path in both directions through an optical fibre which is twisted about its own axis in a known per se manner.
According to another aspect of this invention there is provided a method for sensing the magnitude of current flowing in a conductor by the Faraday effect using an optical fibre forming a loop with at least one turn around the conductor, wherein linearly-polarised electromagnetic radiation is injected into one end of the optical fibre, and the direction of the plane of polarisation of the radiation is analysed after it has traversed the fibre, whereby to sense the magnitude of the current as a function of the angle through which said plane of polarisation is rotated by the magnetic field of the current, characterised in that:
the fibre is twisted about its own axis in a known per se manner to reduce the effect of birefringence; the radiation is injected into and received for analysis from the same end of the optical fibre and coupled thereto via directional coupling means; and radiation reaching the other end of the fibre is reflected back thereinto by reflecting means disposed at that end of the fibre; whereby said radiation traverses said fibre in both directions before its direction of polarisation is analysed, thereby to reduce sensitivity to temperature.
According to a further aspect of this invention there is provided apparatus for sensing the magnitude of current flowing in a conductor by sensing the angle of rotation due to the faraday effect of the plane of polarisation of a beam of linearly-polarised electromagnetic radiation which has traversed a path around the conductor, characterised in that the beam of radiation is caused to traverse the same path in both directions through an optical fibre which is twisted about its own axis in a known per se manner.
According to another aspect of this invention there is provided apparatus for sensing the magntidue of current flowing in a conductor by the Faraday effect, comprising an optical fibre forming a loop with at least one turn around the conductor, a source of linearly-polarised electromagnetic radiation arranged to inject said radiation into one end of the optical fibre, and means arranged to sense said radiation and to analyse the direction of the plane of polarisation thereof after it has traversed the fibre, whereby to sense the magnitude of the current as a function of the angle through which said plane of polarisation is rotated by the magnetic field of the current, characterised in that:
the fibre is twisted about its own axis in a known per se manner to reduce the effect of birefringence; the radiation source and radiation sensing means are disposed at the same end of the optical fibre and coupled thereto via directional coupling means; and reflecting means is disposed at the other end of the fibre to reflect back into the fibre radiation reaching that end; whereby said radiation traverses said fibre in both directions before entering said sensing means, thereby to reduce sensitivity to temperature.
With this invention, the optical activity (circular birefringence) introduced by twisting the fibre itself is retained, thus alleviating the perturbing effects of the (linear) birefringence which results from the nature of the fibre itself and, more significantly, from winding it around the conductor. However, the temperaturesensitive rotation of the plane of polarisation of the radiation by the circular birefringence (which rotation is additional to the Faraday-effect rotation), is a reciprocal effect and is thus exactly cancelled by the two-way journey through thefibre. In contrast, the non-reciprocal Faraday-effect is not cancelled: the magnitude of this rotation is in fact doubled.
A method and apparatus for sensing the magnitude of current flowing in a conductor by the Faraday effect in accordance with this invention will now be described, by way of example, with reference to the accompanying drawings, in which:- Figure 1 is a schematic drawing showing the arrangement of the apparatus; Figure 2 shows a transverse section through a support for optical fibre in the apparatus, in an unexpanded state; Figure 3 shows a transverse section through the support in an expanded state; Figure 4 shows a longitudinal section through the support; and Figure 5 shows an arrangement for simultaneously twisting optical fibre and winding it onto the support.
Referring to Figure 1, the apparatus for sensing the magnitude of current flowing in a conductor 10 (for example an electric power transmission line operating at a tension of, typically, several hundred kilovolts) includes an optical fibre 12, chosen to have as low an intrinsic birefringence as possible, in the form of a loop having at least one complete turn around the conductor 10. If there is more than one turn, a special support 14 is used to carry the fibre 12. One end of the fibre 12 is secured to a source 16 of electromagnetic radiation comprising a helium-neon laser 18 and a quarter-pitch gradient-index rod lens 20.At its other end, the fibre 12 terminates in a plane surface which is normal to the axis of the fibre 12 and which is metallised (as indicated at 22), so that light arriving at that end after traversing the fibre 12 is reflected back into the fibre in the opposite direction. The fibre 12 may be either a monomode fibre or a multimode fibre, but in the latter case the radiation source 16 must be arranged to excite only the fundamental mode of the fibre 12. Between the radiation source 16 and the coiled section around the conductor 10, the fibre 12 is arranged to form part of a directional coupler 24. The purpose of the coupler 24 is to enable light travelling back through the fibre 12 towards the source 16 (that is, after reflection at the matallisation 22) to be extracted and thus separated from the light leaving the source 16, priorto analysis of its polarisation.The light thus extracted is directed via a second length of optical fibre 26 (forming part of the coupler 24 as described below) and a second quarter-pitch gradient-index rod lens 28 into a so-called Wollaston prism 30 which has the property of separating a beam of light into components having orthogonal directions of polarisation. These two components emerging at an angle to one another from the Wollaston prism are sensed by respective photo-receptors 32 (for example, semiconductor photo-diodes), which supply signals indicative of the intensity of each component to signal processing circuitry 34.This circuitry (which may be constructed with known techniques using, for example, digital logic components or a microprocessor) is arranged to derive the magnitude of the current /flowing in the conductor 10 from the intensities of the components I1 and I2 separated by the Wollaston prism, according to the following formula:-
where N is the number of turns of the optical fibre 12 around the conductor 10; and V is the Verdet constant for the particular optical fibre 12 and wavelength of radiation in use.
The directional coupler 24 is made in known mannerfrom the fibre 12 and from a second length of fibre 26 by juxtaposing the fibres 24 and 26 in such a way that their cores effectively form a single integral optical , channel. The spacing between the fibres and the extent of the region of contact is chosen such that light entering the coupler in one of the fibres leaves the coupler approximately evenly divided between the fibres. A directional coupler is preferred to a conventional semi-silvered mirror, since the metallisation of such a mirror severely disturbs the polarisation of light reflected from it, thus complicating the detection of change in polarisation of such light.Both types of coupler reduce the intensity of the wanted beam of light by half ; each time the coupler is encountered, resulting in an output beam for polarisation analysis of one quarter the intensity of the original input beam from the light source 16, so the semi-silvered mirror would offer no advantage in this respect.
The use of a quarter-pitch gradient-index rod lens 20 offers advantages in simplifying the mounting and interconnection of the laser 18, lens 20 and optical fibre 12. It is a property of the lens 20, arising from the variation of refractive index through it, that light entering it from the laser 18 is always brought to a focus at the same point on its right-hand surface (as viewed in Figure 1) irrespective of movement of the laser 18 transverse to the axis of the lens 20. Thus the optical fibre 12 can simply be secured permanently at this point on the lens 20, thereby obviating the risk of relative movement of the lens 20 and the fibre 12 that would exist if a conventional lens were used.Furthermore, provided the lens 20 and laser 18 are adequately supported against bending forces tending to alter the angle between their axes, the optical fibre 12 will be properly illuminated irrespective of translational vibration between the laser 18 and lens 20.
It is well known that the superimposition of a significant level of linear of linear birefringence on a relatively weak optical activity (in this case the Faraday effect) results in a substantial diminution of that optical activity. It has been suggested, as noted above, to overcome this problem by strongly twisting the fibre 12 about its own axis -this has been indicated schematically in Figure 1 by oblique hatching of the fibre 12. Such twisting introduces two additional phenomena:
a rotation of the optical axes of the fibre itself, in accordance with the twisting, in the frame of reference of the overall apparatus; and an optical activity of elasto-optic origin (circular birefringence).
These two phenomena tend to neutralise the deleterious effect of the intrinsic linear birefringence; the extrinsic linear birefringence, and in particular the birefringence induced by bending of the fibre is neutralised by the optical activity alone.
The number of twists per unit length of the fibre is chosen in relation to the value of the intrinsic birefringence of the fibre and to the radius of curvature of the fibre around the support 14. The greater the intrinsic birefrigence and/or the smaller the radius of curvature, the greater must be the rate of twist. It is desirable to use a fibre having as low an intrinsic birefringence as possible, and to wind the fibre on a support 14 of sufficiently large radius, so that the degree of twisting does not exceed the torsional strength of the fibre nor cause gradual creep of the fibre material.
However, the optical activity induced by twisting the fibre is itself sensitive to temperature, so that direction of the plane of polarisation of the light emerging from the fibre also depends on the temperature. Thus, in practice, the derivation of the magnitude of the current flowing in the conductor 10 from the angle of rotation of plane-polarised light traversing the fibre 12 from one end to the other would be severely perturbed by temperature changes.
In this invention, this problem is overcome by arranging that the light traverses the twisted fibre 12 in both directions, that is from the directional coupler 24 to the metallisation 22 and then back again. Reflecting the light back along the same path in this way results, irrespective of the temperature, in exact cancellation of the temperature-sensitive (reciprocal) rotation of the plane of polarisation induced by twisting the fibre 12. This is because the rotation is in the same sense (in the frame of reference of the light beam) for either direction of travel through the fibre, so the total rotation (in the frame of reference of the overall apparatus) for the complete two-way journey is zero.
In contrast, the apparent direction of the current's magnetic field (in the frame of reference of the light beam) is inverted when the light beam is reflected by the metallisation 22. Since the magnetic field direction determines the direction of Faraday rotation, this direction of rotation, relative to the light beam, is also inverted. However, in the frame of reference of the overall apparatus the direction of rotation remains unchanged when the direction of the light beam is reversed at the metallisation 22. Thus, the (non-reciprocal) Faraday rotations induced for each direction of travel through the twisted fibre 12 will augment one another, thereby increasing the angle of rotation for a given current magnitude and facilitating the measurement of the current to a given resolution.
Although the same effect could be obtained by twisting one half of the fibre 12 in one sense and the other half in the opposite sense, as mentioned earlier, such a configuration is more complex to manufacture and would also remain sensitive to temperature difference as between the oppositely-twisted sections of the fibre.
In practice it is envisaged that two systems such as that in Figure 1 would be required in a single current magnitude sensor, one arranged to sense currents in the range 10 to 1000 amperes and the other having fewer turns of optical fibre 12 around the conductor 10 to measure currents above 1000 amperes. Such an arrangement would be desirable to facilitate measurement of the full range of current magnitudes with the required resolution and without ambiguity.
It will be apparent that the metallisation 22 must be perpendicular to the axis of the fibre 12 to ensure that the light striking it is reflected back without loss into the fibre 12. At present this is achieved by the so-called 'score and pull' method to produce a naturally smooth and abrupt end surface.
The material and construction of the support 14 for the fibre 12 must be such as to avoid strain in the fibre 12 arising from expansion and contraction induced by changes in temperature and differences in the coefficients of expansion of the fibre 12 and the support 14. As well as the risk of breaking the fibre 12, such strain would also introduce additional birefringence and its accompanying deleterious effects on the required Faraday effect. Aform of the support 14 which satisfies this requirement by allowing the twisted fibre to rest loosely upon it is shown in Figures 2, 3 and 4.
Referring to Figures 2 and 4, the support 14 comprises a ring 40 of plastics material chosen for its low coefficient of expansion, good electrical insulating properties and moderate elasticity. The ring 40 has three grooves 42, 44 and 46 spaced axially along it, as shown in Figure 4, the central groove 44 being wider than the other two. The diameter of the floor of each groove is slightly less than the required diameter of the coil of optical fibre 12 to be formed on the support 14. A channel (not shown) is formed at an oblique angle through each wall 48 and 50 separating adjacent pairs of grooves 42/44 and 44/46 to permit the optical fibre to extend from one groove to the next. The ring 40 is out through radially at one point 52 around its circumference (see Figure 2).
Priorto winding of the optical fibre 12 onto the support 14, the support 14 is expanded slightly by separating the faces of the cut 52 and placing it on a mandrel 54 having a diameter slightly greater than the inside diameter of the ring 40 - see Figure 3. The elasticity of the material of the ring 40 is chosen appropriately to permit this. After the fibre has been wound onto the ring 40, as described below, the ring 40 is removed from the mandrel 54, whereupon the ring 40 returns to its original, smaller diameter and the faces of the cut 52 can be glued or otherwise permanently secured together.
As a result, the diameter of the turns of optical fibre 12 on the support 14 is slightly greaterthan the diameter of the support 14 itself, so the fibre 12 rests loosely on the support 14 and is unaffected by thermal expansion and contraction or mechanical vibration of the support 14. This is illustrated in Figure 4, the lower half of which shows the position of the fibre 12 while the support 14 is in its expanded state and the upper half of which shows the position of the fibre 12 after the support 14 has returned to its unexpanded state. Furthermore, assuming the fibre 12 is of uniform cross-section, the twist of the fibre 12 will naturally distribute itself uniformly along the length of the free fibre between the points where it is secured to the support 14 at the beginning and end of the fibre coil.Thus, no fixing of the fibre 12 to the support 14 at regular intervals (such as every turn) to ensure even distribution ofthetwist is required.
An apparatus for winding the optical fibre 12 onto the support 14 while at the same time applying the desired amount of twist per unit length to the fibre 12 is shown in Figure 5.
Referring to Figure 5, the apparatus comprises a spool 60 (the size of which is not critical) which carries the required length of optical fibre (already metallised at one end) and which is rotatably mounted between the arms of a yoke 62. This yoke 62 is secured to a pulley 64 mounted for rotation about a horizontal axis perpendicular to the axis of the spool 60. The support 14 is carried, as mentioned above, by the mandrel 54 to which there is secured another pulley 66 rotatable about a horizontal axis perpendicularto the axis ofthe pulley 64. A taut drive thread 68 passes around each pulley 64 and 66, extending therebetween via two guide pulleys 70 whose axes are parallel to the axis of the pulley 66.
To wind the fibre on the support 14, the unmetallised end of the fibre 12 is first threaded through a guide hole 72 in an arm 74 and then secured to the support 14 in one of the side grooves 42 or 46 (see Figure 4). The pulley 66 is then rotated either manually or with a motor (not shown) to wind an initial portion of fibre into the side groove. This first portion will eventually by unwound again for use in connecting the coil of optical fibre to the directional coupler 24 (see Figure 1), and its length is chosen accordingly.
The fibre is then lead through the channel connecting the side groove with the centre groove 44, and, after it has been secured thereto, the required length of fibre is wound into the centre groove 44 to form the optical fibre coil proper. The fibre is again secured to the centre groove 44, lead through the channel which connects with the other side groove, and then the remainder of the fibre is wound into this side groove and the free (metallised) end temporarily secured in place. The portion in this side groove can subsequently be unwound and disposed next to the length of fibre which leads back to the directional coupler 24, so as to close properly the loop formed by the fibre 12 around the conductor 10 (see Figure 1).
It can be seen from Figure 5 that as the pulley 66 is rotated to wind fibre onto the support 14, the drive thread 68 will cause the pulley 64 and yoke 62 to rotate simultaneously therewith, thus rotating the axis of the spool 60 and twisting the fibre 12 about its own axis. The rate of twisting of the fibre 12 per unit length will depend on the diameter of the (expanded) support 14 and the ratio of the diameters of the pulleys 64 and 66.
By way of example, it is envisaged that, using an optical fibre 100 microns (micrometres) in diameter and having an intrinsic birefringence of 0.4 radians/metre, a typical optical fibre coil with a diameter of 16 centimetres would have 10 twists of the fibre about its own axis per metre of its length. The number of turns of the fibre on the support 14 would depend on the current magnitude to be measured, ranging from a few tens of turns for currents of a few tens or hundreds of amperes down to as little as 1 turn for very high current magnitudes.