This invention relates to a method and apparatus for detecting magnetic discontinuities in magnetisable material.

It is known from U.K. Patent Specification 2188156 that discontinuities such as cracks or pits below the surface of a specimen of magnetisable material can be detected by magnetising the material and sensing variations in leakage field near the surface of the specimen. The apparatus disclosed in U.K. Patent Specification 2188156 utilises an array of flux leakage sensors which are moved over the material surface in close proximity thereto. At successive locations in the direction of movement of the array, signals produced by each of the sensors in the array are recorded and subsequently the recorded values compared using appropriate signal processing apparatus, with an "average" of values recorded for each respective sensor at sensor locations successively before and successively after .that particular location. The difference between the recorded value and the "average" gives an indication of magnetic discontinuity at a particular sensor location. Disadvantages with the apparatus and method disclosed in U.K. Patent 2188156 are two fold. Firstly, powerful signal processing apparatus is required to analyse the sensor output signals and produce a "grid reference" model corresponding to the area tested so that locations where a discontinuity has been detected by the sensors can be accurately related to their actual position over the actual area tested. Secondly, arising from the need to compare sensor signals at specific locations with signals from the same sensor at successive locations, and the results subsequently processed, the apparatus cannot be used to give a "real time" indication of discontinuities in the surface under test.

We have now devised apparatus and a method capable of detecting discontinuities in magnetisable material, which overcomes the abovementioned disadvantages.

According to a first aspect of the present invention, there is provided a method of detecting discontinuities in a magnetic material, which method comprises: a) inducing a magnetic field within said material; b) providing an array of sensors at or near the surface of said material arranged to detect flux leakage from said material, said sensors in said array being electrically coupled in respective pairs;

c) producing a differential output signal from a respective pair of said sensors, said differential signal being dependent on the difference between the output signals from the sensors comprising said respective pair; and d) actuating warning means in response to said differential output signal.

Since the sensors comprising each pair are electrically coupled, where both sensors detect the same magnetic flux (i.e. where they are both positioned over homogenous material) the signals will cancel one another out and no differential signal will be produced by the pair.

Where however one of the sensors is positioned over a discontinuity (e.g. a corrosion pit), the change in magnetic flux detected by that sensor will cause a change in the output from that sensor such that a differential output signal will be produced by the pair of sensors of which that sensor is a member.

It is clear that any variation or difference in signals produced by sensors comprising a respective pair will produce a (differential) output signal from the pair, and thereby actuate the warning means. It is preferred that the warning means is arranged to be actuated however when the differential output signal from a respective pair of sensors reaches or exceeds a predetermined (threshold) value. Advantageously, this predetermined (threshold) value is variable. In this instance any differential output signal produced by respective pairs will only actuate the warning means when the output signal is greater than a predetermined threshold level. This enables the method to be "fine tuned" to detect discontinuities representing cracks or pits greater than a pre-determined physical size, and to ignore noise less than a pre-determined level.

Typically, each respective pair is connected to warning means, preferably an audible or visible warning indicator arranged to detect an output signed from a respective pair of sensors. Advantageously each respective pair is connected to a respective warning indicator.

The array is typically mounted on a carriage which is moved over the surface of the material; once a warning indicator has been actuated, the carriage is stopped automatically and the two possible positions (i.e. corresponding to the respective positions of the members of the activated sensor pair) of the discontinuity may be physically marked on the surface of the material immediately above those two possible sites (typically by means of paint mark or the like), either manually or automatically. Subsequently, after the desired

area of the surface of the material has been tested, the marked positions are re-tested by calibrated apparatus such as ultrasound non-destructive testing equipment, so as to give a quantitative value of the depth (and/or size) of the discontinuity (representing the crack or pit) below the surface of the material.

According to a second aspect of the invention, there is provided apparatus for detecting discontinuities in magnetisable material, which apparatus comprises: a) inducing means for inducing a magnetic field in said material; b) an array of sensors, each sensor arranged to detect and produce an output signal relating to, flux leakage at or near the surface of said material, said sensors being electrically coupled in respective pairs each pair of said sensors being arranged to produce a differential output signal dependent on the difference between said output signals produced by the individual sensors comprising each pair; and c) warning means arranged to be actuable in response to a differential output signal produced by a said pair of sensors.

It is preferred that the inducing means, and the array of sensors, (and also preferably the warning means) are mounted on a movable carriage or trolley, which can .be moved relative to a ground surface on wheels or the like. Advantageously, the sensors are arranged in a linear array, with preferably one sensor of a first pair of sensors being provided intermediately between the sensors comprising a second pair of sensors. It is preferred that sensors comprising a pair of sensors are separated from one another by a plurality of intermediate sensors, said intermediate sensors comprising members of other pairs of sensors. It is preferred that the sensors used are solid state, preferably of the hall effect type; alternatively magnetodiodes or induction coil sensors may be used.

Advantageously, calibration means is provided to inhibit actuation of the warning means where differential signals produced by respective pairs of sensors are below a predetermined threshold level. The calibration means is preferably adjustable such that the threshold level may be selectively varied.

Typically, the warning means will be in the form of an audible or visible indicator. Advantageously respective warning means are provided for each respective pair of sensors. In a preferred embodiment, the warning means comprise an array of warning lights, preferably arranged such that a pair of lights corresponds to a respective pair of sensors.

Typically, the magnet will be a permanent magnet, advantageously of "horseshoe" type, although an electromagnet may be used as an alternative.

It is preferred that the array of sensors is provided on a carrier member which carrier member is pivotally mounted on the carriage or trolley. Advantageously, the carrier or trolley is provided with independent drive means, such as a motor or the like arranged to power the apparatus.

The invention will now be further described in a specific embodiment by way of example only, with reference to the accompanying drawings; in which:

Figure 1 is a partly diagrammatic side view of apparatus for detecting magnetic discontinuities;

Figure 2 is a sectional plan view on the line A - A of the apparatus shown in Figure 1 ;

Figure 3 is a plan view of the apparatus shown in Figures 1 and 2; and

Figure 4 is a schematic exemplary view of the sensor arrangement in the apparatus shown in Figures 1 to 3.

Referring to the drawings, apparatus for detecting magnetic discontinuities in magnetisable material (such as corrosion pits or cracks in steel plate 16) is shown. The apparatus, generally designated 1 , comprises a carriage or trolley 2 having wheels 4. A carrier 9 is pivotally mounted to the trolley 2 by means of pivot mountings 5. The trolley 2 comprises two compartments, a forward compartment housing balancing and trigger circuits and visual alarm in the form of an array of L.E.D ^' s (best shown in Figure 3) 14. The rear compartment contains DC batteries 12, a drive motor and gearbox 3, and the trolley- drive wheels 4. The carrier 9 rides on a series of wheels 15 and seats a horseshoe magnet assembly comprising a magnet bridge 1 1 and neodymium-iron-boron rectangular magnets 13. A sensor head 6, also mounted on carrier 9, houses a linear array of thirty six Hall effect 7. It is attached to the magnet carriage by the pivot arms 8 and connected to the balance and trigger circuitry by cables 10.

The magnets 13 are seated in the carrier 9 so as to be close to the surface of a steel plate 16 under examination, the gap between the pole pieces and the surface of the steel plate typically being adjustable. The sensor head comprises a rectangular box the lower half of which is made of a material known commercially as nylatron, chosen for its resistance to wear and abrasion. A printed circuit board (not shown) to which the sensor array is

attached is fitted into the lower box and a lid of material commercially available under the trade mark Tufnol fitted to seal the sensors from dirt and moisture. The pivot arms 8 are adjusted to allow a small gap between the sensor head 6 and the steel plate 16 under examination. They allow the sensor head to ride upwards if the chamfered edges of it the sensor head 7 encounter protrusions or undulations in the steel plate surface.

The sensors in the array are electrically connected in respective pairs (referred to as differential pairs) to balance and trigger circuitry 17 by connection cables 10. As is best shown in Figure 4 (where a part of the sensor array only is shown), the pairs of differential sensors 7, are arranged in "overlapping" relationship with five sensors from adjacent differential pairs being located intermediate the sensors comprising a first differential pair. In the part of the sensor array shown in Figure 4, the sensors are therefore connected in the following pairs: 7a & 7g 7b & 7/7 ; 7c & 7i 7d & 7j 7e & 7k 7/& 71. The spacing between the- centres of adjacent sensors in the array is typically of the order of 7.5 mm, whereas the spacing between sensors comprising a differential pair is typically of the order Of 45mm. It has been observed that the magnetic field extension at the site of a discontinuity in a direction transverse to the direction of movement of the carriage may be 4 to 5 times greater that the dimension of the discontinuity in that direction, the spacing chosen for a differential pair of sensors should ensure that the signals from a typical defect do not tend to be self cancelling.

With the apparatus in position as shown in Figure 1 , the magnets 11 induce a magnetic field in the steel plate 16 under examination. Flux leakage from the plate 16 will vary depending on the presence or absence of magnetic discontinuities (such as pits or cracks) in the thickness of the plate. Each of the Hall effect sensors 7 in the array produces an electrical output signal proportional to the flux leakage at a corresponding point above the plate 16. Since the sensors 7 are arranged in differential pairs as described above, where both sensors comprising a pair lie above homogenous portions of the plate (i.e. where there are no cracks or pits) the respective output signals from each of the sensors comprising that pair will be substantially equal. In this case the differential signal produced by the balancing circuitry for that pair will be zero. Where however one of the sensors 7 comprising a pair lies over a magnetic discontinuity whilst the other sensor 7 does not, each sensor comprising the pair will produce a different output signal in response to different values of flux leakage detected. This difference in individual sensor output signal results in a differential output

signal for that respective pair of greater than zero being output from the balancing and triggering circuitry.

Differential signals which exceed a predetermined calibration threshold (to reduce the effects of noise) trigger an output voltage to two of thirty six bi-colour light emitting diodes 14 corresponding to the differential pair of sensors detecting the discontinuity, changing the colour from green (balanced) to red (discontinuity). A ten turn digital potentiometer is provided to adjust the threshold value during calibration.

In use, the apparatus is moved at a steady speed across the surface of the specimen plate by the motor or manually. In the absence of any discontinuity there will be no output from the differential pairs of sensors to the trigger circuits, and all LED's will show in the green state. As a sensor passes over a discontinuity the balance with its pair is disturbed and an output signal fed to the trigger circuit; if this exceeds the threshold level, the corresponding LED's are changed to red. The operator stops the apparatus and marks the possible location of ι * .discontinuity on the surface of the specimen plate.

In some circumstances, this apparatus has been shown to be capable of detecting corrosion pits penetrating to a depth of less than 30% of wall thickness in 6 to 10 mm plate.