Aircraft fuel wing tanks are designed and built as an integral part of the aircraft frame structure. On completion of the fabrication of the wing section, the inside of the wing is sealed with resins formulated to withstand rapid temperature changes of +/-50 ° Cn whilst remaining sufficiently flexible to withstand the flexing of the wing during flight.

After a period of service small cracks begin to appear in some of the wing tank linings. Aviation fuel then begins to seep between the liner and the metallic wing skin until it eventually finds an egress point. Such egress points are typically a loose rivet or an overlap joint. The aviation fuel spreads over the surface of the wing creating what is termed in the industry as"wingwet".

As the problem becomes more severe, a drip or trickle of aviation fuel emerges from the wing tank when full.

Once the leak has been identified externally, a problem is trying to locate the breach within the resin liner, which is usually remote from the visible leakage point external to the wing structure.

Two methods are currently known for identifying the location of the internal breach in the resin liner. According to one method, a vacuum is induced in the empty wing tank and a penetrating dye is applied to the known external leakage point. With a limited possible vacuum allowed within the wing tank, the process of drawing the dye along the leak path and to the inner breach can last a number of days. Once the dye has reached the breach, the dye is visibly identified through a series of inspection covers located on the underside of the wing tank.

A second and less preferred method is to apply a vacuum to the fuel tank and to introduce helium gas to the outer leak site. With access through the inspection covers, a helium detector is used to attempt to identify first of all the wing tank compartment and then the actual site of the breach in the resin lining. Again, this process has a duration of a number of days and is not always successful.

US patent 5,341,670 describes a method for locating leakage from an above ground tank. The method proposed includes receiving acoustic signals at a first set of points. determining a phase delay between the acoustic signals to define a first arc of possible leak locations, receiving acoustic signals at a second set of points, determining the phase delay between the acoustic signals

received at the second set of points to define a second arc of possible leak locations and locating the actual position of the leak by interception of the first and second arcs. This proposed method, in common with other proposed methods for determining the existence of a leak within a storage tank by acoustic sensing, is unsuitable for identifying the course of a leakage path extending from a known external leakage point through an internal lining or a fluid storage tank.

In accordance with an aspect of the present invention there is provided a method of detecting a leakage path extending from an external leakage point through an internal lining of a fluid storage tank, said method comprising: applying a pressure differential between an inner surface of said lining and an outer surface of said storage tank ; moving acoustic transducing means over said external surface; and detecting, with said acoustic transducing means, acoustic signals characteristic of said leakage path during said movement to detect said leakage path.

By means of this aspect of the invention, the course of a leakage path even in a storage tank having a complex internal construction and having limited internal access can be accurately identified. By identifying the course of the leakage path, a repair along the entire course of one or more leakage paths from the external leakage point to one or more internal breach points can be readily and reliably effected.

In accordance with a further aspect of the invention there is provided apparatus for locating fluid leakage, said apparatus comprising a device to be held by a user, said device comprising two acoustic transducers arranged to separately pick up acoustic signals emanating from mutually spaced points on a test surface when the device is held against said test surface.

This aspect of the invention provides apparatus suitable for following the course of a leakage path being identified below the test surface. By the signals emanating from the points at which both of the acoustic transducers pick up signals as the device is moved over the test surface. the course of the leakage path can be readily identified.

In accordance with a further aspect of the invention there is provided a method of testing an aircraft component in situ, said method comprising detecting acoustic signals emanating from said component under test conditions using sensing apparatus comprising a rigid rod member held by a user, an acoustic transducer attached within a cavity inside said rod member, and monitoring means responsive to said acoustic transducer, said acoustic transducer being arranged in said cavity to sense acoustic vibrations transmitted to one end of said rod member from a test surface in preference to atmospheric vibrations and said method comprising contacting said one end to said test surface during testing of the component.

This aspect of the invention provides a sensitive and accurate method for detecting a defective or failing component in the aircraft by contacting an

acoustic sensing device. adapted to focus on sounds emanating from the component, with a test surface.

Embodiments of the invention will now be described, by way of example only, with reference to the accompanying drawings, wherein: Figure 1 is a cross-section of an aircraft wing which leaks fuel; Figure 2 is a schematic illustration of components of a test kit used in this embodiment; Figure 3 is a schematic illustration of further components of the test kit being used in this embodiment; Figure 4 is a cross-section of an embodiment of acoustic sensor which may be used to detect the course of a leak path in accordance with this embodiment; and Figure 5 is a cross-section of a further embodiment of acoustic sensor which may be used to detect the course of the leak path.

Referring to Figure 1, a common structure of aircraft wing 2 includes an outer metal skin 4, which is internally lined with a resin composition layer 6. The wing is divided into several fuel tank compartments 8, by a series of baffles 10, which include connecting passages 12 to allow the transfer of fuel between the compartments at a slow rate. The purpose of the baffles is to prevent a rush of fuel and thereby an imbalance or large transfer of weight from one side of the aircraft to the other during banked flying operations. The

compartments when full to capacity normally contain a majority of liquid fuel 8, and a small gas-filled headspace.

Figure 1 shows an example of a typical leakage path which has developed in the wing structure after a period of service. An external leakage point 14, for example a loose rivet on the underside of the wing allows fuel to escape from the tank via an internal breach 16 in the resin lining 6 and a leakage path 18. The leakage path 18 lies between the resin lining 6 and the outer skin 4 of the wing.

Figure 2 shows components of a test kit according to this embodiment of the invention. A control case 70 houses a vacuum pump 24, connected to an exhaust outlet 26, a manually-operable vacuum regulator 28, a manually- operable pressure regulator 30, a mechanical positive/negative pressure gauge 32, and two electronic pressure transducers 34,36. A canister 38 of oxygen- free nitrogen, or other suitable inert gas, is connected to an inlet port 40 of the control case. The inlet port 40 is connected to a manually-operable three position selector valve 42. In a first position, the selector valve 42 connects the inlet port 40, via the vacuum regulator 28, to the vacuum pump 24. In a second position, the selector blanks off the inlet port 40. In a third position, the selector valve 42 connects the inlet port 40 to the pressure regulator 40. A manually-operable two position selector valve 44 connects an outlet port 46 either to the vacuum pump 24 or the pressure regulator 30, depending on its position. The pressure gauge 32 is connected to the outlet port 46.

One of the pressure transducers 34 is connected to a first pressure return port 48 via a manually-operable open/close selector valve 52, depending on its position. The other pressure transducer 36 is connected to a second pressure return port 54 via a further manually-operable open/close selector valve 56, depending on its position.

The pressure transducers 34,36 are electronic pressure transducers having a displayed read out accurate to 0.1 Mb, such as a Druck (trade mark) pressure transducer, Model No DPI 700 IS. A coil of neoprene tubing 58,60, connects each pressure transducer 34,36 to its respective selector valve 52, 56. The purpose of this tubing coil is to provide a damping action in relation to the incoming pressure data, such that the coils absorb instantaneous transient fluctuations in pressure which would destabilize the reading on the pressure transducers 34,36. The coils move, i. e. expand and contract, to provide the damping action. A similar effect could be achieved by using coils of other flexible materials, such as copper, or by arranging for an amount of liquid to be present in the pressure return path close to the pressure transducers 34,36, for example liquid mercury, which would damp such incoming pressure fluctuations.

Next, the testing procedure will be described. It is a characteristic of the test that it is preferably carried out with a tank containing its normal liquid contents, i. e. aviation fuel. The tank (s) being tested are preferably filled to approximately 90% full capacity.

In the first stage of the testing procedure the outlet port 46 and the first pressure return port 48 are sealingly connected, via a T-junction and flexible pipelining, to the fuel tank vent outlet in the underside of the wing. This vent outlet is connected to a vent box. located centrally in the fuselage, via a vent line, which in turn vents the gas pressure in each of the compartments of the fuel storage tank in each wing. In normal use, the vent serves to equalise the gas pressure within the fuel storage tanks with atmospheric pressure. During the testing procedures being described, the vent port is used to alter the pressure within each of the compartments of the fuel storage tank with respect to atmospheric pressure.

During testing, the second pressure return port 54 of the control case 20 is connected, by means of a specially-adapted connector to the refuelling port, thereby to sense the liquid pressure within the fuel storage tanks, at the base of the liquid. The specially-adapted connector is similar to a standard fuelling nozzle, but connects a pressure return line to the fuelling port rather than a fuelling hose.

Once connected as described, the control case 20 is set. using the selector switch 42, the pressure regulator 30 and two-way selector switch 44, to generate a positive pressure at the outlet port 46. In this manner, the gas pressure within the fuel storage tank is increased to a predetermined positive pressure in the region of 20 to 200 Mb, preferably 50 to 150 Mb. With the tank at a positive internal pressure, the locations of external liquid leakage

points are more readily visually identified than at atmospheric pressure, due to the resulting increased liquid leakage rate. At this stage, the locations of all visually identified external liquid leakage points are marked on the wing, for example with soft chalk, by the test operator, the markings being used during a subsequent procedure to be described below.

The fuel storage tank is next sealed for a monitoring period, to determine a rate of leakage. The gas pressure within the storage tank is monitored on the first pressure transducer 34. whilst the liquid pressure at the base of the tank may be monitored with the second pressure transducer 36.

The rate at which the excess pressure within the fuel storage tank is lost is indicative of the size of the leak. A calibration with respect to the known gas volume within the fuel storage tank at maximum capacity provides a calibrated leak rate.

Furthermore, by comparing the rate of pressure loss within the gas- filled headspace with the pressure change observed in the liquid-filled portion of the tank (the respective readouts of the first pressure transducer 34 and the second pressure transducer 36), it can be identified whether an internal breach 16 is within the gas-filled headspace in the storage tank or within the liquid- filled space. If the internal breach is within the gas-filled headspace, the decrease in pressure monitored on both of the pressure transducers 34, 36 is substantially equal. On the other hand, if the internal breach is in the liquid- filled space, the pressure drop monitored on the second pressure transducer 36

(that indicating the pressure change in the liquid-filled space) is considerably greater than the pressure drop within the gas-filled headspace. This is because the pressure monitored on the second pressure transducer 36 is the sum of the pressure within the gas-filled space and the weight of the liquid above the point of monitoring. As liquid escapes, both the pressure within the gas-filled headspace and the weight of liquid above the monitoring point for the second pressure transducer 36 decrease.

In order to identify the actual height of the inner breach within the tank, the pressure within the gas-filled headspace of the tank is reduced, by appropriate setting of the switches 42,44 and the vacuum regulator 28 in the control case 20, to a negative pressure in the region of-20 to-200 Mb, preferably-50 to-150 Mb. The tank is then sealed, and the gradual reduction in gas pressure is monitored on the first pressure transducer 34. In the case of a single internal breach 16, the pressure will gradually decrease and tend towards a given negative pressure, which is lower than atmospheric pressure.

At this pressure, the height of the fuel above the intemal breach 16 counter- balances the negative pressure within the gas-filled headspace, and no further ingress of air occurs. The end vacuum pressure is proportionate to the depth of the breach below the liquid surface, which may thus be readily calculated.

It is possible that more than one breach 16 exists within the lining of the tank. In the case of a number of internal breaches (the leak path may be branched) the initial leakage rate under vacuum consists of a combination of

the leakages into each of the internal breaches 16. As the vacuum gradually decays, the lowest breach will at one point stop admitting air, due to the liquid height/gas pressure balancing effect described above. Meanwhile, the higher internal breach will continue admitting air. Thus, by monitoring the rate of leakage over a period of time, it is possible to distinguish internal breaches at a plurality of different heights within the fuel storage tank. Around each of the above-described balance-point vacuums, a charge in the gradient of the rate of leakage will indicate the presence of an additional internal breach 16.

This allows the approximate depth of each breach to be calculated.

Thus, using these procedures, the following information can be identified: (a) the locations of any external liquid leakage points; (b) a total leakage rate, indicating the seriousness of leakage from the tank : (c) whether any internal breach is in a part of the tank containing liquid or containing gas during the test; (d) in the case of a liquid leak, the depth (s) of one or more internal breaches within the liquid.

In the case of a combination of breaches spaced throughout the inside of the storage tank, the various rates of decay at both positive and negative test pressures and above the liquid surface and below the liquid surface may

be used by a test operator, using the indications referred to above, to identify the various breaches 16 and their heights within the tank.

The next stage of the testing procedure concerns identifying the course of any leakage paths within the resin lining 6. In this stage of the testing procedure, the control case 20 is set to produce a constant negative pressure, in the region of-20 to-200 Mb, preferably-50 to-150 Mb, within the gas- filled headspace of the fuel storage tank. Thus, the fuel storage tank is not sealed during the second stage of the testing procedure, but rather is subject to constant vacuuming by the vacuum pump 24, at a pressure set by manipulation of the vacuum regulator 28. This ensures that a characteristic signal, to be described below, is maintained at a substantially constant frequency during this stage of the testing procedure.

The further components of the test kit used in this stage of the procedure, referring to Figure 3, are a dual-microphone acoustic sensor 62 and/or a sensing rod 84, a two channel signal receiver and processor 64, a pair of stereo headphones 66, a signal recorder 68 and a two channel oscilloscope 70.

As shown in Figure 4, the dual-microphone acoustic sensor 62 includes two microphones 72,74 picking up signals from a test surface, in this case the wing skin 4. The microphones are spaced apart on two sides of a hand grip 76. A circular rubber flange 78 surrounds the space in front of each microphone 72,74, in order to block atmospheric sound from the listening

space defined between the microphone 72 and the test surface 4. In addition, the space adjacent to and behind each microphone 72,74 is filled with a silicone gel 80, which also tends to block atmospheric sounds reaching each microphone 72,74. Both microphones 72,74 are connected to a two channel radio transmitter 82.

The receiver 64 contains a two channel amplifier and a variable frequency selector. By varying the frequency range selected by the receiver 64, specific frequencies characteristic of the leakage path can be selected for monitoring on the headphone 66, the signal recorder 68 and the oscilloscope 70.

In particular, when the fuel tank, having a breach as illustrated in Figure 1, is depressurised, the sound of incoming air flowing along the leakage path 18 connecting the external leakage point 14 with the internal breach 16 has a characteristic frequency which is relatively high (between 10- 20 kH depending on the negative pressure used and the materials forming the wing and lining through which the leakage path 18 passes).

In order to identify the course of a leakage path 18 within the resin lining 6, the acoustic sensor 62 is first placed adjacent to the known external leakage point 14, which has been identified and marked previously by the test operator, and the frequency selector is tuned until the characteristic frequency of the incoming air flow is detected. Once the receiver 64 is tuned to a frequency range including this characteristic frequency, the acoustic sensor 62

is moved to track the course of the internal leakage path 18. The course of the path is tracked by balancing the signals picked up by each respective microphone 72,74 whilst the test operator holds the acoustic sensor 62 against the wing skin 4 as illustrated in Figure 4. The balancing of the signals picked up by the two microphones 72,74 whilst the acoustic sensor 62 is held orthogonal to the presumed course of the leak path 18 acts to maintain the leak path 18 centrally between the two microphones 72,74.

If, as the operator moves the acoustic signal 62 along the leakage path 18 being charted, the volume of sound heard in one side of the headphones 66 exceeds the volume of sound heard in the other, and/or correspondingly one of the signals shown on the oscilloscope display has a greater amplitude than the other, this indicates that the leak path is no longer central between the two microphones 72 and 74. The operator correspondingly adjusts the position of the acoustic sensor 62 to maintain balance between the two signals. During identification of the course of the leakage path 18, the operator can mark the course, for example with soft chalk, on the outside of the wing 2. The course is then charted until the internal breach 16 is reached. By, for example, rotating the acoustic sensor 62 about an axis normal to the wing surface whilst listening for airflow along a leakage path, the test operator can readily identify that the end of the leakage path has been found. The receiver 64 may also be tuned to a different frequency range, somewhat lower than the previous frequency range, to listen to a signal characteristic of bubbles of air detaching

from the inner surface of the tank within the liquid content, also indicating the end of the leakage path.

Optimally, the microphones 72,74 are spaced by approximately 20 cm, although other spacings are also possible.

A further embodiment of acoustic sensor which may be used for confirming the course of the leakage path 18 is illustrated in Figure 5. This is a sensing rod 84 having a microphone which is also be connected to the arrangement of the radio transmitter 82, the receiver 64, headphones 66, recorder 68 and oscilloscope 70 as described in relation to the dual- microphone acoustic sensor 62. In this case. only one channel is required as only a single signal is to be sensed.

This sensing rod 84 includes a cylindrical body 86 having a solid end portion 88 with a rounded base 90 and constructed from a material being non- magnetic stainless steel. The rounded base 90 is designed to pick up acoustic vibrations by direct contact with a test surface. The opposite end of the body 86 receives a microphone 92, connected as described above to a radio transmitter 82. The body has a length of approximately 30 cm and an outer diameter of approximately 13 mm. Between the microphone 92 and the solid end 88, a cylindrical resonance cavity 94, having a length of approximately 15 cm and a diameter of approximately 8.5 mm extends along the length of the body 86. The microphone is threaded into the body 86, such that vibrations transmitted along the cavity 94 are transmitted directly to and picked up by

the microphone. Furthermore, the microphone 92 is directed towards the cavity 94 within the body 86 such that sounds transmitted along the resonance cavity 94 are picked up with great sensitivity. On the other hand, atmospheric sounds emanating from outside the body 86 are not readily picked up by the microphone 92.

The sensing rod 84 provides a means of increased sensitivity for detecting vibrations within the test surface 4, in this case the aircraft wing, against which the end of the body 86 is held in contact. The sensing rod 84 senses the characteristic sound of air flowing along the leak path 18. With the receiver 64 tuned to select a frequency range spanning the characteristic frequency of the air flow along the leakage path 18, a test operator can readily detect proximity of the leakage path by contacting the rounded base with the wing surface and identify the course of the leakage path 18 by moving the sensing rod 84 over the outer surface of the wing skin 4, preferably starting from the known external leakage point 14 and ending at the internal breach 16.

Both the dual-microphone acoustic sensor 62 and the sensing rod 84 may be used in combination in order to positively identify the course of the leakage path 18. First, the dual-microphone sensor 62 is used to coarsely plot the leakage path 18. Once so plotted, the sensing rod 84, with its higher sensitivity, is used to diagonally criss-cross in a zig-zag fashion across the coarsely charted leakage path, to finely locate the path of leakage.

Once the leakage path has been identified and marked along the external surface of the tank, a corresponding part of the inner lining can be cut out and repaired once the tank is emptied, with great confidence that the leakage path 18 will have been repaired along its whole length, and that then no further leakage will occur from the external leakage point 14.

In the above-described procedure, the characteristic signal monitored on the acoustic sensors is that of air travelling along the leakage path 18. It is possible to increase the amplitude of the characteristic signals by applying liquid (preferably aviation fuel) to the external leakage point 14 during acoustic sensing, to ensure that the fluid passing along the leakage path consists of a mixture of air and liquid. This may be achieved, for example, by spraying the liquid onto the wing surface at the external leakage point during acoustic monitoring.

Other modifications and variations are possible within the scope of the invention. Furthermore, arrangements similar to the embodiments of acoustic sensor described may be used in other types of leakage sensing operations. In particular, the sensing rod described may be used to test the condition of components of an aircraft other than the fuel tanks. Condition testing may be achieved by, using the rod in combination with amplifying and recording equipment, taking recordings of the audio characteristics of these components during normal operation at periodic intervals to detect any deterioration in their condition. The components identified which are testable in such manner

include the aviation fuel systems (fuel tanks, fuel transfer lines, fuel pumps and valves, non-return valves, fuel injectors, fuel feed lines) and the hydraulic systems (valves, small and large bore pipes, solenoid valves, telescopic rams, seals. etc and other connectors and fittings). Using the detector described, the tests may be carried out with the components in situ, that is to say when normally fitted to the aircraft. This allows for regular maintenance checks to be performed conveniently.