BONDS BETWEEN COMPONENTS IN A WIND TURBINE BLADE

Technical field

The invention relates to a method for assessing bond lines that connect structural components in a wind turbine blade. The invention also relates to a device and system that may be used in the method.

Background to the invention

The use of composite materials is now common in wind turbine installations, and this is particularly so with wind turbine blades. The design of wind turbine blades strikes a balance between aerodynamic and structural priorities. The drive is now towards larger blades in order to maximise energy capture of the wind turbine installation. Composite materials are ideally suited as a construction material since they provide high strength with relatively light weight.

The assembly of constructional elements of composite wind turbine blades is now generally carried out using adhesives. Adhesive joints have the benefit of uniform stress distribution through the joined components and less stress concentration in the bonded interface. However, such bonded joints are vulnerable to defects in the bond line or to improper curing of the adhesive. This is a problem in all joined components, but is particularly critical in the case where the adhesive is used to bond two structural components together.

A modern utility-scale wind turbine blade comprises a structural beam that typically includes a two-part hollow shell. The blade is stiffened to prevent it from bending excessively and, usually, each shell half may be bonded to a central longitudinal reinforcing spar or to one or more relatively stiff longitudinal strips or 'webs' that run along the length of the blade. In a so- called structural shell construction, to provide the blade with the necessary strength to withstand the shear forces acting on it during operation, the opposing shell halves are interconnected by a construction called a shear web. There are numerous approaches to achieving this design, and one of these approaches is shown in Figure 1.

In Figure 1 , a wind turbine blade 2 has a hollow shell structure comprising an upper half shell 4 and a lower half shell 6 that are united to form the complete shell having an aerofoil cross section. Each half shell may be a composite structure comprising inner and outer laminate layers or 'skins' 8, 10 of material, for example fibre reinforced plastic (FRP).

The upper and lower half shells 4,6 may each include a strengthening structure comprising longitudinal reinforcing elements often known as spar caps 12, each of which may run along all or part of the spanwise length of the blade 2 from the blade root towards the blade tip. The spar caps 12 may also be known by other terminology in the art such as 'beams' or 'girders'. It is preferable for the blade 2 to be extremely stiff and lightweight and for this reason the spar caps 12 may be fabricated from infused stacks of carbon fibre pultruded strips that are bonded with the outer blade skin 10, by a suitable engineering adhesive. Carbon fibre is not essential, however, but it generally preferred due to its very high strength to weight ratio. In this blade 2, the spar caps 12 are embedded in the laminated FRP layers and so form an integral part of the shells 4,6. Such a blade design is sometimes referred to as a 'structural shell'. Certain regions of the blade incorporate lightweight cores 14 such as structural foam or balsa wood that are sandwiched between the outer and inner skins 8, 10 and located in between the spars 12. Such a sandwich panel construction improves bending stiffness and reduces the risk of buckling in these regions. Similar blade structures are also known having single or multiple spar caps. Shear webs 16 extend between opposing spar caps 12 and provide a load path between them. Together, the spar caps 12 and the shear webs 16 are the main structural load bearing spine of the blade 2.

The blade is typically made using a resin-infusion process in which components of the blade are laid up in a blade mould. Firstly one or more layers of dry or pre-preg glass-fibre fabric are arranged in the mould to form the outer skin of the blade. Then, elongate panels of structural foam are arranged on top of the glass-fibre layer to form the sandwich panel cores 14 referred to in Figure 1. The foam panels are spaced apart relative to one another to define a pair of channels in between them for receiving respective spar caps.

To assemble the spar caps, a plurality of pultruded strips of CFRP may be stacked in the respective channels. It is also possible to lay pre-assembled stacks into the channels. Once the spar caps are in place, more pre-preg or dry glass-fibre fabric layers are arranged on top of the foam panels and the spar caps. This forms the inner skin of the blade. Next, vacuum bagging film is placed over the mould to cover the layup. Sealing tape is used to seal the vacuum bagging film to a flange of the mould and a vacuum pump is used to withdraw air from the interior volume between the mould and the vacuum bagging film. Once a suitable partial vacuum has been established, resin is introduced to the sealed volume at one more insertion points. The resin infuses between the various laminate layers and fills any gaps in the laminate layup. Once sufficient resin has been supplied to the mould, the mould is heated whilst the vacuum is maintained to cure the resin and bond the various layers together to form the half-shell of the blade.

The other half-shell is made according to a substantially identical process. Suitable strengthening webs are arranged with respect to the spar caps. Adhesive is then applied along the leading and trailing edges of the shells and the shells are bonded together to form the complete blade.

Other examples of rotor blades having spar caps integral with the shell are described in EP1520983, WO2006/082479 and UK Patent Application GB2497578.

Referring again to Figure 1 , the shear webs are bonded to the interior surface of the blade shells using a suitable engineering adhesive, such as an epoxy resin. Since the blade half- shells and the webs are structural components of the blades, it is critical that these components are bonded correctly and that no defects exist in the bond line. Existing techniques for evaluating the integrity of the bond line usually involve a manual process in which a technician enters the interior of the blade to inspect the bond line visually or with non-destructive test tools. In some cases, bond integrity can be inspected by passing various forms of radiation through the blade. However, such techniques are time consuming, present a risk of damaging the interior of the blade and are vulnerable to human error.

It is against this background that the invention has been devised.

Summary of the invention

In a first aspect, the invention provides a method of assessing an adhesive bond between first and second components in an interior volume of a wind turbine blade, the method comprising: moving a vehicle to a test site provided on a surface of one of the components, using the vehicle to apply an impulse force to the test site, using the vehicle to record an acoustic response data set corresponding to the impulse force applied to the test site, analysing the acoustic response and determining whether there is a defect in the bond between the components at the test site based on the analysis. The method thereby allows bond inspection, in particular bond integrity inspection. The invention has a particular benefit in circumstances in which the vehicle is used to crawl within the interior volume of a wind turbine blade in order to assess the adhesive bonding between components, particularly structural components, which would otherwise be very difficult to reach given the restricted space within the blade. The invention therefore makes the final inspection of the wind turbine blade more efficient for technicians and also avoids the risk of causing damage to the blade, which may occur when the interior of the blade has to be treated as a crawlspace. Although the vehicle may be used to assess the bond at a single test site, it is envisaged that the invention will be most helpful to assess a bond line that extends along the length of a structural component, for example a bond line that connects a shear web to a shell portion. A method according to the invention is defined in appended claim 1. Further optional aspects thereof are defined in appended subclaims 2- 10.

The invention also resides in a vehicle that is configured to be used in the method as described above, thereby for inspecting an adhesive bond joining two components in a wind turbine blade. The vehicle comprises a chassis carried on a rolling arrangement so that the chassis is movable across a testing surface; a striker carried on the chassis and configured to apply an impulse force to the test surface; and an acoustic detector carried on the chassis and configured to detect an acoustic response from the test surface following application of the impulse force on the test surface by the striker. A vehicle according to the invention is defined in appended claim 1 1. Further optional features thereof are defined in appended claims 1 1-14.

According to aspects of the method, the test vehicle may be driven to the test site using a remote control means. This may take different forms. In one embodiment the remote control means includes an extendable push rod arrangement in which multiple rods may be connected to one another and used to push the vehicle inside the blade, and also to retrieve the vehicle from the blade. The rods may carry a marking pattern so that the distance that the vehicle is pushed into the blade to the test site from a reference position can be measured.

The remote control means may alternatively take the form of a cable drive arrangement. This may be controlled by a technician or by a computer. If manually controlled, a suitable marking pattern may be provided on the cable so that the vehicle's position can be determined. Alternatively, if computer controlled, the spool out of the cable may be monitored via a computer. In a further aspect, the vehicle of the invention may be configured to co-operate with a cable drive arrangement. By way of example, the vehicle may comprise a releasable cable attachment element at one or both ends thereof. An attachment element may comprise a loop or engagement pin or threaded element or a hook or any other suitable connector. In embodiments, the inspection method may include the step of moving the vehicle in one direction using a cable drive system connected to the vehicle and then moving the vehicle in an opposite direction using the cable drive system. The cable drive system may thereby comprise a cable or a set of cables capable of pulling the vehicle in a first direction and in a second direction opposite to the first direction.

Another alternative is for the vehicle to be equipped with an electric drive system that is operated by a remote control unit. The vehicle can therefore be driven and steered inside the blade as desired.

As mentioned, the remote control means may be used to measure the distance between the test vehicle and a reference position. This measurement can be used to determine whether the vehicle is in the correct position for performing a test on the component. The vehicle can be used to perform a test at a single test site at a known location, or at a plurality of test sites in sequence. The vehicle may travel along the test surface to provide a 'full coverage' of the integrity of the bond line, so the measured distance defines where the test sites will be. Alternatively, the location of the test sites may already be defined, for example by a test schedule that defines locations along the bond line that should be tested. According to the method of the invention, the vehicle may be used for carrying out a set of inspection tests at bond location. A bond location may in particular be a bond area or a bond line. A bond line inspection test at a blade may include running inspections along multiple bond lines. In some cases, a single bond line may give rise to two tests, one along each lateral side of a bond line, perhaps separated by a panel rising up along a generally central region of a bond line. A test vehicle may be moved from one bond line to a successive bond line by an operator. Apparatus may be in place in a wind turbine blade allowing a vehicle to be drawn or pushed along respective bond lines.

In aspects of the method, accuracy of the testing method may be enhanced by introducing a gas into the interior volume of the blade. In one embodiment a gas, which may be air, is introduced into the blade in order to increase the pressure inside the blade. This will tend to drive gas into any defects in the bond thereby improving their detectability. In other, alternative embodiment, gas may be introduced to the blade without increasing the interior pressure, and a gas other than air may be used. In further optional aspects, the testing method may be particularly suited to assessing or inspecting bonds between structural components since failure of these components is a critical factor assessing blade life. In the illustrated embodiment, the components are associated with a web such as a shear web and spar cap structure of the blade, such that the first component is a web such as a shear web and the second component is a shell portion of the blade, which may include spar caps within the body of the shell.

Where they are bonded together, the first component and the second component may be generally parallel in the vicinity of the test site.

Where the first portion is a web, this component may extend transversely to the shell portion. In aspects of the method of the invention, the method may include rotating the wind turbine blade such that a channel is formed at least partially by the first component, along which the test vehicle can be moved. Preferably, a bond line to be tested lies at or nearby a nadir region of the channel. Preferably, the test vehicle is maintained in the channel and in the vicinity of a relevant bond line by the effect of gravity on it. The method may additionally include moving the test vehicle along the channel thereby formed and performing a bond assessment or inspection at one or more bond test locations along the channel. As mentioned, preferably, the bond line to be inspected may run along a nadir region of the channel formed by the rotation of the blade. Advantageously, the effect of gravity on the vehicle acts to keep it on a path following the bond line as it advances along the channel. In embodiments, rotating the blade so that a channel is formed by surfaces running adjacent a bond line may ensure that gravity acts on the vehicle to maintain its position in relation to a surface which runs along one side of the bond line thereby effectively using that surface as a guide for the vehicle in relation to the bond line as it progresses along the bond line. In essence, in the context of a channel formed between a shear web and a flange, the effect of gravity causes the vehicle to 'hug' the web as it moves along the channel, and so guards against the vehicle veering off course. Optionally, a channel may be defined by the first component only, for example as would be the case if the first component has a cross section that is somewhat L-shaped in form, or between the first component and the second component, or between the first component and a further component, that is to say, another component that supports part of the vehicle but which is not a component that is bonded to the first component. In aspects of the invention, data processing may be carried out remotely from the vehicle, for example at a test computer. Thus the vehicle may be configured to transmit test data to the test computer. Therefore, in a further aspect, the invention may comprise system as defined in appended claim 15.

Since data processing can be carried out remotely from the vehicle, the vehicle may optionally be equipped with a control unit configured to control the striker and to process the acoustic response into a data set. The size and weight of the vehicle can therefore be kept low.

In alternative aspects, in order that the vehicle keeps on the right path as it travels along the surface, the rolling arrangement may be biased to steer the vehicle to one side as the vehicle is moved. The is particularly useful in the case where the vehicle is driven along next to a shear web, as the vehicle will gently turn into the web and will not veer off course.

The rolling arrangement may comprise at least one pair of wheels, each wheel having a respective rolling axis. The rolling axis of one wheel of the pair of wheels may be angularly offset from the rolling axis of the other wheel of the pair. The angular offset may be an inclined angle so that one wheel is tilted off the vertical plant relative to the other wheel.

Brief description of the drawings

Reference has already been made to Figure 1 , which illustrates a known construction of a wind turbine blade. In order for the invention to be more fully understood, it will now be described by way of example only with reference to the following drawings, in which:

Figure 2 is a perspective view of a test set up that includes a wind turbine blade that has been configured for structural testing;

Figure 3 is a view of the interior volume of the wind turbine blade in Figure 2 and shows a test vehicle travelling along a web of the blade;

Figure 4 is an enlarged view of the joint between the web and the blade in Figure 3;

Figure 5 is a side view of the test vehicle in Figure 3, in communication with an associated test computer; Figure 6 is a top view of the vehicle in Figure 5;

Figure 7 is a top view, like that of Figure 5, of an alternative test vehicle that features steering;

Figure 8 is an end view of an alternative embodiment of test vehicle;

Figure 9 is a perspective view of another alternative embodiment of test vehicle shown travelling along a web, whereas Figure 10 is a corresponding end view of the vehicle in Figure 9;

Figure 1 1 is a perspective view of another alternative embodiment of test vehicle similar to that of Figure 9, whereas Figure 12 is a corresponding end view of the vehicle in Figure 1 1 ;

Figure 13 is a flow diagram of an embodiment of a test method; and

Figures 14a and 14b are plots illustrating acoustic responses measured at a fault bond and a good bond, respectively.

Detailed description of embodiments of the invention

With reference to Figure 2, a wind turbine blade 20 is shown installed in a test set-up 22 in which the blade 20 is supported in an off-ground position by first and second supports 24, 26. The supports 24, 26 are spaced apart along the length of the blade 20, one being located at a root end 28 of the blade 2 and the other being located approximately two-thirds along the spanwise length of the blade 20 towards a tip end 30. For the purposes of this description, the blade 20 can be considered to be of the type constructed in accordance with the so-called 'structural shell' concept as described previously with reference to Figure 1 although the invention may be applied to any type of blade construction in which bonds or bond lines are to be inspected.

The test set-up 22 allows a variety of external finishing procedures to be applied to the blade 20 as required. For example, any areas of damage to the blade's surface when it is de- moulded can be filled and finished and visual inspection of any lightning protection equipment such as receptors and metallic meshes can be carried out. The test set-up 22 also enables technicians to enter the interior volume of the blade 20, for example in order to carry out visual checks to various internal components.

The supports 24, 26 are configured so that the blade 20 can be rotated about its major axis L that extends along the spanwise length of the blade. This enables technicians to access the entire external surface of the blade 20 for inspection and repair purposes. Although not shown in detail here, it should be noted that the supports 24, 26 comprise a suitable clamping mechanism to hold a respective portion of the blade securely. The clamping mechanism is then rotatable by an electric drive system with respect to a stationary outer frame to achieve rotation of the blade. Only one of the supports 24, 26 needs to be driven. In this test set-up 22 it is envisaged that the first support 24 supporting the root end of the blade is electrically driven whereas the second support 26 simply allows the blade to rotate in response to the driving of the first support 26. Such supports are generally known in the field of composite wind turbine blade production.

The test set-up 22 includes a structural component testing system 32 which, in overview and also with reference to Figure 3, includes a movable testing vehicle 34, a vehicle drive means 36 and a test computer 38.

The vehicle 34 is movable along a test surface 40 in the interior volume of the blade 20. In this embodiment, the test surface 40 is a surface of a structural component, more specifically a flanged web 42 which may be called a shear web. The web 42 is fixed by bonding to the interior surface of blade 20 in the same way as has been described previously with reference to Figure 1 , although Figure 4 shows the adhesive joint between the web 42 and the interior surface of the blade 20 in more detail.

In Figure 4, the blade 20 is represented by a section of blade shell 44, which includes an elongate spar structure or 'spar cap' 46 that extends along the blade shell 44 in a spanwise direction, as is common in structural shell blade designs. The web 42 is located in line with the spar cap 46 and is configured so that force is transferred from the spar cap 46, through the web 42 and to the opposing spar cap (not shown) located in a corresponding position on the other blade shell (not shown).

The shear web 42 in this embodiment is shaped like an I-beam in cross section and so comprises a central wall 48 that is capped by upper and lower transverse flanges 50, although only the lower one of the two flanges 50 can be seen in Figure 4. It should be noted here that the flange 50 in the illustrated embodiment is bisected by the web 42, but it is also possible for the flange 50 to extend only on one side of the web 42, such that the flange 50 and the web 42 together form an L-shape.

The web 42 and the blade shell 44 are bonded together by a suitable bonding agent, such as an epoxy resin, at a bond line 52 which in this embodiment extends along the length of the interface between those two components. As can be seen in Figure 4, the flange 50 of the web 42 and the blade shell 44 are generally parallel in the region of the bond line 52. The strength of the bond line 52 is critical to ensure that the spar cap 46 and the web 42 function together as a structural load bearing member for the wind turbine blade 2. Any defects in the bond line 52 may reduce its strength and affect the life of the joint. Examples of defects that are seen in such bond lines are voids, in which small air pockets are created during formation of the bond line, and so-called 'kissing bonds' in which regions of the two components are adhesively bonded but with very little strength. Kissing bonds are also sometimes known as 'zero volume bonds' and may be caused by incorrect surface preparation, contamination of the adhesive, and also by partial curing of the adhesive before the components are pressed together. It is to the assessment of such defects in the bond line 52 which the invention is directed.

To this end, the vehicle 34 is movable along the flange 50 and is equipped with test equipment to assess whether the bond line 52 sandwiched between the flange 50 and the blade shell 44 has an identifiable defect.

As can be seen in Figure 3, the flange 50 provides the test surface 40 along which the vehicle is moved to test various predetermined points or 'test sites' 54. In principle it is possible for the vehicle 34 to be moved to a single test site 54 to assess the bond line 52 at one particular point along its length. However, it is envisaged that a more practical approach will be to move the vehicle 34 along the flange 50 along the test sites 54 to assess each one in sequence. The separation of each test site 54 along the length of the flange 50 can be selected to provide a full 'picture' of the integrity of the bond line 40 and, accordingly, the test site density can be selected depending on the exact nature of a particular inspection task. In some cases it may be required that test sites should cover the entirety of a bond area and, in others, that the test sites cover only a portion of the bond area. Assessment of each test site 54 is performed using the vehicle 34 which is equipped with acoustic test equipment. In broad terms, the vehicle 34 is used to apply an impulse force to the test site 54 and then to monitor the acoustic response. The acoustic response is then analysed to distinguish between a 'good' bond between the flange 50 and the blade shell 44 and circumstances where a defect exists. Positioning of the vehicle 34 may be controlled by a drive means 36 which, in this embodiment, takes the form of a winch 56 controlled by the test computer 38 and which operates a cable 58 to drive the vehicle 34 along the flange 50 in either direction. Alternatively, the drive means 36 may comprise a cable 58 connectable to the vehicle 34. The cable 58 may also be used to monitor the distance between the vehicle and a reference location or position that may be near to the root end of the blade. Alternatively, the test computer 38 may also monitor the position of the vehicle 34 as it travels to the test site from the reference location. Alternative embodiments are envisaged in which the vehicle 34 has an on-board drive motor which is controlled by a radio transmitter or in which the vehicle 34 is driven manually by a series of rods 57 which can be connected together to form an extendable rod so as to push the vehicle along the flange 50 to any desired position. The rods also allow the vehicle's position along the flange 50 to be calculated. Other examples include using regular markings on the cable 58 connected to the vehicle 34 or using an optical distance sensing means (for example, a laser distance sensor or laser triangulation means) to measure the distance or position of vehicle in relation to the reference location.

An embodiment of the vehicle 34 is shown in more detail in Figures 5 and 6. In overview, the vehicle 34 includes a chassis 60 having a rolling arrangement 62 and a plurality of electrical components 64. The rolling arrangement 62 supports the chassis 60 on the test surface 40 so it can roll along it. In the illustrated embodiment, the rolling arrangement 62 comprises a pair of front wheels 66 and a pair of rear wheels 68 that are located close to the respective ends of the chassis 60. At each end of the chassis 60, a cable connection point 69 is provided for attaching the cable 58 to the vehicle 34.

The electrical components 64 may comprise a control unit 70, a battery pack 72, an electromagnetic hammer or 'striker' unit 74, an acoustic detector or 'microphone' 76. Optionally, a video camera 78 may be included. Those components are connected to the control unit 60 by suitable power and data leads as appropriate.

The main role of the control unit 70 is to control the application of an impulse force to the test surface 40 by the electromagnetic striker unit 74, and also to process the acoustic response of the flange 50 as detected by the microphone 76. In this embodiment, the control unit 70 does not carry out both the acquisition and analysis of the acoustic data captured by the microphone 76, but instead serves to capture the acoustic data and process it into a digital data set via a suitable sound card interface unit 80 that can be processed by the test computer 38. For convenience, the control unit 70 is provided with a wireless LAN adapter 81 and transmitter 82 through which means it is able to communicate with the test computer 38 to send acoustic data for analysis, and also to receive control commands from the test computer 38 for controlling the striker unit 74. It should be appreciated that a wired connection could also be used as an alternative to wireless. Although not used in the acquisition of acoustic data, the camera 78 may be used to inspect parts of the blade visually and also for viewing the way ahead.

The control unit 70 can be any suitable computing device. By way of example, a suitable device is a single board computer in the form of a Broadcom™ BCM2835 'system on a chip', which is more commonly known as the 'Raspberry Pi'™.

The striker unit 74 may comprise a solenoid driven striker piston 84 which may be in the form of a solid metal rod. The striker piston 84 is propelled towards the test surface 40 such that is strikes the surface with a predetermined impulse force. The striker unit may alternatively be any suitable device which, typically, employs a spring-loaded striker or 'hammer' which may be calibrated to apply a predetermined impact to an underlying surface. These devices are typically known in the art as electromagnetic hammers or strikers, an example of which is disclosed in US6848321.

The application of the impulse force to the test site 54 generates a shock wave that travels into the flange 50 producing an acoustic response which can be captured and analysed. By way of example, the acoustic response may be graphically illustrated. It has been observed that defects in the bond line 52 such as voids, contaminated regions and kissing bonds, have a discernible effect on the acoustic response that is detectable at the surface of the flange 50 by the microphone 76. Therefore, the microphone 76 should be selected to have a suitable sensitivity in a relevant acoustic frequency range to pick up the acoustic response, in which particular frequencies of interest are between approximately 200Hz and 20kHz although other frequencies may be used including frequencies outside the audible range for humans. Suitable microphones are available on the market as off the shelf components, as would be understood by the skilled person. The control unit 70 receives the raw acoustic response data from the microphone 76 and constructs a corresponding data set for a time window that starts at the point the striker unit 74 is triggered and extends for a predetermined time period in order to capture accurately the acoustic response of the flange 50. An example of a suitable time window length is around 0.05s to 0.5s, although longer and shorter time ranges may be used depending on the type of material and its acoustic properties. It will be appreciated that the precise time range that is chosen will be determined by the characteristics of the material being tested, for example the material composition, e.g. glass or carbon fibre composite, the thickness of the material, the type of resin used and so on, since these characteristics may affect the acoustic response. Note that it is envisaged that one technique may be to select the time period to be relatively long and then to target a sub-range of the time period using sampling software. Once the data set has been acquired, it may then be sent to the test computer 38 for analysis in order to identify any defects that may exist.

An example of a suitable test procedure conducted according to an embodiment of the invention ad using the vehicle 34 will now be described with reference to Figure 13.

The test computer 38 may preferably initiate the test procedure at step 200 by sending a command to the control unit 70 to trigger the striker unit 74. The control unit 70 triggers the striker unit 74 at step 202 and then enables the sound card interface 80 to acquire an acoustic response data set at step 204. At this point, the sound card interface 80 applies pre-processing to the acoustic data set in order to remove unwanted frequencies that may obscure the frequencies of interest. For example, in a relatively noise factory environment, the pre-processing may filter the data set to remove frequency components of below approximately 300Hz, although this is simply given here by way of example. It should be noted here that although the pre-processing step has been described as occurring at the control unit 70, it is also envisaged that pre-processing may be carried out by test computer 38 as part of the main data analysis step, as described below. There may also be carried out data sampling as part of this pre-processing step, according to which certain portions of the detected acoustic response signal are ignored, such as, for example, a beginning and end of the acoustic response signal, in cases where the signal which may be of interest is expected to lie, or does lie, somewhere in between the beginning of the response signal and its end.

Once an acoustic response data set has been acquired, the control unit 60 sends the acoustic data set to the test computer 38 at step 206 via the wireless LAN adapter 81 which is then received at the test computer at step 208. This and any other data may alternatively be transferred by a wired connection. At this point, the test computer 38 tags the acoustic data set with an identifier which may optionally correlate the data set to the test site, for example by assigning a number corresponding to the test site number. As a result, the data set can be linked to the correct span wise position of the vehicle 34 on the flange 50. This is made easier in cases where the test computer 38 also controls the positioning of the vehicle 34 along the flange 50 by way of the optional drive means 16. It should be noted however that if the vehicle 34 is driven manually, the test computer 38 may be provided with a data input facility so that a technician is able to input the test site number and distance along the flange 36 manually. The data set is then stored as a file in computer memory 210 for later processing. Note that in other embodiments, a predetermined sequential number may be assigned to successive data sets which may match predetermined sequential test site numbering.

Following the data acquisition, the test computer 38 may determine whether there are more test sites available at decision step 212. This may be achieved through a test plan internal to the computer 38 which dictates how many test sites need to be assessed and also the distance between those test sites. Alternatively, it could be a manual process in which a technician enters a command into the test computer 38 that the vehicle should move to the next test site, or move a predetermined distance.

If it is determined that there is another test site available, the test computer 38 may move on to step 214 at which, if applicable, it may move the test vehicle 34 to the next test site using the drive means 36. Once the vehicle 34 is at the next test site, the process moves to step 200 to command the striker unit to apply an impulse force, as before. The process loops through steps 200 to 214 until results have been obtained from each of the relevant test sites.

Once all test sites have been visited or tested, the process moves out of the loop at decision step 212 to analyse the data sets at step 216. Here, each data set that is stored in memory 210 is retrieved and processed to convert the raw data set into an RMS energy value indicative of the acoustic response. The RMS energy value is then compared to a predetermined threshold in order to make a decision on whether the test site is associated with a defective bond, possibly indicating a defective bond line or a defective portion of a bond line. Alternatively, a determination of whether a bond at a given test site is defective may be made immediately after each or any individual test. Figures 14a and 14b are provided by way of example to illustrate the difference between the acoustic response of a reference signal, which indicates a good bond line, and an acoustic response in the case where the bond at the test site is faulty. The reference acoustic response may be made on a test rig in which a comparable bond area is established using the same materials and test equipment. A comparison may therefore be made between the measured acoustic response and the reference acoustic response in order to evaluate the difference between them. If the two acoustic responses differ by more than is deemed acceptable, then it may be inferred that the bond is either defective at the location in question, or that it is at least not as structurally sound as the reference article. Figure 14a represents the acoustic response measured at a test site of a faulty bond, where it can be seen that the response has spiked in the frequency region of 1 kHz at a energy level of approximately -25dB. Figure 14b represents the acoustic response measured at a test site of a good bond, where it can be seen that at the same frequency range of around 1 kHz, the energy level is around -50dB. Thus, since there is an approximately 25db difference in the two test measurements, it is feasible to distinguishing between good and faulty bond lines. An appropriate threshold may be set at a level at which a reliable determination can be made between measurements made of good and faulty bonds.

Once all of the data sets have been analysed, the plurality of RMS energy values associated with each test site may advantageously be compiled into a test report at step 218 which indicates where the bond line has been assessed as defective.

The process described above illustrates how the vehicle 34 can be moved between a sequence of test sites in order to assess the underlying bond. It should be noted that the process could be repeated for adjacent structural components, for example in a structural shell blade which includes two, three or even more shear webs that run along the blade. In such a case, the blade may include integrated pulleys at the far end of each structural component which support a suitable cable arrangement so that the vehicle can be attached to a suitable cable and moved along each component in turn for testing.

The above discussion explains how the testing process is conducted by the vehicle 34 and test computer 38 in order to identify defects along the bond line 52. Although the applicant has determined that it is viable to distinguish between the acoustic response of a defective part of the bond line 52 and a 'good' part of the bond line 52 whilst the internal volume of the blade is at atmospheric pressure, it is also proposed to pressurise the internal volume of the blade to above atmospheric pressure in order to improve the transmission of the acoustic response to the microphone 76. This is shown in the illustrated embodiment in Figure 2, wherein it can be seen that a blower unit 90 feeds into the internal volume of the blade 20 through an air tight bulkhead 92 located at the root end 28 of the blade 20. In particular, by increasing the ambient pressure inside the blade shell, there may be achieved a better penetration of gas such as air into voids within the bond line. This improves the acoustic response which may be obtained when carrying out testing at or near a void in the bond. In some tests, pressure increases of between about 0.05 bar to 1 bar above atmospheric have been found to be useful, although alternative pressure levels may be used depending, for example, in the nature of the materials being tested.

In this embodiment, the blower unit 90 feeds air into the blade. However, it would also be feasible for gases other than air to be introduced into the blade, in which case the blower unit 90 could be swapped for, or supplemented with, pressurised gas tanks for example. It is also envisaged that gas could be introduced to the blade without raising the pressure. In alternative embodiments, a pressurising pump may be installed in place of or in addition to a blower.

In the above process, Figures 5 and 6 illustrate one embodiment of a vehicle that could be used. Variants of the vehicle will now be described with reference to Figures 7 to 12. Where appropriate, the same reference numerals will be used to refer to same or similar components.

Figure 7 illustrates one way in which the vehicle 34 may be provided with a steering function so that the vehicle 34 is directionally biased towards one side as it moves along the test surface 40. With a wall portion or raised portion along one side of the bond line under investigation, this ensures the vehicle maintains its position in relation to the wall or raised portion and thereby also in relation to a bond line running along or in some predetermine relation to the wall. This is to guard against the vehicle 34 inadvertently rolling off the test surface 40 thereby spoiling the test data. In this embodiment, the vehicle 34 includes a front wheel bogie 100 and a rear wheel bogie 102, each of the wheel bogies supporting a pair of wheels. Advantageously, one or both bogies may be pivotably mounted to adjacent ends of the chassis 60 at a respective pivot 104 which may e.g. define a vertical axis. A steering angle which the bogies 100,102 may define with respect to the chassis 60 may be set e.g. by a lockable fastener 106 at the pivots 104 that secures the bogies 100,102 to the chassis 60. The steering angle of the bogies may therefore be adjusted and set by appropriate tightening of their respective lockable fastener 106 or other suitable means. Here it is shown that both bogies 100,102 are pivotably mounted, although it should be appreciated that it would also be acceptable for one or the other of the two bogies to be pivotably mounted.

It should be appreciated that the steering function may be achieved in other ways. For example, the nearside wheels, that is to say the wheels that are nearest to the central wall 48 of the web, as marked on Figure 7 as '110', may be configured such their rolling resistance is greater than that of the offside wheels 112. This will tend to make the vehicle steer slightly towards the central wall 48. It is also envisaged that a similar effect could be achieved by the nearside wheels 1 10 having a slightly smaller diameter than that of the offside wheels 112. More complex steering mechanisms would also be possible, for example servo-operated steering mechanisms.

A further alternative vehicle is shown in Figure 8, which is substantially the same as the vehicle in Figures 5 and 6, apart from the configuration of the nearside wheels 110. In the vehicle of Figure 8, the nearside wheels 1 10 are not parallel to the offside wheels 1 12 as in the previous embodiment, but instead the rolling axis X1 of the nearside wheels 110 is inclined with respect to the rolling axis X2 of the offside wheels 1 12 which is parallel to the flange 50. The effect of this is that the nearside wheels 1 10 are able to track along the internal elbow defined between the flange 50 and the central wall 48 whilst the chassis 60 stays oriented in parallel with the test surface 40.

Another vehicle variant is shown in Figures 9 and 10. Here, the vehicle 34 has a chassis 60 which is similarly configured to the chassis in Figures 5 and 6. However, in this embodiment, the nearside wheels 1 10 are oriented so that they run along a central wall 48 or lateral raised portion along the bond line rather than the flange 50, whilst the offside wheels 1 12 still run along the flange 50. To this end, the nearside wheels 1 10 are preferably mounted in respective axle blocks 1 14 located at front and rear end of the chassis 60. Expressed another way, the rolling axis X1 of the nearside wheels may be generally parallel to the central wall 48, or non-parallel to the flange 50, while the rolling axis of the offside wheels 1 12 may be generally parallel to the flange 50 and, thus, also to any more or less planar surface under the vehicle 34.

The vehicle of this embodiment is adapted to be used in variant of the testing method described above in that its configuration is optimised to run along a channel 120 defined by the central wall 48 and the flange 50. The channel is generally V-shaped since the central wall 48 and the flange 50 extend transversely to one another. In other embodiments, it is envisaged that a similar channel could be formed between two different components, for example the central wall and the underlying blade which are the two bonded components. It is also possible for the vehicle 34 to be supported partly by a further component which is not part of the two bonded components. For example, the vehicle 34 may be of a size that the offside wheels extend beyond the edge of the flange 50 and are supported either by the blade shell 44 or by a protective sheet of some sort which is not itself bonded to either of the first or second components.

To form the channel, the blade 20 may be rotated along its major axis L for example by way of the rotatable supports 24,26 so that the flange 50 of the web 42 is inclined in relation to the horizontal plane H, as is illustrated in Figure 10 by the reference A. This measure ensures that gravity keeps the vehicle 34 located above or adjacent the bond line 52 when moving from one test site to the next and, as a result, the vehicle 34 tracks along the channel 120 and does not veer off course. A suitable inclination between the flange 50 and the horizon plane H may be approximately 10° or more, for example 20° or 30°, or approximately 45°. Any angle of inclination between the flange 50 and the horizontal plane H may be selected, provided the angle would ensure that the vehicle 34 will be kept, by the effect of gravity, in a desired positional relationship to the bond line as it travels along the bond line from one test site to another.

A variant is illustrated in Figures 11 and 12, in which embodiment the chassis 60 of the vehicle 34 comprises first and second elongate chassis portions 122, 124 that are orientated at right angles to each other, thereby defining an L-shape in cross section. It should be noted that the provision of vehicle wheels whose rotational axis lies at an angle offset from the main axis of a vehicle chassis is not an essential aspect of the invention. Satisfactory results may also be obtained using a vehicle with wheels aligned more conventionally along a same or similar axis of rotation. For example, satisfactory results may be obtained in an inspection method in which a measuring vehicle has conventionally arranged wheels. Advantageously, the blade structure may be presented to the inspection vehicle at an angle which places the bond line to be examined at the bottom or 'nadir' of a trough/channel created by virtue of the angle at which the blade is held. This mode of operation may be beneficial in connection both with a conventional vehicle arrangement and in connection with a vehicle wheel arrangement in which one or more rotational axes are offset from each other or from a principal plane of the chassis. In this embodiment, the first chassis portion 122 is aligned with the central wall 48 of the web 42 and is supported by the nearside wheels 1 10 that are mounted to the first chassis portion 122 on respective axle blocks 130. Similarly, the second chassis portion 124 is aligned with the flange 50 and is supported by the offside wheels 112. However, it should be noted that in this embodiment the offside wheels 1 12 are located at an inboard position on the second chassis portion 124 and are rotatably supported in wheel slots 132 defined therein. Changing the location of the offside wheels 112 to an inboard position enables the striker unit 74 to be positioned so that the striker piston 84 makes contact with the test surface 40 at an outboard position relative to the wheels. Applying the impulse force using the striker unit at a position towards the edge of the flange 50 may provide a clearer acoustic response in certain circumstances, depending on the material being tested, for example. In other embodiments, it is envisaged that a clear enough acoustic response may be obtained by striking the wall or flange at a point adjacent a bond line even where it is not directly at or on a bond line.

In this embodiment, it is also possible to configure the offside wheels 1 12 so that they are inclined relative to the flange 50 in a manner similar to the embodiment of Figure 8 so that the offside wheels 112 track the internal elbow defined between the central wall 48 and the flange 50. This alternative wheel configuration is shown in Figure 12 by the wheel in dashed lines denoted by reference numeral 134.

Some variants of the specific embodiments described above have already been explained. However, the skilled person will appreciate that other modifications may be made to those embodiments without departing from the invention as defined by the claims.

In the above process, as illustrated in Figure 13, the adhesive bond is situated between the shear web 42 and the blade shell 44 and extends between the two components in a line, so it can be considered to be a bond line that is seam-like in form, as has been described above. As such, the vehicle is moved along the test surface 40 provided by the flange 50 between the test sites 54 in order to assess the region of the bond line that underlies a particular test site 54. However, it should be appreciated that the adhesive bond need not be formed in a line, although it is most appropriate for the illustrated embodiment. In other embodiments, the adhesive bond between the two components may be in the form of discrete bond regions whether they are small regions of bonds, or different parts of adjacent bond lines. Hence, the term test site should be interpreted as not being limited to only testing different parts of one bond line, but rather as testing an underlying bond region, being part of a common bond line or otherwise.