Portable Shearography and Portable Electronic Speckle Pattern Interferometry: a Presentation of their Capabilities

Transcription

1 Portable Shearography and Portable Electronic Speckle Pattern Interferometry: a Presentation of their Capabilities D Findeis & J Gryzagoridis Department of Mechanical Engineering University of Cape Town, South Africa dfindeis@ebe.uct.ac.za profg@ebe.uct.ac.za ABSTRACT Optical interference techniques such as electronic speckle pattern interferometry (ESPI) and Digital Shearography have been shown to be of great value for the Non-destructive Evaluation of many new age materials. The use of composites in motor vehicles, sea craft and aircrafts including the possible use of Glare in the new A380 Airbus has resulted in the capabilities of optical interference techniques for NDE purposes receiving increased attention. The NDT Laboratory in the Department of Mechanical Engineering at the University of Cape Town has for a number of years now been involved with the research, development, and applications of optical NDE techniques such as Holographic interferometry, ESPI and Digital Shearography. This first led to the development of a portable inspection unit based on Digital Shearography and recently a portable prototype based on the principles of ESPI. This paper reports on the development of the portable ESPI unit and compares the performance of this unit with that of the portable shearography unit. This is accomplished through the inspection of selected acceptable test specimens of composite aircraft components using both systems. The results are presented and comparisons are drawn highlighting the advantages and disadvantages of these two optical NDE techniques. INTRODUCTION Non-destructive testing and evaluation techniques are widely used in the manufacturing, power, petrochemical and Aircraft industry, both as a tool to certify the integrity of manufactured components as well as for routine maintenance inspections. There are many methods employed and include dye penetrants, eddy current, ultrasound, x-ray, thermography as well as optical interference inspection techniques. The Department of Mechanical Engineering at the University of Cape Town began researching optical interference techniques, some 25 years ago. The initial work focused on the application of holographic interference techniques to inspect objects for material defects. Because the technique utilized holographic emulsions to capture either the double exposure or real-time interference patterns, the process was laborious and time consuming. With the development of computer based video capturing systems, the research shifted from the holographic techniques to the digital equivalents, namely Electronic Speckle Pattern Interferometry (ESPI) and Digital Shearography.

2 THEORY ESPI Electronic Speckle Pattern Interferometry, as the title implies, is the electronic processing of speckle patterns produced by a suitable interferometer 3. A single mode monochromatic laser is used to produce the interference speckles. The interferometer consists of 2 laser beams, generated by passing a single laser beam through a beamsplitter. One of the beams is called the object beam and is passed through a beam expander and used to illuminate the object. The second beam, called the reference beam, is directed via a set of mirrors onto the image recording medium, typically a CCD camera. This camera is focused onto the object to be tested. This will cause the beam paths of the object and reference beams to overlap. If the beam path difference between object and reference beam is within the coherent length of the laser beam, the two beams will interfere and produce a speckle pattern, which is recorded by the CCD. With the aid of a suitable framegrabber, these images are captured and stored in a PC. The set-up, as described above is depicted in Figure 1. In order to use this technique to inspect objects for defects, the object needs to be stressed during the inspection process. The methods used to stress an object include pressure, thermal, mechanical and vibration means. When the object is stressed, its surface deforms. This causes the beam path length of the object beam to change, which in turn causes the interference pattern between the object and reference beam to alter. If the speckle image captured before an object is stressed is compared with the speckle images of the object after it is stressed, areas of correlation and decorrelation can be determined. By plotting these areas, an image can be formed, similar to the one displayed in Figure 2 alongside, which reveals the familiar zebra-like fringes indicating the magnate Computer + Digitiser Video Monitor Laser Partial Camera Object Figure 1. Typical ESPI Set-up Figure 2. Typical ESPI fringe pattern around a crack in a cylindrical surface Beam Splitter Beam Expander of displacement the object underwent during the period of stressing. Equation 1 as indicated below can be used to represent this displacement 4. nλ d = (1) [ cos α + cos β ] where: d = out of plane displacement of the object due to the applied stress, α = angle between the direction of object displacement and camera viewing angle, β = angle between the direction of object displacement and object beam. λ = wavelength of the laser beam n = no of fringes counted If an object were to possess a defect, the material in the vicinity of the defect is weakened. When stressed, the defective zone of the object responds more readily to the applied stress, causing a greater

3 localized displacement. When inspecting the generated fringe pattern, localized fringe irregularities or concentrations, highlight the location of the defect. As the above technique generates a speckle pattern based on the interference of two laser wavefronts, any change in either one of the wave fronts will alter the speckle interference pattern. This implies that the reference beam path length has to remain constant during an inspection procedure. Unfortunately environmental and airborne vibrations are very often a reality during an inspection procedure. In order to eliminate the influence of these disturbances the optics and inspection objects are mounted on special vibration isolation marble or metal tables, which are suspended on a set of air cushions. Shearography Like ESPI, shearography is also a laser based interferometric inspection technique. The fundamental difference between ESPI and shearography is that Shearography uses only one laser beam which illuminates the object to be inspected, i.e. there is no reference beam. The laser light reflected off the object is then imaged through a shearing device and digitized by a video camera. The function of the shearing device is to take the incoming image and split it into two. One of the two images is then offset either horizontally or vertically with respect to the other, before being recombined back into one image. Due to the two images being offset with respect to each other, two different parts of the image of the object overlap 6 and produce a speckle interference pattern. Figure 3 depicts a typical Shearographic setup. The method used to conduct an inspection for defects within an object using Shearography is the same as previously described for ESPI. However, the fringes obtained using Shearography represent the rate of surface displacements and hence differ from the fringes produced using ESPI. A typical Shearography fringe pattern result is depicted in Figure 4. As the speckle interference patterns are formed due to the interference of the laser light reflected off two different regions of the surface of the illuminated object, this technique is less sensitive to environmental and airborne vibrations. Any global movement of the object does not influence the relative displacement of the two surface points producing an interference pattern. The technique is also capable of Computer + Digitiser Video Monitor Camera Laser Partial Object Beam Expander Figure 3. Typical shearography set-up Figure 4. Typical shearography image of an internal void tolerating minor laser beam path length variations, as there is only one laser beam used for the interference process. Mathematically 4 the fringes can be represented by Equation 2 as: 4π d φ = S λ x where: (2) φ = correlation phase,

4 d/ x = rate of displacement, S = magnitude of shear, λ = wavelength of the laser light, Equation 2 above indicates that the correlation fringes along which φ is constant, represent lines of constant displacement rates. The spacing between adjacent fringes is a function of the displacement gradient according to Equation 3, d x nλ = 2S (3) where: n = no of fringes. This implies that for a given object surface area, an increase in displacement gradient will produce a corresponding increase in number of fringes. PORTABLE NON-DESTRUCTIVE TESTING PROTOTYPES The two inspection methods described above have been proven to be most suitable for the inspection of a wide variety of objects and materials for defects which include cracks, internal voids, debonds and delaminations 1,2,3 in composites as well as wall thickness thinning due to corrosion. This work had until recently been carried out in our research laboratory on a vibration isolation table, using optics mounted on sturdy magnetic bases. It was however soon realized that there was great potential in developing portable NDT equipment based on Digital Shearography and ESPI. Portable Shearography Prototype Figure 5. Portable shearography prototype With the financial support from Armscor via a research contract it was decided to first develop a portable NDT inspection system based on Digital Shearography. As indicated earlier this configuration is less susceptible to environmental vibration and was considered to be more suited as a portable optical interference technique. The final configuration, is depicted in Figure 5 above. As can be seen, the system consists of a tripod mounted Shearography head unit, which contains a proprietary shearing configuration to allow an image to be sheared in any direction via two mechanical controls. The laser unit is also mounted into the shearography head and the illumination direction can be adjusted. The adjustments for the built in camera module are installed on the top of the removable cover. By mounting all components contribution to the optical interference process onto one common platform, a sturdy and dimensionally stable construction was achieved. Two power supplies, one to drive the camera and one to control the laser power output can be seen on the table in Figure 5. A framegrabber equipped PC with custom written software which, besides controlling all image acquisition, framegrabber brightness, contrast, and Lookup Table configurations, image processing and

5 image storage routines, also allows for double-exposure or real-time inspection modes, acts as the interface between the operator and the inspection camera. Portable ESPI Prototype Following the successful completion of the Portable Shearography prototype, the development of a ESPI counterpart was initiated. The project was broken up into three phases. The first challenge was to investigate methods to meet the vibration isolation requirements of ESPI and reduce their impact on the successful operation of the technique beyond the laboratory table. The results of these investigations were reported on at the 2002 SPIE NDT conference 2. The findings indicated that two methods showed success. The first was a laser chopping technique using a Lithium Niobate crystal which was pulsed in synch with the camera acquisition rate. This produced a pulsed laser beam with reduced optical output. The optimum driving voltage needed to change the polarity of the crystal was in the order of Volts. The second successful technique was to use a variable shutter speed CCD camera 5. With this approach the CCD integration period could be reduced via a shorter shutter period, thereby effectively freezing the object during the image acquisition time. Because this approach did not result in a loss in laser power and there were no dangerously high voltages required, it was concluded that this method of vibration isolation should be adopted. The second phase focused on developing a micro interferometer complete with laser diode, camera, beamsplitting optics, object beam expander and optical fiber for the reference beam. By using thick aluminium cross-sections and keeping the distances between the optics to a minimum, the impact of environmental vibration was further reduced. The final configuration ended up as a double tier configuration with the interferometer mounted above the camera. The diode laser was mounted at the rear of the micro interferometer. The final phase was to manufacture vibration isolation pads for the camera tripod legs. The design consisted of a disk and cup between which an aircushion was wedged. The disks were connected to each tripod leg and a layer of rubber on the underside of the cup completed the isolation pad. For additional support, the isolation pads were also connected to each other via telescopically adjustable arms. Figure 6 reveals the combination of the three phases to produce the resultant prototype. The small interferometer and camera are mounted on top of the fully adjustable tripod, which is secured to the vibration isolation pads. The same framegrabber equipped PC and custom written software as used for the Digital Shearography prototype is used. Figure 6. Portable ESPI prototype Figure 7. Pressure chamber samples.

6 RESULTS In order to compare the capabilities of the prototypes, suitable test samples with known defects had to be manufactured. It was decided to construct two pressure chambers, one defective and one defect free as seen in Figure 7 alongside. In addition, a test panel from an Oryx composite main rotor blade was prepared. A 500mm length of the rotor blade was cut off to manufacture the test sample. Five defects were introduced into this sample. Across the width of the blade three 48mm circular defects were created as follows: a 48mm circular hole saw was used to cut through and remove the skin off one of the sides. For the first hole closest to the leading edge, no honeycomb was removed. At the second hole where the blade thickness was 40 mm, 20 mm of the thickness of the honeycomb was removed. The entire honeycomb was removed beneath the third hole, which was closest to the trailing edge. Figure 8 clearly shows the three circular defects. In addition, using a steel ruler to separate the honeycomb from the upper skin in one of the defects and the lower skin in the other defect created two internal delaminations. The two marker pen lines on the honeycomb cross-section in Figure 8 points to the location of the delamination defects. The inspections were conducted outside the laboratory environment. The pressure chambers were clamped in a vise and the test sample was placed on a wooden table with the defect-free skin facing towards the prototypes and weighed down with a metal block placed on top. A retort stand was connected to the leading edge in order to prevent possible toppling of the sample during the stressing phase of the inspection process. No special form of vibration isolation was applied to either the object or the prototype units. During all inspection procedures, a hot-air blower was used to stress the object. The adjacent Figure 9 shows the inspection environment as described above. Figure 9 below shows the results obtained from the inspection of the defect-free pressure chamber. A small amount of air was removed during capture of images one and two. The ESPI result obtained is shown in Figure 9A. The symmetrical, circular fringes indicate a uniform response to the applied stress with no apparent weakness in the structure. This is supported in the fringe patterns obtained from the shearography inspections, which are shown in Figures 9B and C. The fringe pattern is the same for both, but seem to be rotated by 90 degrees between the two images. This is because for Figure 9B, the image shear was in the horizontal direction resulting in the horizontal displacement gradient being recorded, and for Figure 9C the image was sheared in the vertical direction causing the vertical displacement gradient to be recorded. As the object displacement is symmetrical, similar shearographic fringe patterns are to be expected. Figure 8. Test panel with man-made defects Figure 9. Inspection set-up

7 A B C Figure 9. Fringe pattern results of the pressure chamber. A) ESPI, B) Shearography, horizontal shear, C) Shearography, vertical shear The fringe pattern obtained from the inspection of the flawed pressure chamber on the other hand, differs from the one obtained in Figure 9A, as can be seen in Figure 10. The fringe pattern is still circular, with the exception of the small circular fringe irregularity on the central fringe, at approximately 5 o clock. This indicates the presence of a defect, which weakens the structure and causes an increased localized deflection. A similar phenomenon can be observed in the results of the investigation using Digital Shearography, as shown in Figure 11 below. The side by side double bulls-eye fringe pattern in the bottom portion of the horizontal shear fringe pattern, and the vertical double bullseye in the vertical Horizontal shear shear fringe pattern of Figure 11 reveal the location of the structurally weakening void in the end cap of the pressure chamber. Figure 10. ESPI Result, pressure chamber with defect. Vertical shear The Oryx test panel was first tested using the Portable ESPI prototype. It was soon Figure 11 Shearography results of the pressure chamber with defect. established that by capturing the first speckle pattern, selecting the real-time image processing routine and then heating up the test panel and inspecting the generated fringe patterns did not produce satisfactory results. The resultant fringe spacing was too dense to adequately reveal any clear fringe irregularities. A more suitable inspection procedure was to first heat up the test panel for approximated 6 seconds, capture the first speckle pattern and then select the real-time inspection routine. This approach captured the fringe patterns produced due to the slow cooling down of the heated panel.

8 Figures 12 and 13 depict the ESPI fringe patterns produced whilst inspecting the three varying depth circular defects. Figure 10 shows the results of the deepest and intermediate defect and Figure 13 shows the result of the intermediate and shallowest defect. The location of the defects is clearly identified by the circular fringes in both figures. It is also clearly visible that there is a relationship Figure 12. ESPI result, deepest and intermediate defect Figure 13. ESPI result, intermediate and shallow defect between the resultant fringe spacing and the depth of the defect. The far left fringe pattern in Figure10 locates the defect closest to the trailing edge and the far right fringe pattern in Figure 11 identifies the shallowest defect, closest to the leading edge. The fringe patterns on the right of Figure 12 and on the left of Figure 13 are of the intermediate defect. The results show that the spacing between fringes reduces as the defect depth increases, indicating that relative to the blade thickness, as more honeycomb material is removed, the structure is increasingly weakened. The next 2 Figures, Figure 14 and 15 are the results obtained from the inspection of the front and back delaminations. For these defects, the location of the defect cannot be exactly identified. The fringe patterns do however reveal that there is a structural defect present. In Figure 14, the right angled kinks Figure 14. ESPI result, front delamination Figure 15. ESPI result, back delamination in the fringes as well as the circular fringe in the central upper region of the image are caused by an irregular bulging of the blade skin, perfectly acceptable if one considers that the skin s thermal expansion and contraction is not restricted by the honeycomb substructure. The fringe pattern in Figure 15 also indicates a bulging of the test panel as indicated by the circular dark fringe in the upper right section of the fringe pattern. In addition, the fringe pattern also captures a bending of the delaminated section of the sample about the vertical axis; the two central vertical fringes show this. Notice the

9 closely spaced horizontal fringes along the top of the test panel in Figure 15, which are not present in Figure 14, most probably caused by the expansion of the honeycomb along the free edge. Following the ESPI inspection, the test panel was inspected using the Portable Shearography prototype. The panel was placed in the same position as for the ESPI inspections. It was also attempted to duplicate the method and duration of thermal stressing during the test procedure. However, because this was a manual process it would be fair to assume that variations occurred. The difference in the testing methodology between ESPI and Shearography was that it was possible to capture the first speckle image before the panel had been heated, and with the software set to real-time mode, observe the object s response as the heat was applied. It was of course also possible to follow the testing procedure as previously described for the ESPI inspection process. Figure 16 shows the results obtained from the inspection of the three circular defects of varying depth. Image i) on the left is of the deepest, image ii) of the intermediate and image iii) of the shallowest. As can be seen, the double bulls eye reveals the presence and location of the defect in all three images. The shallowest defect produced the least amount of fringes, as was the case for the results obtained from the ESPI inspection. Unlike ESPI, there was however no clear distinction between the i ii Figure 16. Shearography inspection results for i) Deepest, ii) Intermediate, iii) Shallowest defect intermediate and deepest defects via the number of fringes recorded. The deepest defect also produced the unusual fringe irregularity, which could be as a result of residual resin on the inside of the rotor blade skin. iii Figure 17. Shearography inspection result, front delamination. Figure 18. Shearography inspection result, back delamination. Figure 17 above is the result obtained from the inspection of the area containing the delamination between the honeycomb and the front skin. The 5 irregular fringes are as a result of localised abnormal deformation produced by the delaminated area. It is interesting to note that the rest of the surface is free of fringes indicating the absence of a displacement gradient on the remaining surface. When

10 comparing this result with the ESPI result in Figure 14 one can see that the two sets of results are comparable, bearing in mind that the ESPI fringes are a record of the surface displacement and not the displacement gradient. The result obtained from the inspection of the back skin delamination is shown in Figure 18 above. To the left of the centre of the image one can clearly see two vertical fringes, indicating a bending of the delaminated area about the vertical axis. This ties up with the findings of the ESPI investigation as seen in Figure 15. The discontinuity in the vertical fringe indicates a change in the vertical displacement, as captured in the ESPI result. The space between the two semicircular fringes along the top edge of the sample indicate a local displacement maxima, something which can be seen in the ESPI result as a semicircular fringe on the top edge, but which can easily be missed. Other than the horizontal fringe along the top right edge of the test panel image, the horizontal displacement recorded in the ESPI result in Figure 15 is not clearly shown in the Shearography result, as there is not much of a displacement gradient variation in that displacement field. DISCUSSION AND CONCLUSION From the results it is clear that both Portable ESPI and Portable Shearography prototypes are capable of successfully detecting defects in engineering components. Both systems performed well outside the laboratory environment without any special additional vibration isolation requirements. The results also highlight the difference in the information that the two inspection methods provide. Those obtained from the inspection of the pressure chamber and the circular defects clearly show that both methods are capable of indicating the location of the defects. The images further indicate that the fringe density of the ESPI results, and to a lesser extent the Shearography results, are capable of indicating the severity of the defect. Both techniques were also able to detect the presence of the internal delaminations. In this case however, the boundary of the defect could not be determined with either of the two methods. The fringes obtained from the ESPI results are possibly more informative than the Shearography results due to the fringe irregularities produced in the defect area. The fringe densities for the Shearography results are very low and care has to be taken not to mistake them for environmental influences and thus discard them. When considering the implementation of the two techniques, it is interesting to note that for ESPI the best results were obtained when the cooling process was observed. Shearography on the other hand generated good results by either viewing the effect of applying heat to the test panel or by viewing the cooling down process. REFERENCES 1. D Findeis, J Gryzagoridis, Inspection of Aircraft Components with the aid of Portable Digital Shearography, Proceedings COMADEM 1999, J D MacIntyre, R B K N Rao, pp , Coxmoor Publishing, Oxford, 1999.

11 2. D Findeis, J Gryzagoridis, D Reid Rowland, Vibration Isolation Techniques suitable for Portable Electronic Speckle Pattern Interferometry, Proceedings of Nondestructive Evaluation and Health Monitoring of Aerospace Materials and Civil Infrastructure, A L Gyekenyesi et al, Vol 4707, pp , SPIE, Washington, R K Stanley, P O Moore, P Mcintire, Nondestructive Testing Handbook, 2 nd Ed, American Society for Nondestructive Testing, R Jones, C Wykes, Holographic and Speckle Interferometry, 2 nd Ed. Cambridge University Press, R Spooren, Standard Charge-coupled Device Cameras for Video Speckle Interferometry, Optical Engineering, Vol 33, No 3, pp , G Claud, Optical Methods in Experimental Mechanics, Part 8, Experimental Techniques, Sept/Oct 2003,pp 15-18, 2003.