Center for Nondestructive Evaluation Conference Papers, Posters and Presentations Center for Nondestructive Evaluation 7-2011 Ultrasonic testing of adhesive bonds of thick composites with applications to wind turbine blades Sunil Kishore Chakrapani Iowa State University, csk@iastate.edu Vinay Dayal Iowa State University, vdayal@iastate.edu Ryan Krafka Iowa State University Aaron Eldal Iowa State University Follow this and additional works at: http://lib.dr.iastate.edu/cnde_conf Part of the Materials Science and Engineering Commons, and the Structures and Materials Commons The complete bibliographic information for this item can be found at http://lib.dr.iastate.edu/ cnde_conf/35. For information on how to cite this item, please visit http://lib.dr.iastate.edu/ howtocite.html. This Conference Proceeding is brought to you for free and open access by the Center for Nondestructive Evaluation at Iowa State University Digital Repository. It has been accepted for inclusion in Center for Nondestructive Evaluation Conference Papers, Posters and Presentations by an authorized administrator of Iowa State University Digital Repository. For more information, please contact digirep@iastate.edu.
Ultrasonic testing of adhesive bonds of thick composites with applications to wind turbine blades Abstract This paper discusses the use of pulse echo based ultrasonic testing for the inspection of adhesive bonds between very thick composite plates (thickness greater than 30 mm). Large wind turbine blades use very thick composite plates for its main structural members, and the inspection of adhesive bond-line is very vital. A wide gamut of samples was created by changing the thickness of plate and the adhesive. The influence of experimental parameters such as frequency on measurement is studied in this paper. Two different frequencies are chosen, and the measurement error bars are determined experimentally. T-Ray measurements were used to verify and correct results, and conclusions are made based on the combined results. Keywords adhesive bonding, composite materials, inspection, plates (structures), ultrasonic applications, wind turbines, nondestructive evaluation, QNDE, Aerospace Engineering Disciplines Aerospace Engineering Materials Science and Engineering Structures and Materials Comments Copyright 2012 American Institute of Physics. This article may be downloaded for personal use only. Any other use requires prior permission of the author and the American Institute of Physics. This article appeared in AIP Conference Proceedings 1430 (2012): 1284 1290 and may be found at http://dx.doi.org/10.1063/1.4716366. This conference proceeding is available at Iowa State University Digital Repository: http://lib.dr.iastate.edu/cnde_conf/35
Ultrasonic testing of adhesive bonds of thick composites with applications to wind turbine blades Sunil Kishore Chakrapani, Vinay Dayal, Ryan Krafka, and Aaron Eldal Citation: AIP Conf. Proc. 1430, 1284 (2012); doi: 10.1063/1.4716366 View online: http://dx.doi.org/10.1063/1.4716366 View Table of Contents: http://proceedings.aip.org/dbt/dbt.jsp?key=apcpcs&volume=1430&issue=1 Published by the American Institute of Physics. Related Articles A study of pressure-shear vertical wave propagation in periodically layered fluid and piezoelectric structure J. Appl. Phys. 113, 054903 (2013) Lamb wave near field enhancements for surface breaking defects in plates J. Appl. Phys. 111, 104906 (2012) Extraordinary acoustic shielding by a monolayer of periodical polymethyl methacrylate cylinders immersed in water J. Appl. Phys. 110, 014509 (2011) Meandering collimated beam of surface acoustic waves on a trigonal crystal ball Appl. Phys. Lett. 98, 123508 (2011) Interferometric imaging of surface acoustic waves on a glass sphere J. Appl. Phys. 108, 123508 (2010) Additional information on AIP Conf. Proc. Journal Homepage: http://proceedings.aip.org/ Journal Information: http://proceedings.aip.org/about/about_the_proceedings Top downloads: http://proceedings.aip.org/dbt/most_downloaded.jsp?key=apcpcs Information for Authors: http://proceedings.aip.org/authors/information_for_authors
ULTRASONIC TESTING OF ADHESIVE BONDS OF THICK COMPOSITES WITH APPLICATIONS TO WIND TURBINE BLADES Sunil Kishore Chakrapani 1, Vinay Dayal 1, Ryan Krafka 1, and Aaron Eldal 1 1 Center for Nondestructive Evaluation and Department of Aerospace Engineering, Iowa State University, Ames, IA 50011, USA ABSTRACT. This paper discusses the use of pulse echo based ultrasonic testing for the inspection of adhesive bonds between very thick composite plates (thickness greater than 30 mm). Large wind turbine blades use very thick composite plates for its main structural members, and the inspection of adhesive bond-line is very vital. A wide gamut of samples was created by changing the thickness of plate and the adhesive. The influence of experimental parameters such as frequency on measurement is studied in this paper. Two different frequencies are chosen, and the measurement error bars are determined experimentally. T-Ray measurements were used to verify and correct results, and conclusions are made based on the combined results. Keywords: Ultrasound, Adhesive Bonds, Wind Turbine Blades, Thickness Measurement PACS: 43.35.Zc, 43.35.Pt INTRODUCTION A typical wind turbine blade is manufactured as two halves and bonded together using high performance adhesive. When the two halves are joined, a shear web is placed in between two spar caps to create what is essentially an I-beam configuration along the entire span of the blade. This is done in order to provide required bending stiffness to the blade. As such, bond-lines extend along the trailing edge, leading edge, and the shear web/spar cap joints. The shear web and spar cap joints are widely considered to be the most vital joints in the structure of wind turbine blades. The inspection of such a joint for lack of adhesive, or thickness of adhesive, becomes very important, since a failure in one of these structures could result in loss of blade bending stiffness and potentially break the blade into two halves. Hence, the study of bond integrity in these joints becomes vital. Another challenging aspect of inspection is bond thickness variation. Uneven adhesive distribution will result in a poor joint between the structures. Hence, measuring the thickness of the bond becomes important. The use of nondestructive evaluation to test and evaluate bonded joints is very common. A common NDE technique used in factory is contact ultrasonics. Most of the previous work has been performed in the area of thin laminates or thin composites, and in conventional materials such as steel and aluminum [1][2]. There are two methods which have been proven in theory to work well in bond thickness and strength measurement, which are Lamb wave based approach, and pulse echo using bulk waves [3][4][5]. Lamb waves and Rayleigh waves have been shown to work effectively to measure bond thickness in both Review of Progress in Quantitative Nondestructive Evaluation AIP Conf. Proc. 1430, 1284-1290 (2012); doi: 10.1063/1.4716366 2012 American Institute of Physics 978-0-7354-1013-8/$30.00 1284
bonded composite and metal plates. This practice is limited to laboratory measurements and is difficult to implement on shop floor. On the other hand, the use of pulse echo technique is the simplest for both evaluation of bond strength or integrity, and bond thickness measurements. Additionally the technique is field implementable. Velocity measurements and scattering of ultrasonic signals reveal the strength of the bond, while, the time of flight data reveals the thickness of the bond. The same concept has been extended to thick composite spar caps of thickness greater than 30 mm. Although the concept can be extended, testing thick bonded composites becomes challenging due to attenuation and damping characteristics of composites. The objective of this research is to find an optimum frequency for inspection of bonded joints in wind turbine blades. Measurement of bond thickness to determine bond thickness variation is focused in this paper. To facilitate this, a scanning method using contact transducers was devised and tested. Furthermore the same method can be extended to perform C Scans on the bonded joint sections to determine defects such as lack of adhesive, voids in adhesive, bond thickness variation. METHOD AND THEORY Optimization of frequency was the first step in the methodology. Contact transducers of frequency range from 100 KHz to 2.25 MHz were used to optimize the ultrasonic signal. At 500 KHz, a strong signal with less attenuation was observed in the spar cap sections. A schematic of the spar cap/shear web bond is shown in Figure (1). Pulse echo mode of measurement was used to inspect this joint. By impinging the surface with an ultrasonic wave, stress waves are created in the structure which travels through the thickness of the structure. A portion of these waves gets reflected at the interface and the reflected signal returns as the echo for the initial pulse (shown as A in Fig. (2)). The other portion gets transmitted through the bond-line, and gets reflected at the end of the shear web. This is obtained as the second signal in time domain (shown as B in Fig. (2)). By measuring the difference in the time of arrival between the two signals we can obtain the time of flight (TOF) in the adhesive region. Knowing the speed of sound in adhesive, we can Shear web (foam core) Adhesive bond-line Spar cap FIGURE 1. Schematic of shear web/spar cap joint bond-line. 1285
Adhesive bond-line Shear web (foam core) Spar cap B A T/R To pulse-receiver FIGURE 2. Schematic of pulse echo method used for inspection. determine the thickness using TOF. This method has been consistently used for bond-line measurement, but using it for thick composite plates, (spar cap sections of more than 30 mm) is the challenge. Figure (3) shows the time domain signal form the adhesive and spar cap interfaces as explained in Figure (2). The peak to peak difference in time was measured to give the difference in TOF. The velocity of sound in adhesive was measured by using accurately machined cubes of adhesive and repeating pulse echo measurements to obtain the TOF. Time in microseconds Adhesive signal Amplitude in Volts Spar cap (composite) signal FIGURE 3. The adhesive and spar cap signal in time domain. 1286
FIGURE 4. Sample set used for experimentation. There are 10 samples per set with a total of 60 samples. SAMPLE PREPARATION & EXPERIMENTAL SETUP A set of standardized samples were created to facilitate experiments. In a typical wind turbine blade, the laminate thickness varies from 1.4 mm to 45.5 mm along the length of the blade. Since only the outside of the blade is accessible for measurement, the ultrasonic signal has to travel through the entire thickness of the spar cap twice. Ideally the adhesive thickness varies from 5 mm to 10 mm. Based on the thickness of adhesive and thickness of spar cap, 60 samples were fabricated to create a sufficient data set. The thickness of adhesive was controlled by placing the machined cubes (also called spacers) on the corners of laminate. Hence it could be assumed that the adhesive would be as thick as the spacers. But this thickness cannot be guaranteed as the laminate will not be perfectly flat and when the pressure is applied on the laminates for a good joint, there can be variations in the adhesive thickness. This variation is assumed in the samples and measurements made. The sample set is shown in Figure (4). An Olympus EPOCH XT was used as a pulse/receiver and display unit for measurement of bond-line. The pulse-receiver was operated at 100 KHz PRF, and 40DB gain. Individual A-Scans could be stored and retrieved, but the online-display of EPOCH XT can also display the thickness directly with the help of measuring GATES, which made the measurements faster. Together the system was portable and easy to use. 500 KHz and 2.25 MHz contact transducers were used for all experimentation. RESULTS Since there are two parameters to be considered here, two sets of experiments were done, and two different conclusions are possible. A third parameter was introduced in the form of ease of measurement, i.e. the ability to differentiate peaks and measure the time of flight difference easily. Although this parameter cannot be measured physically, it depends on the person measuring and the time it takes to obtain data. One has to consider all these parameters to able to make a conclusive decision. The results for all the 60 samples are shown in Table (1). It shows that for any given adhesive thickness, if the spar cap thickness is 1.4 mm, the measurement error is ±8%. The measurement error is based on assuming the adhesive thickness to be the same as spacer thickness. Similarly, the measurement error for all laminate thickness can be obtained. Table (1) also shows the results obtained using 2.25 MHz contact transducers. As seen in these results, 500 KHz and 2.25 MHz have similar error bars in the 10.5 mm to 45.5 mm. Below 10.5 mm, 2.25 MHz is much more sensitive. 1287
TABLE 1. Measurement errors for laminate thickness, for any give adhesive thickness. 500 KHz 2.25 MHz Laminate Measurement Laminate Measurement thickness in mm Error thickness in mm Error 1.4 ±8% 1.4 ±2.78% 2.8 ±9% 2.8 ±3.63% 4.2 ±8.46% 4.2 ±3.26% 5.6 ±6.21% 5.6 ±2.94% 7 ±5.13% 7 ±4.61% 10.5 ±5.03% 10.5 ±1.04% 17.5 ±5.34% 17.5 ±2.85% 28 ±5.15% 28 ±6% 35 ±8.48% 35 ±4.92% 45.5 ±3.72% 45.5 ±3.44% TABLE 2. Measurement errors for adhesive thickness, for any give laminate thickness. 500 KHz 2.25 MHz Adhesive thickness in mm Measurement Error Adhesive thickness in mm Measurement Error 5 ±6.19% 5 ±5.52% 6 ±7.5% 6 ±3.1% 7 ±7.32% 7 ±3.56% 8 ±6.35% 8 ±4% 9 ±7.44% 9 ±2.48% 10 ±3.05% 10 ±2.63% Table (2) shows the measurement results for any given laminate thickness, i.e. for any given laminate thickness, the measurement error for measuring a 5 mm thick adhesive bond is ±6.19%. Here the adhesive thickness is assumed to be the same as spacer thickness. Since both the thicknesses are high compared to wavelength of the ultrasonic wave, the error bars are similar for both the 500 KHz and 2.25 MHz. The ease of measurement factor is the same for both frequencies. Further to verify if the assumption of adhesive thickness being the same as spacer thickness, T-Ray was used to measure the bond thickness. T-Ray operates at much higher frequencies, i.e. terahertz (10 12 Hertz), and works by the principles of electromagnetic waves and not ultrasound. This high resolution will help us determine the thickness with a higher precision compared to any other NDE technique. The results of T-Ray investigation are shown in Table (3). These results show that an assumed 5 mm thickness can have a variation of ±1.68%. 1288
TABLE 3. Results of standard sample thickness measurement using T-Ray. The error bar obtained from T-Ray measurements was used for correcting the results. Thickness average in mm T-Ray %error Average 4.908375-1.8325 ±1.68% 5.090786 1.815725 5.111453 2.22905 5.157378 3.14755 4.992048-0.15905 5.01501 0.3002 5.097675 1.9535 5.05175 1.035 5.134415 2.6883 TABLE 4. Corrected confidence levels for laminate thickness, for any give adhesive thickness. 500 KHz Corrected results 2.25 MHz Corrected results Laminate Confidence level Laminate Confidence level thickness in mm thickness in mm 1.4 90.35% 1.4 95.535% 2.8 89.36% 2.8 94.68% 4.2 89.85% 4.2 95.05% 5.6 92.10% 5.6 95.75% 7 93.18% 7 93.7% 10.5 93.28% 10.5 97.27% 17.5 92.97% 17.5 95.46% 28 93.16% 28 92.22% 35 89.93% 35 93.4% 45.5 94.59% 45.5 94.87% TABLE 5. Corrected confidence levels for adhesive thickness, for any given laminate thickness. 500 KHz Corrected results 2.25 MHz Corrected results Adhesive Confidence level Adhesive thickness Confidence level thickness in mm in mm 5 92.12% 5 92.76% 6 90.8% 6 95.22% 7 90.99% 7 94.75% 8 91.96% 8 94.26% 9 90.87% 9 95.83% 10 95.26% 10 95.69% Based on the T-Ray investigation all the previous results of Table (1) and Table (2) were corrected and the new confidence levels of the measurement are shown in Table (4). 1289
DISCUSSIONS & CONCLUSIONS From the results we can observe that 2.25 MHz works best for all ranges of laminate thickness and adhesive thickness. But the third factor, i.e. ease of measurement determines which frequency can be used for measurements. Although 2.25 MHz has an average of 95% confidence levels for any thickness of adhesive and laminate, it is very difficult to measure at laminate thickness beyond 10.5 mm due to attenuation and damping of the signal. Similarly, the 500 KHz measurements become difficult at laminate thickness below 10.5 mm since the first echo becomes merged in the front wall. Hence an optimum solution would be to have a combined solution, i.e. instead of using one particular frequency both 500 KHz and 2.25 MHz can be used depending on the thickness of the laminate under inspection. If the laminate is less than 10.5 mm thick, one can use the 2.25 MHz to obtain higher confidence levels in measurement, and easier measurements, while if the thickness is greater than 10.5 mm, one can use 500 KHz. With the help of bond thickness variation along the length of the blade, blade manufactures can optimize the manufacturing process to ensure the correct amount of adhesive is applied between the shells. This method can be extended to a complete scanning system, by using ultrasonic transducers to obtain the adhesive signal in pulse echo mode. The absence of a second signal in time domain denotes the absence of adhesive in the particular region. By performing a C-Scan (spatial location vs. amplitude), we can obtain the spatial locations of the defective regions. In the presence of the second signal, we can use GATES to track the signal and obtain the thickness in the region directly. This system will be field implementable and will be considered for future work. REFERENCES 1. Luca Goglio, Massimo Rossetto, Ultrasonic testing of adhesive bonds of thin metal sheets, NDT&E, Volume 32, Issue 6, September 1999, Pages 323-331. 2. D. Hsu, M. Hughes, Simultaneous ultrasonic velocity and sample thickness measurement and applications in composites, J. Acoust. Soc. Am. Volume 92, Issue 2, pp. 669-675 (1992). 3. Sergey A. Titov, Roman Gr. Maev, Alexey N. Bogachenkov, Pulse-echo NDT of adhesively bonded joints in automotive assemblies, Ultrasonics Volume 48, Issues 6-7, November 2008, Pages 537-546. 4. K Heller, L.J Jacobs, Characterization of adhesive bond properties using Lamb waves, NDT & E International Volume 33, Issue 8, December 2000, Pages 555-563. 5. L Singher, Bond strength measurement by ultrasonic guided waves, Ultrasonics Volume 35, Issue 4, June 1997, Pages 305-315. 1290