Jtam.dvi JOURNAL OF THEORETICAL AND APPLIED MECHANICS 49, 2, pp. 399-417, Warsaw 2011 DAMAGE DETECTION IN BEAMS USING WAVELET TRANSFORM ON HIGHER VIBRATION MODES Magdalena Rucka Gdansk University of Technology, Faculty of Civil and Environmental Engineering, Gdańsk, Poland e-mail: mrucka@pg.gda.pl Technical difficulties prevented so far wider applications of higher mode shapes in damage detection. Yet thesemodes carry on a lot of, so much needed, information on damage inflicted to a structure. However, recent scanning laser-basedvibrationmeasurement techniques allowone to uti- lize these highermodes in damage detection effectively. This paper deals with thewavelet-baseddamagedetection technique ona cantilever beam with damage in the form of a single notch of depth 20%, 10% and 5% of the beam height. The purpose of the study is to present the results of experimental and numerical analyses of damage detection based on higher order modes. The first eight modes are considered and the influ- ence of the mode order on the effectiveness of damage detection by the continuous wavelet transform is analysed in detail. Key words: damage detection, wavelet transform, higher vibration mo- des, scanning laser vibrometer 1. Introduction The ability to damage detection and localization at the earliest possible sta- ge becomes an important issue throughout the aerospace, mechanical or civil engineering communities. The existence of damage in a structure results in changes of global dynamic characteristics. Therefore, relatively simple vibra- tion measurements of a structural response and extraction of information on natural frequencies, damping ormode shapes,make damage detection feasible (Ren andDeRoeck, 2002a,b). Dynamic tests are successfully used to damage identification or even to damage reconstruction, especially combined with ge- netic algorithms (Kokot andZembaty, 2008, 2009) or artificial neural networks (Waszczyszyn and Ziemiański, 2001; Kuźniar and Waszczyszyn, 2002; Rucka andWilde, 2010). 400 M. Rucka A relatively recent area of research in damage detection and localization is based on the wavelet transform applied to mode shapes or static deflection line data. The wavelet transform acts as a differential operator and can be applied effectively even for noisy signals. Damage which cannot be identified directly frommode shapes,may be observed onwavelet transforms since local abnormalities in a signal lead to substantial variations of wavelet coefficients in the neighborhood of damage. The literature on wavelet transforms in da- mage detection is very extensive (see e.g. Wang and Deng, 1996; Hong et al., 2002; Douka et al., 2003; Gentile and Messina, 2003; Chang and Chen, 2003, 2004; Knitter-Piątkowska andGarstecki, 2004; Loutridis et al., 2004;Messina, 2004, 2008; Knitter-Piątkowska et al., 2006; Ziopaja et al., 2006; Rucka and Wilde, 2006a,b; Castro et al., 2007; Trentadue et al., 2007; Pakrashi et al., 2007). The previous studies have shown very good accuracy and effectiveness of the wavelet transform although most of the investigations were performed on numerical data without experimental verification. For a practical appli- cation of the wavelet damage detection techniques, research on experimental data is themost important. The applicability of thewavelet damage detection techniques depends on measurement precision and sampling distances. Hong et al. (2002), Douka et al. (2003), Loutridis et al. (2004) as well as Rucka and Wilde (2006a) performed wavelet analyses on mode shapes obtained by standard modal tests using accelerometers and a modal hammer. Trentadue et al. (2007) used a laser sensor for a dynamic modes measurement. Pakrashi et al. (2007) employed a video camera to measure the beam deflection shape. The experimental research carried out so far revealed that only relatively large defects can be detected by thewavelet transform.Hong et al. (2002), Pakrashi et al. (2007) as well as Messina (2008) considered damage of 50% depth of the beam height. Smaller depth of damage was studied by Rucka and Wilde (2006a) – 35%, Douka et al. (2003) – 30% and Loutridis et al. (2004) – 20% and 30%. Most of the reported research based on the wavelet transform are devoted to the analysis of the first mode shape (Hong et al., 2002; Chang and Chen, 2003;Douka et al., 2003; Loutridis et al., 2004;Rucka andWilde, 2006a) or the first threemodes (Gentile andMessina, 2003; Messina, 2004, 2008; Trentadue et al., 2007;Rucka andWilde, 2010). It is sobecause the lowermodes aremuch easier to acquire than the higher ones. However, the measurements of higher order modes have recently becomemuch simpler by applying a modern scan- ning laser vibrometer (Okafor andDutta, 2000; Pai andYoung, 2001;Waldron et al., 2002). This raises a question if the higher order modes can also be ef- fectively applied in damage detection using thewavelet technique. Okafor and Dutta (2000) analysed three experimental and six numericalmodes of a canti- Damage detection in beams... 401 lever beam by the wavelet transform. They concluded that the first and third translational modes showed the damage location clearly, but the secondmode was inconclusive because the damage fell in the vicinity of zero-crossing for the second mode. Gentile and Messina (2003) indicated that damage may occur in locations having poor sensitivity for certain mode shapes. They concluded that neither the fundamental mode shape nor any other higher single mode can be a priori considered as particularly useful in damage detection. They also stated that all available modes or operational deflection shapes should be analysed. Castro et al. (2006) presented a numerical study of the quality of damage detection versus the order of vibration modes for a rod vibrating axially and formulated the conclusions about the influence of the mode order on damage detection. They stated that highermodes aremore appropriate to damage detection and deduced that the modes of displacement that have a node closer to the defect of stiffness type are the best for its detection in the rod. This paper is devoted to damage detection in a cantilever beam with da- mage depth of 20%, 10%and5%of the beamheight by the continuouswavelet transform (CWT) on mode shapes and operational deflection shapes (ODS). Its purpose is to present the results of experimental and numerical analyses of damage detection based on higher order modes. The first eight modes are considered and the influence of themode order on the effectiveness of damage detection by the CWT is analysed in detail. 2. Continuous wavelet transform in damage detection This section shortly recapitulates the theoretical bases of the continuous wa- velet transform and its application to detection of singularities. For a given one-dimensional signal f(x) (here in the form of a beam mode shape, as it is shown in Fig.1) the continuous wavelet transform can be defined as (e.g. Mallat, 1998) Wf(u,s)= 1√ s +∞ ∫ −∞ f(x)ψ∗ (x−u s ) dx (2.1) where x is the spatial variable, Wf(u,s) is the wavelet coefficient for the wavelet function ψu,s(x) and the real numbers s and u denote the scale and the translation parameter, respectively. In the detection of signal singularities, 402 M. Rucka the vanishing moments play an important role. A wavelet has n vanishing moments if the following equation is satisfied +∞ ∫ −∞ x k ψ(x) dx =0 k =0,1,2, . . . ,n−1 (2.2) Mallat (1998) proved that forwavelets having n vanishingmoments and a fast decay, there exists a smoothing function θ(x) with a fast decay such that ψ(x) = (−1)nd nθ(x) dxn (2.3) Therefore the wavelet with n vanishing moments can be rewritten as the n- th order derivative of the smoothing function θ(x), and the resulting wavelet transform can be expressed as amultiscale differential operator (Mallat, 1998) Wf(u,s)= sn dn dun ( f(x)∗θs(x) ) (u) θs(x)= 1√ s θ (−x s ) (2.4) where f(x) ∗ θs(x) denotes convolution of functions. Equation (2.4) reveals that the wavelet transform is the n-th derivative of the signal f(x) smoothed by the function θs(x) at the scale s. Singularities in a signal f(x) can be detected by finding the abscissa where the maxima of the wavelet transform modulus (WTM) |Wf(u,s)| converge at fine scales (Mallat, 1998). The sketch presented in Fig.1 illustrates the scheme of operation of the wavelet-based damage detection technique. Fig. 1. (a) Geometry of the analysed beam; (b) sketch showing the wavelet-based damage detection technique The selection of an appropriate type of the wavelet function and the cho- ice of the number of its vanishing moments is crucial for the effective use of the wavelet analysis in damage detection. The application of wavelets which Damage detection in beams... 403 create themaximumnumber of wavelet coefficients that are close to zero faci- litate damage identification. In this case, strong non-zero values are observed only in places where damage occurs (Fig.1). For the first mode shape of a cantilever or simply supported beam, the wavelet function with 4 vanishing moments should be used. For structural deflection shapes, which can be ap- proximated by polynomials of higher order than 4, the use of wavelets with higher numbers of vanishingmoments is necessary. In this paper, theGaussian wavelets gaus4, gaus6 and gaus8 (Misiti et al., 2007) having four, six and eight vanishing moments, respectively, have been tested as the best candidates to damagedetectionwith theone-dimensional continuouswavelet transform.The advantages of theGaussian wavelets were discussed byMallat (1998), Gentile andMessina (2003) as well as Rucka andWilde (2006a,b). 3. Numerical analysis of damage detection on an example of a cantilever beam A steel cantilever beam with dimensions b = 0.01m, h = 0.01m, L = 0.4m is considered as the testing model (Fig.1a). The experimentally determined material parameters are: the modulus of elasticity E = 198.25GPa and the mass density ρ = 7718.59kg/m3. The beam has introduced a rectangular notch at the distance xd measured from the fixed end to the notch centre. The notch of length 0.002mwas obtained by a high precision cut. The depth of the notch is 20% of the beam height. Numerical simulations were performed on the first eight mode shapes (Fig.2). The mode shapes of the beam were computed by the finite element methodusing the classical Euler-Bernoulli beamtheory.Thebeamwasdivided into 200 elements of length 2mmanddamagewasmodelled as an elementwith a reduced height. Each mode shape line was normalized to 1 and a piecewise cubic spline data interpolation was used to decrease the sampling distance to 1mm. Then each mode was extended outside the original beam span by a cubic spline extrapolation based on four neighbouring points to reduce the boundary effects (cf. Rucka andWilde, 2006a,b). In the first simulation, the notch has been introduced at the distance xd =0.201m. The CWT was conducted on the first eight beammode shapes by the Gaussian wavelet family. The following wavelet functions were applied in the performed analysis: gaus4 (for the 1stmode shape), gaus6 (for the 2nd, 3rd, 4th mode shapes) and gaus8 (for the 5th, 6th, 7th, 8th mode shapes). The chosen wavelet functions create the maximum number of wavelet coeffi- 404 M. Rucka Fig. 2. Numerical mode shapes (black line), curvatures of mode shapes without damage (dashed line) and curvatures of mode shapes with damage (gray line) for the considered cantilever beam cients that are close to zero, and non-zero values dominate at the position of damage. Figure 3 shows the computed WTM for the 1st to 8th mode shape. The modulus maximum value grows with the increasing scale s = 1-15, and the centre notch position can be very well located at xd = 0.201m for the 1st, 2nd, 4th, 6th and 8th mode shapes. However for the 3rd, 5th and 7th modes, the wavelet transform results do not allow one to identify the damage existence in this position, because the damage lies within the zero curvature part of the mode shape. It should also be noted that for the chosen wavelet (e.g. gaus8), the higher mode (see mode 6th and 8th in Fig.3), the higher value of the wavelet transform modulus (for the 6th mode the maximum va- lue of the WTM is 0.0106, while for the 8th mode is 0.0202). This is a result of the property of vanishing moments (Eq. (2.2)). A wavelet having n vani- shing moments is orthogonal to polynomials up to degree n − 1. Since 6th Damage detection in beams... 405 Fig. 3.Wavelet transformmodulus of the first eight numerical mode shapes for damage at the distance xd =0.201m 406 M. Rucka and 8th modes can be approximated by polynomials of higher order than 6 and 8, respectively, hence both produce some non-zero coefficients which are larger for higher order polynomials. Figure 4 presents the WTM of the 2nd and 8th mode shapes computed using gaus4 and gaus6 wavelets. Application of the wavelet function with a smaller number of vanishing moments causes that some non-zero values are observed beyond the defect position. However, the maximum value of the WTM in the defect place has a larger value for the wavelet with the smaller number of vanishing moments than for the wa- velet with the larger number of vanishing moments. For the 8th mode, the maximumvalue ofWTM is 0.0438 when gaus6 is used and 0.0202 when gaus8 is applied. This is because that the wavelet transform is proportional to the nth derivative of a function (Eq. (2.4)). Since in Eq. (2.4) a signal function is smoothed by a smoothing function, values of theWTMare smaller for higher numbers of vanishingmoments n. Fig. 4.Wavelet transformmodulus of the 2nd and 8th numerical mode shapes for damage at the distance xd =0.201m In the second simulation, the notch location xd changes from 0.009m to 0.393m with the step of 0.002m, giving 193 different damage positions. Fi- gure 5 shows the wavelet analysis on the second mode shape for the beam with the notch situated near both ends of the beam, namely at xd =0.021m, and xd = 0.381m as well as with the notch situated within the beam span (xd =0.101m and xd =0.251m). It is visible that for the established wavelet function (here: gaus6), themaximumvalue of theWTMachieves different va- lues depending on damage localization. Figure 6 shows the value of theWTM (for the scale s = 15) at the position of the introduced notch. In this way, we can observe localizations, where the particular mode shape is not sensitive enough to detect the damage. The 1st mode displays only one such area near the free end, whereas the 8th mode has eight such dead areas. The regions where the quality of damage detection by CWT is poor cover of course with Damage detection in beams... 407 zeros of mode shape curvatures. The maximum value of WTM for each da- mage position provides curves (Fig.6) which agree with the absolute values of mode shapes curvatures (Fig.2). Fig. 5.Wavelet transformmodulus of the 2nd numerical mode shape for damage at the distance (a) xd =0.021m; (b) xd =0.151m; (c) xd =0.251m; (d) xd =0.381m 4. Experimental verification using scanning laser vibrometer The experimental setup is shown in Fig.7. Measurements were performed on the previously calculated beam with the notch situated at the distance xd = 0.201m. Three different depths of the notch were considered: name- ly 20%, 10% and 5% of the beam height. The beam was excited using the electrodynamic vibration systems TIRA TV 50009 which consists of shaker S 503 and amplifier BAA 60. The velocity signals were measured by the Po- lytec Scanning Laser Vibrometer PSV-3D-400-M in 101 point distributed at the distance of 4mm along the beam axis. The load cell PCB W208C01 was applied to register the excitation force. In the first step, resonant frequen- cies were found using the frequency response function (FRF). The periodic 408 M. Rucka Fig. 6. Values of the wavelet transformmodulus line (for the scale s =15) at the position of damage (position of damage xd from 0.009m to 0.393m, every 0.002m) Damage detection in beams... 409 chirp signal excitation was applied in the frequency range from 0 to 10kHz. It was possible to identify eight resonant frequencies within this range with the resolution of 0.78125Hz (for example frequencies for the beam with 20% de- fect were found to be: 47.65625Hz, 285.15625Hz, 782.8125Hz, 1577.34375Hz, 2709.375Hz, 3848.4375Hz, 5380.46875Hz, 6796.09375Hz). Each velocity and forcemeasurements were repeated ten times and then the data were averaged in the frequency domain. The estimator H1 for the FRF was used because for scanning measurements the output noise was bigger than the input noise. Next, a sine excitation at eight particular frequencies was used tomeasure the deflection shapes. Themeasured dynamic shapes for the beamwith the notch of depth 20% are illustrated in Fig.8. For the single input force and small damping, if the structure is excited at a resonance, theODS are similar to the mode shapes (Waldron et al., 2002). Fig. 7. Experimental setup The wavelet transform analysis on the measured ODS for the beam with 20% defect were performed using the Gaussian wavelet family (gaus4 for the 1st and 2ndODS, gaus6 for 3rd to 8thODS). Thewavelet transformmodulus results are illustrated in Fig.9. Detection of damage using the 1st ODS was impossible due to insufficient quality of the signal. In the case of 2nd, 4th, 6th and 8th ODS, the dominant maxima lines, corresponding to the defect po- sitions, increase monotonically, and for larger scales they achieve the largest values. The notch presence can be easily recognized and its position determi- ned by thewavelet analysis varies from 0.2m to 0.201m.Thewavelet analysis 410 M. Rucka Fig. 8. Experimentallymeasured operational deflection shapes (black line) and their curvatures (gray line) for the beamwith defect of 20% depth of the 3rd, 5th and 7th ODS makes it impossible or ambiguous to detect da- mage presence and localization because theseODShave zero curvatures in the vicinity of damage location. Figure 10 shows the wavelet transform modulus of 2nd and 8th ODS performed using gaus6 and gaus8 wavelets, respective- ly. Note, that the application of wavelets with a higher number of vanishing moments provides worse damage identification on experimental data, becau- se the wavelet function with smaller numbers of vanishing moments removes small single fluctuations due tomeasurement noise, and the detection of large variations is dominated. Results for thebeamwith thedefect of depth10%are illustrated inFig.11. Only 6th and8thmode enable defect localization using gaus6wavelet. If gaus8 was applied, the identificationwas impossible.Defect of depth 5%of the beam height was impossible to localize, even on higher modes (Fig.12). Damage detection in beams... 411 Fig. 9.Wavelet transformmodulus of the first eight experimental operational deflection shapes (defect with depth of 20%) 412 M. Rucka Fig. 10.Wavelet transformmodulus of the 2nd and 8th experimental operational deflection shapes (defect with depth of 20%) 5. Conclusions In this paper, the wavelet-based damage detection technique was investigated both experimentally and numerically on an example of the cantilever beam with damage in the form of the notch of depth 20%, 10% and 5% of the beam height. The analysis was performed on the first eight mode shapes. Results of the research on the effectiveness of the wavelet-base damage detection techni- que applied to higher vibration modes lead to the following conclusions: • For the established wavelet function, if the mode is higher, the value of thewavelet transformmodulus is also higher, what indicates that higher modes are more sensitive to the presence of the defect. • For the established mode shape, if the number of vanishingmoments of the wavelet is higher, the WTM contains a large number of zero valu- es, what facilitates damage identification. In this case, strong non-zero values are observed only in places where the damage occurs. On the other hand, the application of the wavelet function with a smaller num- ber of vanishingmoments causes that some non-zero values are observed beyond the defect position. • For the establishedmode shape, themaximumvalue of theWTM in the defectplace is larger for thewaveletwith the smaller numberof vanishing moments than for thewaveletwith larger numbersof vanishingmoments. Experimental investigations demonstrated that: • Damage detection by the CWT on the first ODS was impossible, even though the ODSwas determined using very precise apparatus. • Damage detection by the wavelet analysis was more effective on higher experimentally measured ODS. Damage detection in beams... 413 Fig. 11.Wavelet transformmodulus of the first eight experimental operational deflection shapes (defect with depth of 10%) 414 M. Rucka Fig. 12.Wavelet transformmodulus of the 6th and 8th experimental operational deflection shapes (defect with depth of 5%) • The smallest detectable defectwas found to beof depth 10%of the beam height. Localization of this defect was only possible on higher vibration modes (on the 6th and 8th ODS). • For experimental data, the analysis of higher operational deflection sha- pes by the CWT using wavelets with smaller numbers of vanishingmo- ments appeared to bemore effective. Application ofwavelets with higher number of vanishingmoments providedworse or even impossible damage identification. Each mode shape has one or more regions in which it loses sensitivity to damage detection by the CWT. These regions are in agreement with the zeros of mode shapes curvatures. Highermodes aremore sensitive to damage; however they containmore dead zones, where the quality of damage detection by the CWT is poor. Dead zones for particular mode shapes generally do not cover; therefore damage detection based on the CWT is possible for each damage position, if more than one mode is available. For reliable damage localization, minimum two modes are necessary. The future studies will be dedicated todevelopa solution thatwill allow selectingwhichparticularmodes should be measured (instead of measuring all possible modes) in order to identify damage at an arbitrary position by the CWT technique. References 1. Castro E., Garcia-Hernandez M.T., Gallego A. 2006, Damage detec- tion in rods bymeans of thewavelet analysis of vibration: Influence of themode order, Journal of Sound and Vibration, 296, 1028-1038 Damage detection in beams... 415 2. Castro E., Garcia-Hernandez M.T., Gallego A. 2007, Defect identifi- cation in rods subject to forced vibrations using the spatial wavelet transform, Applied Acoustics, 68, 699-715 3. ChangC.-C.,ChenL.-W., 2003,Vibrationdamagedetection of aTimoshen- ko beam by spatial wavelet based approach,Applied Acoustics, 64, 1217-1240 4. Chang C.-C., Chen L.-W., 2004, Damage detection of a rectangular plate by spatial wavelet based approach,Applied Acoustic, 65, 819-832 5. Douka E., Loutridis S., Trochidis A., 2003, Crack identification in be- ams using wavelet analysis, International Journal of Solids and Structures,40, 3557-3569 6. Gentile A., Messina A., 2003, On the continuous wavelet transforms ap- plied to discrete vibrational data for detecting open cracks in damaged beams, International Journal of Solids and Structures, 40, 295-315 7. Hong J.-C.,KimY.Y., LeeH.C., LeeY.W., 2002,Damagedetection using Lipschitz exponent estimated by the wavelet transform: applications to vibra- tion modes of a beam, International Journal of Solids and Structures, 39, 1803-1816 8. Knitter-Piątkowska A., Garstecki A., 2004, Wavelet transformation in damage identification by dynamic tests, Proceedings in Applied Mathematics and Mechanics, 4, 404-405 9. Knitter-Piątkowska A., Pozorski Z., Garstecki A., 2006, Application of discrete wavelet transformation in damage detection. Part I: Static and dy- namic experiments, Computer Assisted Mechanics and Engineering Sciences (CAMES), 13, 21-38 10. Kokot S., Zembaty Z., 2008, Damage reconstruction of 3D frames using genetic algorithmswith Levenberg-Marquardt local search, Soil Dynamics and Earthquake Engineering, 29, 311-323 11. Kokot S., Zembaty Z., 2009, Vibration based stiffness reconstruction of be- ams and frames by observing their rotations under harmonic excitations – a numerical analysis,Engineering Structures, 31, 1581-1588 12. Kuźniar K., Waszczyszyn Z., 2002, Neural analysis of vibration problems of real flat buildings and data pre-processing, Engineering Structures, 24, 1327-1335 13. Loutridis S.,DoukaE.,TrochidisA., 2004,Crack identification indouble- cracked beams using wavelet analysis, Journal of Sound and Vibration, 277, 1025-1039 14. Mallat S., 1998,A Wavelet Tour of Signal Processing, Academic Press 416 M. Rucka 15. Messina A., 2004, Detecting damage in beams through digital differentiator filters andcontinuouswavelet transforms,Journal of Sound andVibration,272, 385-412 16. MessinaA., 2008,Refinements of damagedetectionmethods based onwavelet analysis of dynamical shapes, International Journal of Solids and Structures, 45, 4068-4097 17. Misiti M., Misiti Y., Oppenheim G., Poggi J.-M., 2007,Wavelet Toolbox 4 User’s Guide, TheMathWorks Inc. 18. Okafor A.C., Dutta A., 2000, Structural damage detection in beams by wavelet transforms, Smart Materials and Structures, 9, 906-917 19. Pai P.F., Young L.G., 2001, Damage detection of beams using operational deflection shapes, International Journal of Solid and Structures,38, 3161-3192 20. Pakrashi V., Basu B., O’Connor A., 2007, Structural damage detection and calibration using wavelet-kurtosis technique, Engineering Structures, 29, 2097-2108 21. Ren W.-X., De Roeck G., 2002a, Structural damage identification using modaldata. I: Simulationverification,Journal of StructuralEngineeringASCE, 128, 87-95 22. Ren W.-X., De Roeck G., 2002b, Structural damage identification using modal data. II: Test verification,Journal of Structural EngineeringASCE,128, 96-104 23. Rucka M., Wilde K., 2006a, Application of continuous wavelet transform in vibration based damage detection method for beams and plates, Journal of Sound and Vibration, 297, 536-550 24. Rucka M., Wilde K., 2006b, Crack identification using wavelets on experi- mental static deflection profiles,Engineering Structures, 28, 279-288 25. Rucka M., Wilde K., 2010, Neuro-wavelet damage detection technique in beam, plate and shell structures with experimental validation, Journal of The- oretical and Applied Mechanics, 48, 579-604 26. Trentadue B., Messina A., Giannoccaro N.I., 2007, Detecting damage through the processing of dynamic shapesmeasured by a PSD-triangular laser sensor, International Journal of Solids and Structures, 44, 5554-5575 27. Waldron K., Ghoshal A., Schulz M.J., Sundaresan M.J., Ferguson F., Pai P.F., Chung J.H., 2002, Damage detection using finite element and laser operational deflection shapes, Finite Elements in Analysis and Design, 38, 193-226 28. Wang Q., Deng X., 1996, Damage detection with spatial wavelets, Interna- tional Journal of Solids and Structures, 36, 3443-3468 Damage detection in beams... 417 29. WaszczyszynZ.,ZiemiańskiL., 2001,Neuralnetworks inmechanicsof struc- tures andmaterials – new results andprospects of applications,Computers and Structures, 79, 2261-2276 30. Ziopaja K., Pozorski Z., Garstecki A., 2006, Application of discrete wa- velet transformation in damage detection. Part II: Heat transfer experiments, Computer Assisted Mechanics and Engineering Sciences (CAMES), 13, 39-51 Wykrywanie uszkodzeń w konstrukcjach belkowych za pomocą transformaty falkowej na bazie wyższych postaci drgań Streszczenie Niniejsza praca poświęcona jest technice diagnostyki konstrukcji bazującej na transformacie falkowej. Badany obiekt to belka wspornikowa z uszkodzeniami w fir- mie nacięcia o głębokości 20%, 10% oraz 5%wysokości belki. Pomiary postaci drgań wykonano za pomocą nowoczesnegowibrometru laserowego.Celem pracy jest przed- stawienie eksperymentalnych i numerycznych analiz wpływuwyższych postaci drgań na efektywność wykrywania uszkodzeńmetodą ciągłej transformaty falkowej. Manuscript received September 13, 2010; accepted for print October 28, 2010