You are on page 1of 10

5th International Conference on Structural Health Monitoring of Intelligent Infrastructure (SHMII-5) 2011 11-15 December 2011, Cancún, México

A COMPARISON BETWEEN TWO WAVELET BASED DAMAGE DETECTION OF CRACKS IN A BEAM SUBJECTED TO A MOVING LOAD
A. Khorram1∗, F. Bakhtiari-Nejad2, and M. Rezaian3 PhD candidate, Department of Mechanical Engineering, Amirkabir University of technology, Tehran, Iran 2 Professor, Department of Mechanical Engineering, Amirkabir University of technology, Tehran, Iran 3 Associate Professor, Department of Mechanical Engineering, Amirkabir University of technology, Tehran, Iran
1

Abstract In this study, performances of two wavelet based damage detection approaches to find cracks in a simply supported beam subjected to a moving load are compared. In the first approach, a sensor is assumed to be located at mid span of the beam (fixed sensor at mid span), so in this approach, continuous wavelet transform (CWT) coefficient of time varying deflection of beam at midspan is used . In the second approach, the sensor is assumed to be attached to the moving load (moving sensor), so in this approach, CWT coefficient of moving sensor is analyzed. In both approaches wavelet Gaussian-4 is used. A simply supported beam with a crack is considered. The crack is modeled as a rotational spring which joins the two segments of the beam and its stiffness is obtained from fracture mechanics. Modal expansion theory is used to find the deflection of cracked beam. Keywords: moving load; damage detection; Gaussian wavelet; moving sensor 1. Introduction Arising a crack in structures is a sign of starting failure, so finding crack in a structure at the beginning of its presence is of vital importance in engineering. It has been a long time that many researchers have been trying to find proper approaches for structural health monitoring and damage detection. One of the topics of scientists’ interest is to find the crack effects on vibration response. The idea comes from the fact that each structure can be assumed to include mass and stiffness. Presence of crack induces local flexibility reduction in the structure and so changes the vibration response of the structure due to such damage. In the crack detection of a structure, size and location of crack are of prime importance. These are mostly studied based on free and forced vibration responses of the structure. Vibration caused by moving loads on beams is very common model for many technical components under operating conditions. Metal parts under machining processes, railways and bridges are the common examples of a beam structure subject to a moving
*

Corresponding author’s email: amir_khorram@aut.ac.ir

2. discrete wavelet transform. Sieffert et al (2006).] Some of the noticeable efforts in modeling the response of the cracked and undamaged beams are done by Lee and Ng (1994). Mahmoud and Abu Zeid (2002) and Lin and Chang (2006). wavelet transform has been introduced as a promising method for finding the damage in the structures.The equation of deflective motion for each . Modal expansion theory is used to find the deflection of cracked beam. The location of cracks are determined from the sudden changes in the spatial variation of the transformed response. Vostroukhov & Metrikine (2003).1 has a length . In the second approach. The deflection of the beam is denoted by i=1 for segment is and i=2 for for two intervals of . To estimate the relative depth of the cracks. Parhi and Behera (2001) . as far. it seems to be reported only two papers in damage detection of beam like structures based on the wavelet transform and excitation of moving loads. In their study. However. longitudinal uniform rectangular cross section of height and width with a crack of depth located at is under the excitation of a concentrated moving load with magnitude of and a constant speed of . There are various types of wavelet transform such as Continuous Wavelet Transform (CWT). it is assumed that a sensor is located at mid span of the beam and CWT coefficient of time varying deflection of beam at midspan is used . One of the benefits of wavelet transform is its ability to localize both in time and frequency domains. The advantage of this method was due to preparation of vibration data obtained directly from the moving vehicle. In the first approach.etc. the above two wavelet based damage detection approaches are compared by particulars of a single crack found in a simply supported beam subjected to a moving load. wavelet packet . In other research Zhu and Law (2006) presented another method for crack identification of a beam structure under a moving load based on wavelet analysis. In one research Nguyen and Tran (2010) used a technique for beam damage detection based on the wavelet transform of dynamic response of a bridge loaded by a moving vehicle. cracks could be detected by peaks in the wavelet function ‘‘Symlet’’ for the dynamic response. They analyzed response of beam mid span by using CWT. Theoretical model A simply supported cracked beam as shown in Fig. Garinei & Risitano(2008). Another benefit is its ability to check the smoothness of the curves. it is assumed that the sensor moves with moving load and CWT coefficient of moving sensor is analyzed. In this study wavelet Gaussian-4 is used. In the recent two decades. Cancún. Mahmoud (2001) . In this study.etc that are widely used for damage detection. This property makes the wavelet as a powerful tool for damage detection. México load. It was shown that small cracks with a depth more than 10% of the beam height can be detected. [Cheng et al (2001). It is assumed that the beam is an Euler-Bernoulli in two segments joined by a rotational spring. The crack is modeled as a rotational spring which joins the two sections of the beam and its stiffness is obtained from fracture mechanics. Wang et al (2003).5th International Conference on Structural Health Monitoring of Intelligent Infrastructure (SHMII-5) 2011 11-15 December 2011. spatial wavelet transform. Many of researchers try to improve the usage of wavelet transform in the field of structural health monitoring. they established a damage factor which relates the size of the cracks to the coefficients of the wavelet transform.

conditions of this simply supported beam are: The compatibility equation at the location of crack is: Where. Fig. respectively. which is a function of the crack size and for a single-sided open crack can be defined as Tada et al (1985): Where. México Where and are Young’s modulus.5th International Conference on Structural Health Monitoring of Intelligent Infrastructure (SHMII-5) 2011 11-15 December 2011. Cracked simply supported beam under excitation of a moving load 3. Using the method of separation of variables yields: .1. Eigen-values and mode shapes analysis In order to find the natural frequencies and modal shapes of cracked beam. homogenous solution of equation is found by letting the force term of equation (1) to be zero. is a non-dimensional crack to depth ratio. The boundary inertia of the beam. cross sectional area and moment of is the magnitude of a force which travels along the is denoted as the Dirac delta function. Cancún. is the non-dimensional cracked beam sectional flexibility. beam with a constant speed of V. density.

which are . and multiplying both sides of the above equation by and integrating from 0 to yields: . Cancún. México Substituting equation (6) into equation (5) yields: Where: The solution of equation (7) is: There are 9 parameters in equation 9 that must be found. Forced response analysis Using the modal expansion theory. is N-th eigenfunction of the cracked beam in equation(9) and is N-th generalized coordinate. the forced response y(x. t) of the system can be expressed as: Where. Substituting equation (11) into equation (1).5th International Conference on Structural Health Monitoring of Intelligent Infrastructure (SHMII-5) 2011 11-15 December 2011. These parameters can be found by 4 boundary conditions given in equations (2) and 4 compatibility conditions given by equations (3a) and (3b) and following equation called normalizing equation of mode shape as: 4.

So the response of undamaged beam subject to moving load is: .5th International Conference on Structural Health Monitoring of Intelligent Infrastructure (SHMII-5) 2011 11-15 December 2011. assuming is: and Where: As a special case. Cancún. in the case of undamaged beam under the excitation of moving load and other coefficients are zero. the initial displacement and velocity is zero. México Where: In this study. Which yield .

we try to choose the velocity in such a way that the response of beam become independent of beam parameters. A discussion on choosing the velocity of excitation As can be seen from equations and .( ) yields: . We define as: And also define: Where is an arbitrary velocity coefficient and is called as a non-dimensional time coefficient which varies from 0 to 1 when the load passes the beam.5th International Conference on Structural Health Monitoring of Intelligent Infrastructure (SHMII-5) 2011 11-15 December 2011. Mathematical manipulations. México Where equal to is N-th natural frequency of undamaged simply supported beam which is 5. i. Cancún. there are many parameters which affect the beam deflection. and normalizing the deflection given by equation 17a by static deflection of the beam at mid span due to a constant load at middle of the undamaged simply supported beam. In order to generalize our study.e.

this idea can be applied to the case of cracked beam to generalize the study the response. 2. This figure matches the figure represented in Mahmoud (2001) which shows the accuracy of the model used in this study. This holds in the case of cracked beam with a good approximation.5th International Conference on Structural Health Monitoring of Intelligent Infrastructure (SHMII-5) 2011 11-15 December 2011. Validation of modeling As a validation to this study. México In the above equation. the non-dimensional time and number of mode shapes used to approximate the solution. This is favorably declared that the response is mainly affected by crack size and location which is our favorite goal. place of moving load 7. shows the normalized place of moving load which is equal to Vt/l. The displacement of the beam is normalized by as mentioned above. Comparison of two methods of damage detection based on CWT The continuous wavelet transform (CWT) of a function f(t) by definition is: Where is complex conjugate of wavelet .Normalized deflection of mid span vs. we suppose a simply supported cracked beam with . So the response of undamaged beam is independent of beam parameters. 6. Cancún. must satisfy the following admissibility condition: . the normalized deflection at distance the beam is only a function of the velocity coefficient coefficient from the left support of . Deflections of the beam at mid span versus the position of moving load for two speeds of and are depicted in Fig. a is scale and b is translation. 2. The horizontal axes. Fig. So.

One may verify this fact by comparisons as presented in cases shown in Fig.4.e. Cancún. The above inequality implies that: is the coefficient of wavelet transform. As a sample. 4 shows comparison between two methods when frequency of sampling is 400 . the spike can clearly be seen. In in equation (22) are compared: a) this study. changes in the dynamic response at the point of crack. However. 5 shows comparison between two methods when frequency of sampling is 100. and Fig. The other parameters are the same as used to obtain Fig. In this study. CWT of two functions of Displacement at mid span during the time (we called it as fixed sensor (we at mid span) and b) Displacement of moving load during the time called it as moving sensor) The CWT coefficient of above cases a and b under various excitation velocities of such as were investigated.5. 4. and at scale . the crack can be found using CWT of mid span displacement but in the case of moving sensor. CWT coefficient reported by moving sensor is greater than the other which means CWT coefficient in method of moving sensor is more sensitive and shows the crack size better. México Where is the Fourier transform of . Fig. Gaussian-4 is used which is fourth derivate of and is a constant which can be found in such a way that the norm of wavelet equals to one. and scale at . .this sharp changes can be seen by the wavelet and a large amount of -like a spike. In this case.can be seen in the plot of wavelet coefficients. When sharp changes occur in the - i. Investigations over several cases show that the method based on moving sensor is much more effective than fixed sensor in terms of detectability. Fig.5th International Conference on Structural Health Monitoring of Intelligent Infrastructure (SHMII-5) 2011 11-15 December 2011. As can be seen both methods find the place of crack accurately.

Moving sensor Comparison between CWT coefficients obtained by moving sensor and fixed sensor at mid span and 7. K. 2001. K. .4 Comparison between CWT coefficient obtained by moving sensor and fixed sensor at mid span and a. Lee HP. These approaches are : a) using a sensor which was located at mid span of the beam.Vibrations of railway bridges for high speed trains under moving loads varying in time. Engineering Structures..5th International Conference on Structural Health Monitoring of Intelligent Infrastructure (SHMII-5) 2011 11-15 December 2011. Au F. México Fig. Conclusion In this study. By comparing Gaussian-4 CWT coefficients for responses in both approaches. Dynamic response of a cracked beam subject to a moving load. two wavelet based damage detection approaches to find a crack in a simply supported beam subjected to a moving load were studied and compared. b) using a sensor which was attached to the moving load and moves along the beam (moving sensor). S. References: Cheng Y. 2008... Acta Mechanica. Cheung Y. T. Fixed sensor at mid span b. 1994. it was found that CWT coefficients obtained from moving sensor rather than a fixed sensor is more sensitive to crack depth and more detectable for crack location. Engineering Structures. Ng TY. Risitano G. Cancún.Vibration of railway bridges under a moving train by using bridgetrack-vehicle element. 23: 1597-1606 Garinei A. 106:221–30. 30:724-732.

28: 1543-1554 Tada H.. F. México Lin H. Missouri: Research Corporation. Proceedings of the Institution of Mechanical Engineers: F—Journal of Rail and Rapid Transport. International Journal of Mechanical Sciences. T. 2006. V. Ramondenc Ph. P... L. 40: 5723-5752 Wang J.Q . 2002. Y.. 2006.. Sieffert Y. and Irwin G. Proceedings of the Institution of Mechanical Engineers: C. V. The Stress Analysis of Cracks Handbook. Periodically supported beam on a viscoelastic layer as a model for dynamic analysis of a high-speed railway track. International Journal of Solids and Structures. Multi-cracks detection of a beam-like structure based on the on-vehicle vibration signal and wavelet analysis. Michel G.Identification of crack location in vibration simply supported beams. Jullien J. Cancún. Law S.S.. 2003. Metrikine A. Forced responses of cracked cantilever beams subjected to a concentrated moving load. M. Journal of Sound and Vibration. 215:207–15. Journal of Sound and Vibration.. Engineering Structures.5th International Conference on Structural Health Monitoring of Intelligent Infrastructure (SHMII-5) 2011 11-15 December 2011. International Journal of Solids and Structures. . Dynamic deflection of a cracked beam with moving Mass. 329 : 4455–4465 Parhi D. 2003.. V. Paris P. F. 2006. A. R. Effects of the diaphragm at midspan on static and dynamic behaviour of composite railway bridge: A case study. St Louis. Tran H. 172 (4):549–558. Journal of Mechanical Engineering Science. 1994. Chen B. 1985. A. C. 256(4): 591-603 Mahmoud MA. Dynamic response of a beam with a crack subject to a moving mass... Journal of sound and vibration . Effect of cracks on the dynamic response of a simple beam subject to a moving load. Nguyen K. 2001. Narkis.215:207–15.. 2001. Abou Zaid. second edition Vostroukhov A. Vibration suppression for high-speed railway bridges using tuned mass dampers. Ch. 40: 465-491 Zhu X. Chang Sh. K. International Journal of Solids and Structures.. 2010. Lin C. Behera A.Wavelet-based crack identification of bridge beam from operational deflection time history. 43: 2299–2317 . 48: 1456-1463 Mahmoud M.