Font Size:
VIBRATION-BASED DAMAGE DETECTION OF STRUCTURE’S JOINTS IN PRESENCE OF UNCERTAINTY
Last modified: 2017-11-17
Abstract
A considerable number of different kinds of joints and fasteners are used in any built-up structure such as a road vehicles. Early damage detection of such joints is essential in order to ensure the integrity of structures. The damage can be identified by a vibration based damage detection where a change in dynamic response of the structure is used to identify the damage. Any technique that is used for such a purpose requires dealing with the variability inherent to the system due to manufacturing tolerances, environmental conditions or aging. The level of variability in vibrational response can be very high for mass produced complex structures that possess a large number of components. In this study, a novel time frequency method is proposed for detection of damage in connecting joints. A model of two plate connected by a series of mounts are used to examine the effectiveness of the method where the uncertainty in mount properties are taken into account to model the variability in the built-up structure. The motivation behind the simplified model is to identify the faulty mounts in trim-structure joints of an automotive vehicle where a large number of simple plastic clips are used to connect the trims to the vehicle structure.
The method used here is based on singular spectrum analysis (SSA). It can be used for damage detection and evaluating its severity in structures. A reference space is made from the eigenvectors of the healthy samples. The method involves subjecting the lagged version of the vibration signal to principal component analysis (PCA) and building a new variable space. Any new signal can be compared with the healthy one by projecting it on the reference space. By setting thresholds based on healthy signals the damage can be identified.
A simple build-up structure comprising of two plates connected by eleven mounts is used in this study and the proposed method is used to detect an arbitrarily chosen broken mount. The uncertainties in effective stiffness and damping of the mount were considered to model the variability. The statistical data for the mount properties are adopted from a previous study where the variation in the effective stiffness and damping of trim-structure mounts of an automotive vehicle’s door are evaluated and it was shown that the uncertainty in the mount property have a noticeable effect on the variability of the dynamic response. The uncertainty is mainly due to the boundary condition at mount-trim interface. This allowed assessing the performance of the method in detection of faulty mounts in a realistic scenario.
The plates were considered made of steel with a thickness of 1.5 mm and 1.3 mm and dimensions of .5 m × .35 m and .6 m × .45 m respectively. Eleven mounts, located randomly, connect two plates. FRFs in form of mobility between an excitation point on one of the plates and a response point on the other plate are obtained analytically by considering out of plane deformations. These are used then to simulate time domain responses numerically using inverse Fast Fourier Transform. Twelve set of response are simulated where in the first set all the mounts were connected (healthy samples). In each of the remaining sets, one of the mounts is removed. Each set is comprised of 250 realizations where the mount properties is randomly changed for all the mounts. The first three principle components are used to detect the faulty mounts. A threshold can be set and the method proofed to be effective to detect a broken mount.
The method used here is based on singular spectrum analysis (SSA). It can be used for damage detection and evaluating its severity in structures. A reference space is made from the eigenvectors of the healthy samples. The method involves subjecting the lagged version of the vibration signal to principal component analysis (PCA) and building a new variable space. Any new signal can be compared with the healthy one by projecting it on the reference space. By setting thresholds based on healthy signals the damage can be identified.
A simple build-up structure comprising of two plates connected by eleven mounts is used in this study and the proposed method is used to detect an arbitrarily chosen broken mount. The uncertainties in effective stiffness and damping of the mount were considered to model the variability. The statistical data for the mount properties are adopted from a previous study where the variation in the effective stiffness and damping of trim-structure mounts of an automotive vehicle’s door are evaluated and it was shown that the uncertainty in the mount property have a noticeable effect on the variability of the dynamic response. The uncertainty is mainly due to the boundary condition at mount-trim interface. This allowed assessing the performance of the method in detection of faulty mounts in a realistic scenario.
The plates were considered made of steel with a thickness of 1.5 mm and 1.3 mm and dimensions of .5 m × .35 m and .6 m × .45 m respectively. Eleven mounts, located randomly, connect two plates. FRFs in form of mobility between an excitation point on one of the plates and a response point on the other plate are obtained analytically by considering out of plane deformations. These are used then to simulate time domain responses numerically using inverse Fast Fourier Transform. Twelve set of response are simulated where in the first set all the mounts were connected (healthy samples). In each of the remaining sets, one of the mounts is removed. Each set is comprised of 250 realizations where the mount properties is randomly changed for all the mounts. The first three principle components are used to detect the faulty mounts. A threshold can be set and the method proofed to be effective to detect a broken mount.