Vibration based damage identification in highway bridge with generalized curvature quotient difference method
Sara Zalaghi,Armin Aziminejad,Hossein Rahami,Abdolreza S. Moghadam,Mir Hamid Hosseini
Abstract
This paper presents a new damage detection technique for simply supported beams and highway bridges, utilizing the generalized curvature quotient difference method. A convolutional neural network (CNN) system was developed in in conjunction with this method to effectively identify damage locations and intensities in steel girder highway bridges, even amidst noise interference. The proposed damage index is calculated using the stiffness matrix of an intact element, allowing it to detect various damage scenarios in both simply supported beams and the validated finite element model of the I-40 bridge. Simulations were conducted using different bending mode shapes, specifically the first mode for the simply supported beam and the second and third modes for the I-40 bridge model. The results illustrate that the combined approach of the proosed index and the CEEMD noise-canceling method effectively identifies damage locations under both noisy and noise-free conditions. After omitting noise-polluted data, this index was used as input data to train the CNN system. The trained CNN system rechecked the damage locations and achieved precise intensity estimations for multiple unspecified damages (up to four occurring simultaneously), even under noisy conditions. This technique, along with the CNN system, addresses previous research limitations such as the impact of noise (especially near the supports), low speed, low precision, huge input data, and time-consuming training network. The outcomes of this method indicate clarity and accuracy in determining the location and intensity of either multiple or single damage scenarios, even in the presence of noise up to 15%.
Key Words
damage detection; deep convolution neural network; generalized curvature quotient method; noisy condition; steel girder bridge; structural health monitoring
Address
(1) Sara Zalaghi, Armin Aziminejad, Mir Hamid Hosseini — Department of Civil Engineering, SR.C., Islamic Azad University, Tehran, Iran
(2) Hossein Rahami — School of Engineering Science, College of Engineering, University of Tehran, Tehran, Iran
(3) Abdolreza S. Moghadam — International Institute of Earthquake Engineering and Seismology (IIEES), Tehran, Iran.
PDF Viewer
Preview is limited to the first 3 pages. Sign in to access the full PDF.
