Damage detection in concrete gravity dams using Hilbert-Huang Method, as one of the most common signal processing techniques, is studied in this research. After considering a typical geometry for dams, damage is modelled by a reduction in the modulus of elasticity in the dam’s body (in three levels of damage) and in different areas of the structure. The dam is excited by a horizontal earthquake and the primary natural frequencies of the dam are calculated by applying Hilbert-Huang Method to the measured response, which is the acceleration of five points in the system. Based on the changes in the frequencies of the damaged and undamaged structure, a parameter, called relative frequency error, has been introduced. The results show that the proposed criterion used in this study can not only properly identify the location of damage but also predict the severity of the structural damage in concrete gravity dams accurately.

In structural engineering, damage detection in concrete gravity dams (CGDS) is a practical problem. Dam destruction can have severe financial consequences and may even lead to fatalities. Therefore, structural health monitoring in advance is crucial. In this regard, a well-known CGD, namely the Pine Flat Dam, has been chosen for the Finite Element Modeling. In this paper, damage is induced in the dam neck through elasticity modulus reduction by 40 % and 80 %. In addition, after applying Northridge earthquake, the acceleration in structure nodes for intact and damaged cases are recorded in vector formats. Using various methods, such as Discrete-time Fourier Transform (DTFT), Wavelet transform and Wiener transform, the differences between these two signals are investigated. The standard deviation (S.D.) of variations is chosen as the performance metric and is applied to the signal amplitude between intact and damage observations/signals. The reason why several signal processing algorithms are used is finding an approach which shows more clearly the differences caused by the destruction. This is evaluated via S.D. values for different algorithms. The results confirm the superiority of DTFT over other given algorithms. DTFT has a negligible outperformance (approximately zero dB) with respect to the Wavelet transform in both the crest and the lower nodes of the dam. This rate for DTFT and Wavelet is 10dB higher than that of Wiener and 35 dB in comparison with the simple amplitude difference. Moreover, the detection thresholds for the given methods are compared, and it is verified that the DTFT and Wavelet indicate the best performance.

In this article, to detect dam destruction, statistical analysis of recorded observations was used. The efficiency of the introduced method is estimated by numerical simulations. The acceleration measurements in different nodes are obtained in both damaged and intact cases after implementing of three earthquake records with low, middle, and high frequencies, respectively. The statistical segmentation of measurements is performed based on a histogram framework. Consequently, a new automatic method is devised based on the zero-order and first-order cumulative momentums to determine the presence of structure destruction. The proposed histogram-based dam detector has been analyzed and simulated in the MATLAB simulator. To do so, the Peak-to-Sidelobe Ratio (PSLR) of the histogram curves is considered as a performance assessment criterion. According to the simulation results, the proposed method has an efficient performance to detect destruction and determine its position in the dam.

No abstract

Detection of damage in concrete gravity dams (CGDs) is one of the challenges that need to be overcome since dam failure may lead to irreversible consequences. This research aims to detect structural damage within CGDs by wavelet analysis. From a structural point of view, stiffness is an important factor in the dynamic behaviour of concrete gravity dam systems. Any sudden change in the stiffness leads to alteration in the dynamic response of the structures. The proposed analysis of such a condition will help to investigate the responses before and after the occurrence of any structural damage. The main contributions of this paper are to detect the existence of any damage in the dam structure and determine the damage location along the height of the dam. In order to achieve these purposes, three finite element models of the Pine Flat, Bluestone, and Folsom dams are chosen as case studies. These dams have been modelled for both intact and damaged states, and their geometrical, physical, and mechanical characteristics are defined by SAP2000 software. A series of modal analyses was performed to determine the frequencies and shapes of the structural motions. After reduction of the elasticity modulus by 20% and 50%, the Discrete Wavelet Transform (DWT) was applied to the difference between the intact and damaged observations. Then, the DWT outputs were analysed to get information about the existence of damage as well as its location in the dam structure. Overall, from the obtained results, the main finding of this study states that the location and severity of the structural damages have been efficiently detected according to the significant amplitude variations in DWT diagrams. ABSTRAK: Pengesanan kerosakan pada empangan graviti konkrit (CGDs) adalah salah satu cabaran yang perlu diatasi disebabkan kegagalan empangan yang boleh membawa kepada akibat buruk. Kajian ini bertujuan bagi mengesan kerosakan struktur dalam CGDs menggunakan analisis wavelet. Dari sudut pandang struktur, struktur yang kukuh adalah faktor penting dalam sifat dinamik sistem empangan graviti konkrit. Sebarang perubahan secara tiba-tiba pada struktur bangunan membawa kepada perubahan tindak balas dinamik struktur. Analisis yang dicadangkan terhadap keadaan ini membantu dalam memberi tindak balas sebelum dan selepas jika berlaku sebarang kerosakan struktur. Sumbangan utama kajian ini adalah bagi mengesan jika terdapat sebarang kerosakan pada struktur dalam empangan dan menentukan lokasi kerosakan sepanjang ketinggian empangan. Bagi mencapai matlamat ini, tiga model unsur terhingga daripada empangan Pine Flat, Bluestone dan Folsom telah dipilih sebagai kes kajian. Kesemua empangan ini dimodelkan bagi kedua-dua keadaan iaitu ketika baik dan rosak. Ciri geometri, fizikal dan ciri-ciri mekanikal juga telah ditakrif menggunakan perisian SAP2000. Satu siri model analisis telah dijalankan bagi menentukan frekuensi dan bentuk gerakan struktur. Selepas pengurangan modulus keanjalan sebanyak 20% dan 50%, Transformasi Wavelet Diskret (DWT) telah digunakan bagi mengesan perbezaan antara keadaan baik dan rosak. Kemudian, hasil dari DWT ini dianalisis bagi mendapatkan maklumat mengenai kewujudan kerosakan pada empangan dan juga lokasi kerosakan dalam struktur empangan. Secara keseluruhan, hasil kajian berjaya menentukan lokasi dan tahap kerosakan struktur dengan cekap mengikut variasi amplitud ketara dalam rajah DWT.