Research On Vibration-based Damage Identification Methods Of Concealed Engineering Structures

Civil engineering structures can normally be divided into two sectors:one is called exposing engineering structure which can be directly touched or observed; the other is called concealed engineering structure which can hardly be touched or observed directly. Concealed engineering structure is an important component of large scale civil engineering structure, which possesses a vital position in the whole structure. However, the damage in concealed engineering structure is difficult to be discovered for its concealed position, which lays a concealed security risk to the whole engineering structure. If the concealed engineering structure damage (especially damage in early stage) can be effectively identified timely, the security assessment of whole engineering structure can be accurately conducted, thus sky-high expensive infrastructure can escape from suffering potential huge loss. Therefore, concealed structure damage diagnosis, confirmation of the existence of concealed structure damage, identification of damage position, evaluation of damage degree and damage variation tendency become the most important issues in health monitoring and security assessing system of large scale engineering structure. In this paper, the exploratory research on damage identification of concealed engineering structures based on vibration was carried out, based on the project of "Synchronous recognition Method of Structural damage and load under ambient excitation"(No. Y201223039) for research projects of Zhejiang Provincial Education Department and the project of "Damage identification method of scour bridge and its application based on support vector machine method"(No.13KC05) for Hunan Province Research Center for Safety Control Technology and Equipment of Bridge Engineering, the main contents of the research were the basic theory of structural damage identification of concealed engineering structure, the exploratory research focuses on the application of modal strain energy efficient, wavelet theory, substructure, time-varying autoregressive moving average model, support vector machines and adaptive Kalman filtering for damage diagnosis of concealed engineering structures under ambient excitation. In this dissertation, the main original work includes:(1) The proposed method identified damage of structures using efficient modes which could be continuous or discontinuous. The method was completely different from the traditional damage identification method which generally used low-level and continuous modes. Based on efficient modal strain energy, a damage identification method of concealed structure was established, which theoretically established the evaluation index and identification equation for the concealed structure damage under the condition of efficient modal. In order to analyze the sensibility of structure damages changing upon efficient modal, by analyzing the eigenvalue change rates before and after structure damages emerge, this paper specifically deduced the theoretical equation for modals which possess higher sensibility to the strain energy variation of damaged elements. In order to judge the effectiveness of this identification method, this paper deeply analyzed the diagnosis effect for single damage and multiple-site damages of concealed engineering structure under working conditions with various degrees of noises.(2) A damage identification method of concealed structure was established based on wavelet analysis. To compensate the deficiencies of a single fingerprint to identify damage of structures which couldn’t guarantee the accuracy of identification, the wavelet analysis technology and efficient mode were integrated to identify the concealed structural damage. The positions of tiny structural damages were identified accurately by taking the advantage of the sensibility of element modal strain energy on tiny structure damage, and efficient partial analysis ability of time-frequency domain of wavelet analysis. The equation of damage identification was used to identify the concealed structure damage based on the damage localization. In order to judge the effectiveness of the identification method, numerical algorithm and engineering case were adopted to deeply study identification procedure, identification effect and impact factors of this method.(3) Based on the interface parameters of relative substructure, a damage identification method was established to identify the partial damage of the concealed substructure. In this method, the sensitivity of relative substructural parameters was used to identify the concealed structural damage. In the research process, a series of researches were analyzed in depth including the identification theory of interface parameters, the relations between concealed structure damage and variation of substructure interface parameters, concealed structure damage identification in terms of damage position and damage degree, and the robustness of the identification method.(4)A damage identification method of structure was established based on ARMA and SVM. For solving the problems of non-stationarity of signals and very limited damage samples, the damage identification method of concealed structure based on Autoregressive moving average model (ARMA) and support vector machines (SVM) was proposed. A series of topics were researched in depth including damage identification theory based on combination of ARMA and SVM, identification procedure and identification accuracy for concealed structure damage of large scale engineering structure under ambient excitation, and the robustness of the identification method.(5) Based on AKF (Adaptive Kalman Filter) and substructure, an identification method of structural damage was established to solve the challenging issue of simultaneously identifying concealed structure and exposed structure in service under ambient excitation only using interface parameters of relative substructure. An exploratory study was conducted including the algorithm theory of AKF, Substructural damage diagnosis theory using AKF, method of determining forgetting factor of AKF, the accuracy and robustness of damage diagnosis by using this identification method.