Research On The Health Monitoring Methods Of Bridge Subjected To Moving Load Based On Hilbert-Huang Transform

The bridge engineering is the public engineering that serves travelling needs of many people in the society. Furthermore, it contributes significantly to the development of economy and society. If one day, suddenly the bridges have to stop work when it will adversely great affect to society, especially dangerous if one suddenly collapsed in the service process. To avoid the risk of pity is an extremely important work that needs to be done regularly bridge health monitoring. Moreover the difference between the design and the fact that the truth cannot be avoided, the need to identify the original actual parameter of the bridge structure is extremely important to help the design and management works later. Faced with these problems, the dissertation deeply studied a small range in structural health monitoring that is to identify the actual structural parameters and identify incipient damage for the bridges. To be able to consider whole bridges (all sections), the author have chosen using a the movement vehicle over the bridge to excite bridge structure, then take the displacement obtained results from the accelerometer analyze and estimate to determine the damaged section and the main typical parameters of the bridge structure.The dissertation present the method of using Hilbert Huang transform and dynamic response of bridge subjected to moving load for bridge health monitoring. In this dissertation, author focuses on the investigation of two main problems. The first is the damage detection of bridge structure. The second is the parameter identification of bridge structure. The problem of damage detection in this dissertation only interests crack detection in bridge beam. Not only simply supported beam is applied for crack detection but also multi span continuous beam and cable-stayed bridge are applied for crack detection. Successful application for cable-stayed bridge has an important mean because cable-stayed bridge is complex bridge structure. With this success, we completely can apply for other bridge structures while we will still use the proposed method without any the improving. The problem of parameter identification in this dissertation only focuses on identifying vibration frequency and damping ratio. Because vibration frequency is one of the most important parameter of bridge dynamic. If we have the abnormal vibration frequency of bridge, almost we know that the bridge is working abnormally. The damping ratio of bridge structure also contains important information about working of bridge material. The abnormal changes in damping ratio can give us information about abnormal material. The abnormal material will lead to inceptive cracks in bridge. In this dissertation identifies vibration frequency and damping ratio of simply supported beam, multi span continuous beam and cable-stayed bridge. To clearly understand proposed method, the dissertation will investigate some problems as following presentation1. The first, the author reviews publishing of structural health monitoring, damage detection and parameter identification of the researchers in the world. Though this reviewing, author can clearly see development trend of this approaches, the disadvantage and the advantage of this method of approaches, the restrictions and overcoming ways of the previous methods.2. After reviewing structural health monitoring, damage detection and parameter identification methods, the dissertation deeply focus on two approaches of Wavelet transform based and Hilbert Huang transform based. These approaches are very strong and convenient to damage detection and parameter identification. The two approaches build on modern techniques of signal processing such as wavelet transform and Hilbert Huang transform.3. Although both wavelet transform based and Hilbert Huang transform based approaches have mostly the same advantages, Hilbert Huang transform based approach is more effective than wavelet transform based approach because Hilbert Huang transform can process nonlinear and non-stationary signal. With this advantage the dissertation selects the Hilbert Huang transform based approach. After selecting Hilbert Huang transform based for damage detection and parameter identification, the dissertation continues deeply analyzing shortcomings of Hilbert Huang transform and overcoming approaches. In here, author analyzes some examples of decomposition of two harmonics, from the example the dissertation pointed out that the signals with close mode frequency will decompose very difficultly. Especially, if signals with ratio of consecutive mode frequency are less1.5then EMD cannot decompose this signal. To resolve this problem the dissertation proposes three approaches, the first they use Band Pass Filter method, the second they use HVD method, and finally they use AMD method. As following presenting, the EMD method also will meet difficult problem when signal contain white noise. To de-noising the dissertation introduces EEMD method for signal processing with white noise. Because HVD method also is limited by ratio of consecutive mode frequency and the AMD method needs to determine whole frequency of the displacement result so in this dissertation author selects Band Pass Filter method.4. To effectively apply the EMD method for dynamic response results, we need also to understand formula of displacement results. So the dissertation used improved transfer matrix method for analyzing of dynamic response of cracked bridge subjected to moving load. The traditional transfer matrix method used four unknowns to solve but improved transfer matrix only uses two unknowns to solve. To have formula displacement of complex structures, the dissertation has to analyze these complex structures. In here, the dissertation has analyzed dynamic response of simply supported beam with crack, multi span continuous beam with crack and cable-stayed bridge with crack. These formula results help improving Hilbert Huang transform method later.5. From above displacement results, the dissertation establishes an algorithm of damage detection and parameter identification for simply supported beam, multi span continuous beam and cable-stayed bridge. The algorithm is based on Hilbert transform. The proposed method in this dissertation is combination of the BPF method and the HHT method. In here, the BPF method will play role important to extract the desired vibration components from the displacement result.6. According to the proposed algorithm we can detect damage by IF1curve plot and identify vibration frequency and damping ratio by other plot. The techniques using Hilbert-Huang transform for structural health monitoring under the effect of moving load was presented here. The results of theoretical with numerical simulation show the method can be detected location of damage section relatively accurate. Numerical simulation research the cases of damage in different locations, the different depth of cracks and the load speed of different velocity. The examples of simply supported beam, multi-span continuous beam and cable-stayed bridge have demonstrated proposed damage detection and parameter identification method that can use for actual structures.