Separation Method Of Damage Factor Based On Deflection Measurement Of Prestressed Concrete Bridge

The large span bridge dynamic deflection is the important indicator of analysis and assessment on its function, safety, durability. Obtaining, filtering, extracting and analyzing the dynamic deflection of the large span structure is an interdisciplinary topic,they involve the disciplines of the Civil Engineering Surveying, Signal Processing Method and Scientific Computing.The Digital Photogrammetry System (DPS) is one kind of the high-precision,long-distance, non-contact observation system. Combining the characteristics of the large span prestressed concrete bridge deflection, this paper develops the bridge deflection photogrammetry system by using this technology. It can monitor and record the deflection datas in real-time continuously. This photogrammetry system has been installed on Guangdong Guangming Highway Fuwan bridge and Fujian Pingtan cross-sea bridge, it can monitor the displacement of the midspan on the main span.The whole observation work lasts for 2?5 years. This observation systemcan can obtain many signals of the bridge deformation, through them, it can describe the dynamic change process of the continuous rigid frame bridge.The dynamic deflection that obtained in this paper is the result of the overall performance from multiple factors. The factors that leading to the change of bridge deflection include vehicle load, temperature effect, shrinkage and creep of concrete,loss of prestress and so on. They can be identified as the different compositions of the deflection signals.On the study of the damage diagnosis of bridge, it is first to identify the damage position and the damage extent, afterwards, find out the root of the damage. Now, this paper comes up with the new thought by using the monitoring data, it is seperating the monitoring signals (like the midspan deflection of the bridge) directly,this work can recover each original signal (different compositions), these original signals are unable to be observed directly. Thereby the root of the damage and its proportion can be obtained , it is last to locate the damage.So, this paper tries to look for a way to separate the deflection signals through the methods of signal processing. The technology of EMD (Empirical Mode Decomposition) can adaptively decompose the non-stationary signals. By using it, this paper seperated the instant deflection caused by the vehicle load from the monitoring datas.For the very complilcated long-term deflection signals, after analyzing the deflection of the temperature, concrete shrinkage and creep, loss of prestress, this paper selects the deflection signal of prestress loss as the break through to carry out the test of separation.This paper proposes a new method by using EMD and Independent Component Analysis (ICA), then this paper builds the separation flow chart about the deflection of the prestressed loss. The separation process is mainly divided into two parts. First,determining whether there is loss of prestress on the beam body through the separated deflection signals. And then, it calculates the instant frequency of the beam under prestressed loss, because of the relations of the prestressed and frequency, so it can find out the loss value of prestressed steel through the way of searching for the pre-built prestressed-frequency database.In order to verify the effectiveness of the separation method, this paper designs the deflection signal separation experiment of prestressed loss with the simple support beam. This experiment includes three conditions, the 12 groups deflection signals of the beam were obtained by the DPS. At last, this method can successfully separate the deflection signals of prestressed loss from the mixed signal. Next step it can identify the value of the loss.