Numerical And Model Test Of Health Diagnosis Of Suspension Birdge

In recent decades, with the advances of science and technology, the quantity and size of large bridge structure are growing rapidly, In the large span bridge, the suspension bridge has become the main bridge. Due to the complexity of the structure and environment, the suspension bridge is liable to a variety of high-risk damage and damage accumulation in the actual operation, such as corrosion, aging, damage, cracks of components and so on. Sometimes the structure is subjected to overloading and sudden disasters. If these damages can not be dealed with in time, security accidents and heavy losses would be resulted in.In this paper, some important issues about the health monitoring and diagnosis of the suspension bridge structure have been study primarily based on the entire bridge model test.For health monitor and diagnosis, a 10-meter-long suspension bridge test model is designed and establised. By means of test model, static and dynamic properties of the suspension bridge model have been analyzed. The damage cases involved abnormality, which are possible to occur, have been simulated. Based on the testing, the variations of static and dynamic responses and sensitivity characteristics of suspension bridges have been studied.Based on the finite model which has been revised by test result, the damage detection indices of suspension bridge have been studied. frequency changes ratio index, modal curvature difference index and static strain difference index for girder damage, hanger damage and cable damage have been analysised. From the simulation results, a practical damage detection methodology of suspension bridges is proposed in this thesis by combination usage of those three indices.According to the characteristic of the three indices, a simple and easy sensors' placement method has been put forward and given a detailed discussion of the disposing of the measuring points when detecting static strain,frequency and modal curvature. Sensor placement for static strain relies on the static strain distribution while maximum kinetic energy criterion is adopted to determine the sensor locations for frequency. When locating sensors of modal curvature, successive addition approach is introduced and MAC matrix is used to evaluate nonlinear of singular vectors.According to that the hangers are distributed along the whole suspension bridge, a new damage detection method of suspension bridge based on the combination of hanger's vibration measurement and neural network is proposed. The outstanding feature of the method is that a good result can be obtained by using only fundamental natural frequencies of hangers; the measurement of fundamental natural frequencies of a few hangers is much easier than some other damage-oriented measurements.The large-scale suspension structure has significant geometry nonlinear characteristics, particularly in the damage state, the response signals show great non-linear non-steady characteristics which traditional signal processing techniques and analysising methods are limited greatly to deal with. the application of HHT method to process the non-steady-state singals is dicussed in this thesis.