Damage Identification Of Prestressed Concrete Girder Bridge Using Artificial Intelligence Method

With the development of science & technology and traffic volume, the prestressed concrete girder bridge has been growing up rapidly with it's unique predominance of handy construction, economical cost, reasonable internal force behavior and comfortable traveling. Hence, a great number of bridges in service with different degree of damages and defects are becoming more and more serious. In order to ensure the safety of the bridges and decrease the total fee of maintenance and repair, detecting these bridge's damages in time becomes an important issue in the field of bridge engineering. According to the fact that main damages in prestressed concrete girder bridges lie in two aspects, i.e. crack of concrete and prestress loss, a cracked beam element model and a prestress loss model are built. In this dissertation, a set of damage identification system about prestressed concrete girder bridge is proposed based on artifical intelligence method. The main research work of this dissertation is as follows:The common crack models are reviewed and a three dimensional cracked beam element with an open, one-edge, transverse crack extending uniformly along the width of the cross section is formulated. Considering a cracked prismatic beam loaded with 6 forces and 6 displacements along the direction of forces, the cracked beam element consists of three segments which are two uncracked beam elements connected by a dimensionless crack element. Two uncracked beam elements are treated as tradional beam elements in the finite element model. The dimensionless crack element is represented by a local flexibility matrix obtained from the integration of stress intensity factors. The element stiffness matrix of cracked beam element is obtained from the concept of the transfer matrix.Three major methods of estimating prestress losses are reviewed, i.e. time-step method, sum of the individual components method and lump-sum method. Based on the assumptions that the general direction of prestress loss value is on the increase, a new mathematical model used for estimating prestress losses is established which includes 5 parameters.BP networks and genetic algorithm, two of the computational intelligence methods to accomplish the mathematically diffcult tasks in the structural condition monitoring are summarized. Aiming at the complexity of damage identification about the prestressed concrete girder bridges, the aptitudinal and predictable method based on combining artificial neural networks and genetic algorithm(GA) in structural damage identification is proposed. The procedure of identifying damage can be defined as a minimization problem. The optimum solution can be obtained effectively by using GA. Since GA usually needs a long analyzing and calculating process in use with the finite element method (FEM). But a non-linear mapping function from multiple input data (structural damage parameters) to multiple output data (differences of response between damaged structure and intact structure calculated by FEM) is constructed within BP networks. The ability of constructing a non-linear mapping function within BP networks offers the strong calculation means which can solve the problem of identification of damages with GA .The method and the approach about identification of damages combined BP networks with GA are given in this dissertation.The basal theories to detect damages about girder structure are also discussed in this dissertation. Those theories include element properties analysis and sensitive factors.Using the method combined BP networks with GA, crack location and depth of a T- shaped girder bridge are successfully detected. Another example consists of a continuous girder bridge, the value of prestress loss about strands is obtained access to two successive phases. The results show the effectiveness of this method used in prestressed concrete girder bridge.