Technology And Method Of Health Monitoring For Bridges

Along with the bridge structures collapsing to fall with every kind of reasons and the large span bridges to set up continuously, the health condition of a bridge structure is becoming people's focus of the concern gradually, and the structural health monitoring (SHM) of a large bridge is one of the research heat points of the bridge engineering field currently and is also a very difficult task. Because of the many courses with complexity of the health monitoring system oneself and a bridge structure with environmental complexity with indetermination etc., the study on this field in the world is still at the exploring phase. This text firstly introduces the concept and basic contents of a bridge health monitoring, points out its advantages compares to the traditional monitoring methods, and the state-of-the-arts of structural health monitoring is comprehensively reviewed in recent years. Then the current advanced fiber optic sensors are comprehensively reviewed and its advantages comparing to the traditional inspective sensors are pointed out. Then through the experiment of applying the fiber optic sensors to monitor the strain changes of the bridge pier with the function of horizontal pull, it is studied that the fiber optic sensors' superiorities comparing to the traditional inspective sensors and its feasibility with creditability etc. of applications with development in a bridge structure health monitoring, and the results indicate that the fiber optic sensors are of many advantages, such as the high gage precision, the absolute gage, the immunity to the interference of electromagnetism, the small effect with temperature, the repeating use and so on. In the last, this text studies the damage identification of the bridge pier structure which is used the neural network, the large quantity of the neural network's training samples are gained from the imitate analysis of ANSYS, the strain data measured from the experiment and partial data calculated from ANSYS is separately to be used as the input vector of the neural network to identify the location and degree of the structural damage. The results indicate that the BP neural network which is composed of the structural deformation degree using as the index of the structural damage identification under the loads is more accurate with simulated samples and measured samples, but it will need more data of the structure's key positions, correspondingly, the BP neural network's work of foundation is more; and that the BP nerve network which is composed of the structural proper frequency as the index of the structural damage identification is basically accurate, but the change of proper frequency is no sensitive to small damage, it only uses to the big damage situation, at the same time, the structural proper frequency is more easy to measure, the homologous training sample is smaller, the workload is small but the result is better, it is viable that the neural network applies to the structural damage identification, and its results are also more accurate and authentic.