The Comparative Studies On Seismic Damage Prediction Methods Of Bridge Structures

Bridge structures are the key point of the lifeline engineering, it plays important part in the urgent succor and rebuilding after the tragedy. After the intensive earthquake, bridges suffered serious damage or collapsed, then the traffic breaks off, the succor was blocked, which will bring on more damage and the tragedy will be much more serious. In fact, the number of bridge destroyed in earthquake is much larger than that was destructed by wind or ship bumping. Therefore, the study on seismic damage of bridges and forecasting the damage in the earthquake is needed and it has important actual senses. On the base, this paper studied several means to predict the seismic damage of bridges in these years and put them into the actual bridges forecasting use and then compare with them on the practicability.In this paper, it will choose several bridge seismic predict methods which are statistical analysis method, pushover method and neural net work predicting method to put into use of the seismic damage predicting.Statistical analysis method is a qualitative analysis one on seismic prediction of bridge structures. We select several methods from the statistical analysis ones to evaluate the bridge safety, when the result is different, we can figure out the concrete seismic result with the engineering comprehensive judgement or analysis method.Secondly ,after briefly reviewing the history of seismic design stages, such as static analysis method, response spectrum method, time history analysis and so on, then gives emphasis to static non linear analysis (pushover analysis method),which is applied to seismic damage prediction of bridge structures. We can get the capacity curve with idarc program, judge out the different failure critical state in different seismic intensity, in this way, we will get the seismic prediction feasibility with this method.Third, the paper constitutes an effective and applied method on predicting the seismic damage of bridges. To avoid the BP algorithm's shortcoming of trapping to a local optimum and to take advantage of the genetic algorithm's globe optimal searching, a new kind of hybrid algorithm was formed based on GA and BP in this paper. First, the initialized weights and threshold of BP neural network was optimized with GA, and then train the network with GA-BP. After collecting a great deal of information of girder-style bridges, the hybrid algorithm was introduced into the earthquake damage prediction, then compare with the single BP neural network, the results show that GA-BP algorithm is an effective and accurate method to seismic damage prediction of bridges.Finally, we choose two seismic examples in Tangshan earthquake which are luan he bridge and sheng li bridge and then study the seismic performance with these three analysis methods, at last, we figure out pushover analysis method and GA-BP method can get the much better seismic prediction result of bridge structures, the statistical analysis method needs more improvement and development because of the effects of anthropic factors.