Bridge Safety Evaluation Based On Neural Network

Long-span bridges are the hinges of traffic system and play an important role in national economic and society. Recently, bridge accidents occur from time to time, which bring more and more attention to safety of existing bridges all over the world. Now the safety evaluation of bridge is becoming the hot problem in the research of bridge engineering.The nerve network has the accurate mapping ability to the non-linear problem and it may carry on the free precision to any continual nonlinear function approaching. This characteristic of the nerve network is widely applied to each domain. The high performances of the RBF network are applied in many domains,especially simple structure, the fast training process and the good promoted ability. This article based on the RBF nerve network theory has carry on the appraisal to the old bridges and the appraisal result is accurate and effective.This article has mainly done following several aspects work:1, The nerve network model of reinforced concrete bridge load-bearing rate in bending has been established by the crack width,deflection, the concrete density, the steel bar yield strength, the steel bar diameter,protector thickness as the input level vector, reinforced concrete bridge loading-bearing rate as the output level vector.2, By the results comparation of appraise with the RBF nerve network and the BP nerve network, the BP network is obviously inferior to the RBF network, moreover the BP network training time is bigger than the RBF network.Its training speed is quite slow. The RBF nerve network is one kind partial approaches network and the question has the only determination solution,moreover, the BP net often meets the partial minimum problem.3, Based on the hierarchy assessment model, a neural network method to assess the damage in bridges is put forward. The subsystem neural network assessment model is created by the scores of assessment factors as the input value according to the BRF network high accuracy simulation characteristic.By the scores of assessment factors as the input value and the score of damage assessment as the output value, a multilevel assessment model for bridges is created.This method may be used to evaluate the final result of the bridge as a whole and get the damage rating according to the given standard.