The Bridge Structural Health Evaluation Based On Big Data Analysis

Bridge plays an important role in national transportation.With the development of bridge engineering in China,the construction of bridge has been transferred from the large-scale construction to the management and maintenance.Bridge structural health evaluation has received remarkable attention due to the arising structural safety problems.It is the consensus of bridge construction industry to deploy bridge structure health evaluation system on large bridge.In the background of today's big data,the traditional bridge structure health evaluation system has the problem of failing to store data collected by sensors and unable to analyze data effectively.The purpose of this paper is to solve these two problems.The specific work content and achievements are as follows:(1)Combining the characteristics of bridge big data and Hadoop's advantages in processing big data,the architecture of Hadoop-based health evaluation system for bridge,called BSHE-Hadoop,was designed to solve the problem that traditional bridge structure health monitoring systems cannot store large amounts of data effectively.(2)For the damage identification of bridge structural,the deep belief network is used as the damage identification algorithm for bridge structural.Deep learning has already achieved fruitful research results in large data processing in many fields.However,there are few links between deep learning and bridge health evaluation.In this paper,deep learning is introduced to bridge structural health evaluation.(3)In order to solve the deep belief network from being overfitting during the training process,the dropout algorithm was added to the deep belief network.The deep belief network after adding the dropout algorithm compared to the traditional deep belief network has improved the accuracy of the bridge structure state classification by 1.67%% to 63.20%.(4)The deep belief network was parallelized and the MapReduce program of the deep belief network was completed.The deep belief network under the Hadoop platform can complete the damage identification task of the bridge structure and has solved the problem that the bridge structural damage identification cannot be performed under the big data environment effectively.