Font Size: a A A

SVM Pattern Recognition Method On Frequency Domain Identification Of Bridge

Posted on:2015-03-25Degree:MasterType:Thesis
Country:ChinaCandidate:S FuFull Text:PDF
GTID:2272330452950174Subject:Bridge and tunnel project
Abstract/Summary:PDF Full Text Request
Due to a wide variety of unforeseen conditions and circumstance, it will never bepossible or practical to build a structure that has a zero percent probability of failureduring the bridge operation. Material aging, environmental conditions, and severeoverloading are examples of circumstances that could affect the reliability and the lifeof a structure. The significance of developing a long-term monitoring system for alarge-scale bridge is that it is really able to provide information for evaluatingstructural integrity, durability and reliability throughout the bridge life cycle andensuring optimal maintenance planning. As the core of bridge health monitoringsystem, the damage identification theory is active area of research in recent years.The basic concepts of bridge health monitoring and damage identification arepresented. The contemporary methods of bridge damage identification aresummarized and reviewed. The major difficulties of damage identification are alsoanalysed briefly. First the support vector machines method is adopted in the trussstructure identification. Then the support vector machines method is adopted in thedamage identification of long-span cable-stayed bridge and the good identificationeffect is obtained. The specific studies are listed as follows:1. The basic characteristics of frequency domain damage identification arebriefly introduced, truss structure as an example to choose a more reasonable damageindex.2. The long-span cable-stayed bridge is taken as the example by optimalplacement of sensors, numerical simulation and analysis have been applied on it, thenthe sample sets are obtained. The support vector classification and support vectorregression are adopted respectively to locate and quantify the damage. The damageidentification result which under the different conditions is compared with each otherand the influence of different noise level is also considered.3.The main works and the research results are summarized,and some improvedtesting methods for the bridge damage identification are put forward at last.
Keywords/Search Tags:pattern recognition, support vector machine, frequency domain damageidentification, long-span cable-stayed bridge
PDF Full Text Request
Related items