Research On Durability Assessment Of Existing Reinforced Concrete Bridge Based On Fuzzy Neural Networks

Bridge structure plays a pivotal role in the development of national economy and people’s daily life as an important part of the transport infrastructure. However, a large number of existing reinforced concrete bridges produce concrete spalling, steel corrosion, concrete carbonation, surface cracks and other defects in the process of operation under the influence of external environment and vehicle load. This seriously affects the operational safety and using quality resulting that part of bridge fail prematurely before reaching the design reference period and structural durability problem is very obvious.So, this paper combines engineering practice on the basis of consulting the domestic and foreign research status to formulate the durability evaluation index system and establish the fuzzy neural network system for durability evaluation of existing reinforced concrete bridges. The specific research work is shown as follows:1. This paper systematically analyzes the research background, purpose and meaning of the bridge condition evaluation and expounds the meaning of bridge condition and durability evaluation. Through literature search at home and abroad, this paper clears the research status of durability evaluation to propose the research priorities on this basis.2. Owing to a large number of factors affecting the bridge structures durability,such as concrete carbonation, steel corrosion, concrete freeze-thaw damage, concrete cracks,the paper presents the durability evaluation index system for the reinforced concrete bridge based on AHP following operability, completeness, independence, representativeness, concise and other basic principles.3. On the basis of the reference to domestic and international norms and literatures, this paper determines the evaluation criterion and evaluation levels of the durability evaluation index. This provides the basis of establishing the fuzzy neural network evaluation system.4. This paper elaborates the fuzzy neural network computing theory in order to analyze and research the key technologies constructing durability evaluation system for the reinforced concrete bridge based on the fuzzy neural network. It not only clears the basic processes of the durability assessment but also determines the forms of evaluation index membership function. Even more, this paper builds a fuzzy if-then rule base and determines the fuzzy inference process. These provide a basis for verifying later the fuzzy neural network durability evaluation.5. In order to build the durability evaluation system for existing reinforced concrete bridges based on the fuzzy neural network, this paper selects the assay data and expert opinion survey data as network training and testing samples and normalizes these sample data. The simulation results of the training and testing samples show that the fuzzy neural network system has a good ability to learn and predict. The physical works demonstrate the effectiveness of the method.