Research Of Concrete Structure Durability In The Yellow River Delta Area Based On Artificial Neural Network

Reinforced concrete structures have been widely, for reinforced concrete structures can make the best use of both concrete and steel bars. For a long time, the deficiency of durability appears in lots of existed reinforced concrete structures. It causes many aging problems and has an ill effect on people's productivity and life. Especially in marine environment, corrosion of steel bars is more serious, influenced by carbonation and chloride erosion.Artificial neural network (ANN) is a mathematical model of neural network of brain, and is a information handling system that simulate architecture and function of brain. ANN can realize complicated logic operation and nonlinear relation, and it is very suitable for such problems that are difficult to be established mathematical model but easy to be collected learning set. Structure durability is a matter of systems engineering that involves many factors, and it is difficult to be expressed with a mathematical formula precisely. Therefore, the author put forward the research of concrete durability in the yellow river delta area based on artificial neural network.By borrowing ideas from the research findings on concrete durability around the world, the author studied the failure mechanism of buildings in the yellow river delta area. Incorporating the principle of Artificial Neural Networks, the paper used MATLAB Neural Networks Tool Box to research how to select input variable, network fabric, transfer function and the other parameters, and improved the Disadvantages of BP algorithm.On the basis, base on the general survey of buildings in the yellow river delta area, the typical buildings, constructed in different years, were tested. The contents include concrete strength, rate of steel corrosion, diameter of rebar, concrete cover thickness, carbonation depth, chloride erosion depth, which become training samples and testing samples.This paper uses individually self-adaption learning ratio algorithm of additional momentum, L-M algorithm to train networks, creates the artificial neural network to predict carbonation depth, chloride erosion depth and rate of steel corrosion and analyses the feasibility by the actual data.At last, author proposes some measures and suggestiongs aiming at repair and the reinforcement of the existing concrete structures and design and construction of building structures in the yellow river delta area.