Experimental Research On Resistance Of Concrete Structure Corroded By Seawater Based On Artificial Neural Network

In practical concrete structure projects, more and more engineering accidentshappened because of lacking of durability. Consequently, the research on durabilityof concrete structure is one of currently important issues in structure research. For the reinforced concrete structure in offshore environment, a main factorof reinforcing steel corrosion is the invasion of chloride iron, which can causedurable damage of the structures. According to this, in this paper, using adipping-drying circulation (in artificial seawater) method, a simulated experimenthad been done in laboratory on the concrete samples under the condition ofdifferent water cement ratios and different dipping-drying temperature. Thechloride iron contents of concrete samples were measured by means of titrating.Furthermore, the artificial neural network model of seawater corrosion was set upand corresponsive MATLAB program had been compiled to analyze and predictthe chloride iron contents of concrete samples in all kinds of circumstances. In addition, two kinds of product, MEC cement and RMA admixture, whichare mainly used in offshore engineering, were introduced in this paper and theirresistance of chloride penetration were judged and analyzed by means ofcomparison experiment of coloring and titrating.