Investigation On The Corrosion Behavior And Corrosion Prediction Model Of Engineering Steels Used In Marine Environment

Marine corrosion is the quesion which ocean exploitation must face and resolve. It is meaningful for design, constructing, and operation of marine engineering to research corrosion behavior of common-used steels in seawater and predict their corrosion rates. In this thesis, corrosion behavior of 3C, A3, 16Mn, 10MnPNbRe, 10CrMoAl and D36 steel in different seawater environment was studied, and it was analyzed for relativity between corrosion rates and seawater environmental factors, for data difference and its reason of different ways to measure corrosion rates and for feasibility of corrosion rate prediction. On the basis of above analysis, complex data processing was used to establish corrosion rate prediction model of marine engineering steels, and the prediction was achieved by computer software.It was found that crystal forms of 10CrMoAl corrosion products were mainly made up of FeO(OH), and little of Fe3O4, which caused its corrosion rate distinctly greater than that of other steels. Due to limited experimental conditions, data and corrosion rules by indoor experiments were different from those by locale experiments. Within alloy elements in steels, effect of Si, C, Mn, Cr, Co, and Al on corrosion rate was relatively marked, and effect of Cu, W, V, Mo and Nb was relatively little. This conclusion was agreed with the result that corrosion rate of 10CrMoAl distinctly greater than that of other steels.After the group of seawater environmental factors in correlative data of "seawater environmental factors — steel corrosion rates" was confirmed, more than one hundred thousand of data were measured, contrasted and analyzed, which gave many corrosion rates of 6 types of marine engineering steels correlated with seawater environmental fators. It was found that strong interaction existed between seawater environmental factors, which caused it difficult for the factors to be separated as independent variables or to be converted into else variables, so all environmental fators must be considered at the same time to establish the model. This made it difficult for analytical means to achieve, and complex data processing must be used. In this thesis, marine corrosion rate prediction model of engineering steels was established by grey interrelation analysis, artificial neural networks and data-base, and prediction system was achieved by computer software. Compared with practical results, the error of prediction values was less than 15%. The system could estimate marine environment corrosion, corrosion grade of marine environment, and prediction corrosion rate for long periods, sea areas, different seasons and sea depth.