Derivative prediction method for corrosion of metallic materials

The corrosion prediction of metallic materials is very important in engineering, but there always are big errors of predicting the long-term corrosion performance from its short-term test data. A derivative prediction method for corrosion of metallic materials is presented. In order to enhance the inherent regularity contained in the corrosion data, the original corrosion test data are dealt by the method of weighted accumulative addition, and the three-parameter exponent model is applied to describe the regularity of mapped data. At the mean while, the derivative least square method is used to determine the unknown parameters in the model. The method has the advantage that it can give a well fit to both the corrosion curve and its derivatives, so the corrosion curve and its development trend can be described accurately. Therefore the long-term corrosion performance can be predicted with high precision. A lot of engineering applications show that the derivative prediction method has a higher precision than the traditional methods in corrosion prediction. Two examples are also given.