金属腐蚀性能导数预测方法

DERIVATIVE PREDICTION METHOD FOR CORROSION OF METALLIC MATERIALS

工程中金属材料腐蚀性能的预测具有十分重要的意义 ,但是长期腐蚀性能预测的精度问题始终未能得到很好解决。文中提出一种金属腐蚀性能导数预测方法 ,建立新的三参数指数函数预测模型。该方法首先对原始腐蚀试验数据进行加权累加变换 ,使其最大限度呈现出指数函数规律 ,然后采用三参数指数函数预测模型描述加权累加后数据的变化规律 ,并通过导数最小二乘法确定模型中的待定参数。从而能够同时对腐蚀性能曲线及其导数进行拟合 ,很好把握腐蚀性能曲线的变化趋势 ,实现对金属长期腐蚀性能的高精度预测。大量对比试验结果表明 ,与传统方法相比 ,可以大大提高预测精度。

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.