应用人工神经网络预测碳钢低合金钢的大气腐蚀

应用人工神经网络从已有的大气腐蚀数据中求得气象因素、冶金因素与碳钢、低合金钢的大气腐蚀之间的关系,从而预测大气腐蚀,并探讨具体条件下部分因素对大气腐蚀的影响。结果表明:大气腐蚀与气象因素、冶金因素有明显的联系,可以用人工神经网络来预测大气腐蚀,其准确度取决于训练样本的代表性和准确性。

APPLICATION OF NEURAL NETWORKS TO ATMOSPHERIC CORROSION OF CARBON STEEL AND LOW ALLOY STEELS

Neural network (BP algorithm) was applied to atmospheric corrosion of carbon steel and low alloy steels. After learning, the neural network could reflect relationships between meteoric factors, metallurgical factors and atmospheric corrosion performance of the steels. As demonstration, the effects of some factors were examined. The neural network was also used for prediction of atmospheric corrosion of several other steel. The accuracy of the prediction was dependent on accuracy and typicality of training samples.