基于形貌特征的CO2腐蚀速率预测方法

针对CO2腐蚀过程复杂，难以利用实测数据有效预测腐蚀速率问题，文中以腐蚀形貌图像为对象，利用支持向量机（SVM）构建预测模型，实现对CO2腐蚀速率的预测。对N80钢的CO2腐蚀图像进行灰度处理、灰度增强及二值化处理，提取蚀孔数目和孔蚀面积。经计算获得孔蚀密度及孔蚀率，结合工作温度及CO2分压作为腐蚀速率预测的四维特征向量。以SVM构建预测模型，经测试，可准确预测CO2腐蚀速率，并与神经网络预测结果进行对比，验证了该方法的优越性。

Prediction Method of CO2 Corrosion Rate Based on Images

Aiming at solving the difficulty in predicting the corrosion rate for complex CO2 corrosion process with measured data, this paper constructs a prediction model to predict CO2 corrosion rate based on corrosion images. Pit number and pitting cor- rosion area are extracted from the corrosion images processed by gray processing, gray enhancement and binary processing. A four- dimensional eigenvector to predict the corrosion rate is formed in combination with the work temperature and CO2 pressure. The pit density and pitting corrosion arc obtained by calculation. The prediction model constructed by SMV can predict the corrosion rate accurately and the superiority of this method is verified by comparison with the predicted result of the neural network.