加速腐蚀环境下LD2 CS铝合金腐蚀损伤分析方法

摘要

目的：研究LD2 CS铝合金腐蚀损伤数据的分析方法。方法针对LD2 CS铝合金腐蚀坑深度值，采用最小二乘支持向量机、BP人工神经网络和求和自回归移动平均等3种模型进行回归和预测分析。对比总结各种模型在小样本递增时间序列下的应用特点。结果对于LD2 CS铝合金腐蚀坑深度值预测，最小二乘支持向量机在回归和预测上都有较高的精度。结论最小二乘支持向量机模型最适合LD2 CS铝合金的腐蚀坑深度值预测。%Objective To study analysis methods for corrosion damage data of LD2CS aluminum alloy. Methods Three forecasting models were used to conduct regression and prediction analysis, targeting at the corrosion pit depths of LD2CS aluminum alloy structural material, which were LS-SVM, BP-ANN and ARIMA. Basic forecasting principles and forecast accuracy of the three prediction methods were compared and analyzed. Results The calculation results showed that the pre-diction accuracy and the prediction extensionality of LS-SVM were better than those of the other models for the prediction of corrosion pit depths of LD2CS aluminum alloy. Conclusion LS-SVM was the most suitable model for prediction of corrosion pit depths of LD2CS aluminum alloy.