GM(1,N)模型进行钢材腐蚀速度预测及腐蚀因素敏感性分析  

Prediction of Corrosion Rates and Sensitivity Analysis of Corrosion Factors of Steels Based on GM(1 ,N) Model

作　　者：任竞争[1] 谭世语[1] 周志明[1] 董立春[1] 何思然[1] 

机构地区：[1]重庆大学化学化工学院,重庆400030

出　　处：《材料保护》2010年第8期21-24,共4页Materials Protection

基　　金：重庆市科技攻关计划项目(CSTC;2008AB4119)

摘　　要：特定钢不同环境下的腐蚀速度预测极为复杂。应用灰色系统理论建立钢材腐蚀速度预测及腐蚀因素敏感性分析的GM(1,N)模型(GlayModel),突破了神经网络方法需要多个样本才能预测的局限。通过计算驱动系数来判断腐蚀因素对腐蚀速度的影响,并提出了选取关键因素进行GM(1,N)建模的方法,在保证精度的情况下,减少了计算量。2个实例分析结果证明,该模型能够准确预测钢材的腐蚀速度,并能正确地判断腐蚀因素对腐蚀速度影响的程度和极性,为钢材腐蚀速度的预测及腐蚀因素的敏感性分析提供了新方法。Grey system theory was adopted to establish GM( 1 ,N) model for the prediction of corro-sion rates and sensitivity analysis of corrosion factors of steels,with which the reliance of neural network route on a large number of samples was eliminated.The grey coefficient was calculated to judge the effect of corrosion factors on corrosion rates,and the method of selecting key factors for modeling was proposed,mading it feasible to reduce the computational complexity without damage to precision.Two examples indicated that the established model could be well used to accurately predict the corrosion rate of steels and correctly judge the effect of corrosion factors on corrosion rates in relation to extent and polarity.

关 键 词：腐蚀速度预测 敏感性分析 钢材 GM(1 N)模型 准确性 

分 类 号：TG172.5[金属学及工艺—金属表面处理]