基于SOM-SVM组合分类器的涂层防护性能研究

Protective Performance of Coating Based on Self-organizing Neural Network and Support Vector Machine

目的为避免EIS,EN技术可能出现的问题,建立一个准确、高效的评价模型,以探究现役军用有机涂层防护性能。方法利用电化学阻抗谱(EIS)、电化学噪声(EN)技术分析了两种军车有机涂层在循环暴露试验中的腐蚀行为,提取低频阻抗模值|Z|_(0.1 Hz)与涂层噪声电阻R_n两种电化学评价参数作为自组织神经网络(SOM)的输入训练样本,同时结合支持向量机(SVM)方法建立涂层防护性能组合分类器。结果将涂层失效过程自适应地分为涂层防护性能良好、防护性能下降、基本失效三个阶段。结论所建立的SOM-SVM组合分类器对于辅助分析涂层防护性能具有可行性。

Objective To avoid possible problems of EIS and EN, and establish an accurate and efficient evaluation model to evaluate the performance of active military organic coatings. Methods Through the analysis on corrosion behaviors of two organic coating of military vehicle in the cyclic exposure test, the impedance in low frequency region ｜ Z｜0.1 Hz and the noise resistance Rn, were extracted with EIS and EN. These two electrochemical evaluation parameters were extracted as input training samples of self-organizing neural network（SOM）. At the same time, combined with support vector machine（SVM） method, the coating protection performance classifier was established. Results The failure processes of coating were divided into three stages spontaneously： protective properties being good, being reduced and failure. Conclusion The SOM-SVM combined classifier is feasible for assistant analysis on protective performance of coating