涂层失效过程电化学阻抗谱的神经网络分析

ELECTROCHEMICAL IMPEDANCE SPECTROSCOPY ANALYSIS OF COATING DETERIORATION PROCESS WITH KOHONEN NEURAL NETWORKS

在涂层失效过程的研究中,对长期浸泡于电解质溶液中试片的电化学阻抗谱(EIS)进行分析,提出了利用Bode图幅频特性曲线斜率做为评价参数的方法.以幅频特性斜率曲线作为Kohonen神经网络的输入,将涂层失效过程自适应的分为5个连续子过程,并将学习后的网络用于涂层失效过程阻抗谱分析.结果表明不同类型涂层失效过程不同。

The Electrochemical impedance spectroscopy (EIS)of the simulated specimens with different coatings was studied.The slope of log impedance modulus | Z | versus log f was put forward as characteristic parameter k(f) .The deterioration process was divided into five sub- processes by using Kohonen neural networks.The input of the networks is k(f) of different samples for different immersed time.The coating state can be defined by the output of the networks.Therefore the output can be considered as the quality index of the coating to some extent.The result indicates that the degradation process varies with different coatings and preparation techniques.The network proves to be useful to evaluate the coating performance.