自组织特征映射人工神经网络对有机涂层电化学阻抗谱的分析  

Analysis of Electrochemical Impedance Spectra of Organic Coatings by SOM Artificial Neural Network

作　　者：丁锐[1,2] 桂泰江[2] 蒋健明[2] 余海斌[1] 

机构地区：[1]中国科学院宁波材料技术与工程研究所,浙江宁波315201 [2]海洋化工研究院有限公司,海洋涂料国家重点实验室,山东青岛266071

出　　处：《现代涂料与涂装》2017年第2期8-11,28,共5页Modern Paint & Finishing

摘　　要：应用SOM神经网络,以不同阻抗参数对阻抗谱进行了自动分析。由于相位角和阻抗模值、相位角和阻抗变化率、相位角变化率和阻抗模值的数据组合,既可以反映阻抗谱形状的变化,又包括了阻抗谱大小的变化,应用SOM人工神经网络,获得了较好的分类结果,其中相位角和阻抗变化率的数据组合效果最好。The self-organizing feature map artificial neural network was established to identify and analyze the electrochemical impedance spectra of organic coatings. The experimental results showed that the three kinds of data combinations, which were phase angle and impedance modulus, phase angle and impedance variation rate, phase angle variation rate and impedance modulus, included both the change in shape and magnitude of the impedance spectrum, and could get a better recognition effect. The data combination of phase angle and impedance variation rate got the best effect. The network technology can provide the coating research institutions with convenient impedance spectrum analysis means for examining the variation of coating state.

关 键 词：SOM神经网络 有机涂层 电化学阻抗谱 

分 类 号：TQ630[化学工程—精细化工]