【摘要】 利用Kohenen的人工神经网络（artificial neural networks, ANN）技术进行膜层分析，研究了模拟海洋大气环境中镀锌钢表面三价铬基转化（trivalent chromium conversion, TCC）膜的腐蚀失效演变规律.以电化学阻抗谱（electrochemical impedance spectroscopy, EIS）低频膜阻值(|Z|_{0.1 Hz})及低频相位角(θ_{0.85 Hz})两种特征参数作为评价指标，对涂层性能变化过程进行研究，TCC膜的腐蚀过程大致经历5个阶段，并且|Z|_{0.1 Hz}的数据更具有代表性及合理性.将伯德图中全频阻抗变化率k（f）作为ANN的样本输入，5个膜层失效过程对应腐蚀初期、腐蚀前中期、腐蚀中期、腐蚀后中期及腐蚀后期.利用实验检测手段（SEM和EDS）,验证了自组织ANN对TCC膜的各腐蚀阶段分类结果，分别是腐蚀阻隔阶段、膜层微蚀阶段、腐蚀产物沉积阶段、腐蚀拓展阶段、膜层失效阶段.利用ANN分析膜层全频阻抗变化率可实现对涂层性能状态的快速有效判断.更多还原

【Abstract】 Corrosion failure evolution of trivalent chromium conversion（TCC） film on galvanized steel in a simulated marine atmosphere has been studied by utilizing Kohenen’s artificial neural networks（ANN） film analysis. The corrosion process of TCC film has been investigated, using two characteristic parameters of electrochemical impedance spectroscopy（EIS）, low frequency film resistance(|Z|_{0.1 Hz}) and low frequency phase angle(θ_{0.85 Hz}) as evaluation indexes. It has been confirmed that the corrosion process of TCC film roughly undergoes five stages, and the data of |Z|_{0.1 Hz} is more representative and reasonable. The full-frequency impedance change rate in Bode diagram is taken as the input sample of ANN, and the five film failure processes correspond to the initial corrosion stage, the pre-middle corrosion stage, the middle corrosion stage, the post-middle corrosion stage, and the late corrosion stage. The classification results of TCC films by self-organized ANN have been verified by the experimental analysis（SEM and EDS）, which are the corrosion barrier stage, the film micro-etching stage, the corrosion product deposition stage, the corrosion expansion stage, and the film failure stage. It could be concluded that the performance state of the coating can be judged quickly and effectively by utilizing ANN to analyze the full frequency impedance change rate of the film.更多还原