基于BP神经网络的脉冲电沉积Ni-TiN纳米镀层腐蚀行为预测研究  

Corrosion behavior forecast of pulse electrodeposited Ni-TiN nanocoatings based on BP neural networks

作　　者：许毅[1] XU Yi(Laiwu Vocational and Technical College,Laiwu 271100,China)

机构地区：[1]莱芜职业技术学院,山东莱芜271100

出　　处：《兵器材料科学与工程》2018年第6期49-53,共5页Ordnance Material Science and Engineering

基　　金：国家自然科学基金(51474072)

摘　　要：采用脉冲电沉积方法在A3钢表面制备Ni-TiN纳米镀层,利用扫描电镜、电子分析天平、X射线光电子能谱分析仪等设备对Ni-TiN纳米镀层腐蚀行为进行研究,用BP神经网络模型预测Ni-TiN纳米镀层腐蚀失质量。结果表明:当脉冲频率为500 Hz、TiN粒子的质量浓度为8 g/L和电流密度为4 A/dm2时,Ni-TiN纳米镀层镍晶粒细小、结构精细、致密性较高、耐蚀性较好,镀层中元素Ti和Ni的原子数分数分别为20.1%和51.4%;BP神经网络预测Ni-TiN纳米镀层腐蚀失质量的相对误差较小,最大均方误差(MSE)仅为9.8%。The Ni-TiN nanocrystalline coating was prepared on A3 steel surface by pulse electrodeposition. The corrosion behavior of Ni-TiN nanocrystalline coating was studied by means of scanning electron microscope （SEM） and X-ray photoelectron spectroscopy （XPS）. And BP neural network model was used to predict the weight loss of Ni-TiN nano coating. The results show that, when the pulse frequency is 8 g/L and the current density is 4 A/dm^2, the Ni-TiN nanocrystalline coating with fine grain, fine structure, high density and good corrosion resistance can be prepared. The atomic fraction of Ti and Ni in the coating are 20.1% and 51.4%, respectively. The relative error Of predicting the corrosion weight loss of Ni-TiN nanocrystalline coating by BP neural network is relatively small, and the maximum mean square error （MSE） is only 9.8%.

关 键 词：BP神经网络 Ni-TiN纳米镀层 腐蚀 预测 

分 类 号：TG174.4[金属学及工艺—金属表面处理]