基于人工神经网络预测Ni-W合金镀层的硬度和耐腐蚀性能

周琼宇, 谢蔚, 王小芬, 王操, 胡义锋

表面技术 ›› 2016, Vol. 45 ›› Issue (12) : 140-146.

PDF(10567 KB)
PDF(10567 KB)
表面技术 ›› 2016, Vol. 45 ›› Issue (12) : 140-146. DOI: 10.16490/j.cnki.issn.1001-3660.2016.12.023
表面失效及防护

基于人工神经网络预测Ni-W合金镀层的硬度和耐腐蚀性能

  • 周琼宇1, 谢蔚2, 王小芬2, 王操2, 胡义锋2
作者信息 +

Artificial Neural Network-based Prediction of Hardness and Corrosion Resistance of Ni-W Alloy Coating

  • ZHOU Qiong-yu1, Xie Wei2, WANG Xiao-fen2, WANG Cao2, HU Yi-feng2
Author information +
文章历史 +

摘要

目的 预测Ni-W合金镀层的硬度和耐腐蚀性能,优化Ni-W合金镀层的电沉积工艺。方法 在柠檬酸-硫酸盐溶液体系中直接沉积制备Ni-W合金镀层,并将实验所得镀层数据作为学习样品,利用BP神经网络对建立了Ni-W合金电沉积过程参数对镀层硬度和腐蚀电流密度之间的映射关系。结果 低碳钢表面所沉积的Ni-W合金镀层表面均匀致密,与基体结合良好,能够有效地对基体起到保护作用。第二隐层的加入使得3-7-15-2四层网络达到网络收敛的训练次数(1 215 365次)远小于3-7-2三层网络的训练次数(239 950 000次)。四层网络预测所得镀层的硬度和腐蚀电流密度与实验值十分相近,其相对误差≤5.03%。结论 BP神经网络能够准确建立电沉积Ni-W合金镀层的工艺条件和目标性能之间的映射关系,在本文所用的沉积体系和参数范围内,Ni-W合金镀层的显微硬度在296~982HV之间,其硬度最大时所对应的电沉积工艺条件为:pH=7.2,电流密度8 A/dm2,WO42+浓度为0.46 mol/L。Ni-W合金镀层的腐蚀电流密度在7.3~100 μA/cm2范围内。镀层耐蚀性能最好时,即镀层腐蚀电流密度最小时的电沉积工艺条件为:pH=6.4,电流密度0.36 A/dm2,WO42+浓度为0.34 mol/L。

Abstract

The work aims to predicte hardness and corrosion resistance so as to optimize the deposition process of electrodeposited Ni-W alloy coating. Ni–W alloy coating was prepared by direct deposition in aqueous citrate-sulphate solution system. Coating statistics obtained by means of experiment shall be taken as samples to be studied. A neural network was used to establish electro-deposition process parameters of Ni-W alloy, so as to reflect mapping relation between coating hardness and corrosion current density. Ni-W alloy coating deposited on surface of low-carbon steel was uniform and compact, provided with good adhesion to the substrate. Hence it could protect the substrate effectively. With the addition of second hidden layer, 3-7-15-2 four-layer network reaches training times of net work convergence (1 215 365 times), far less than that of 3-7-2 three-layer network (239 950 000 times). The predicted values of hardness and corrosion current density (Jcorr) were close to the values got by experiment, and the relative error was≤ 5.03%. An accurate mapping relation between process conditions of electrodeposited Ni-W alloy coating and target properties can be built by BP neural network. The microhardness of Ni–W alloy coating was within 296~982HV. Electrodeposition process conditions corresponding to maximum hardness are as follows: pH value of 7.2, deposition current density of 8 A/dm2 and WO42+ content of 0.46 mol/L. Corrosion current density of the Ni–W alloy coating is within 7.3~100 μA/cm2. Electrodeposition process conditions correponding to lowest corrosion current density, i.e., best corrosion resistance, are as follows: pH value of 6.4, deposition current density of 0.36 A/dm2 and WO42+ content of 0.34 mol/L.

关键词

Ni-W合金;镀层;人工神经网络;BP网络;硬度;耐蚀性

Key words

Ni-W alloy; coating; artificial neural network; BP network; hardness; corrosion resistance

引用本文

导出引用
周琼宇, 谢蔚, 王小芬, 王操, 胡义锋. 基于人工神经网络预测Ni-W合金镀层的硬度和耐腐蚀性能[J]. 表面技术. 2016, 45(12): 140-146
ZHOU Qiong-yu, Xie Wei, WANG Xiao-fen, WANG Cao, HU Yi-feng. Artificial Neural Network-based Prediction of Hardness and Corrosion Resistance of Ni-W Alloy Coating[J]. Surface Technology. 2016, 45(12): 140-146

基金

国家自然科学基金(51504104);江西省自然科学基金(20151BAB216012,20161BAB206141); 江西理工大学博士启动基金(3401223204)

PDF(10567 KB)

Accesses

Citation

Detail

段落导航
相关文章

/