WE43镁合金微弧氧化复合工艺及膜层组织与耐蚀性能研究

发布时间：2018-05-28 08:15

本文选题：WE43镁合金 + 激光重熔　；参考：《江苏科技大学》2017年硕士论文

【摘要】：镁合金具有比强度高、密度低、电磁屏蔽性和减震性好等优点,其作为一种轻量化材料,已被各行各业广泛采用,但镁合金较差的耐蚀性能严重限制了其应用。为了提高镁合金的耐蚀性能,对其表面进行处理是最简单、有效的方法之一。为此,本文以WE43稀土镁合金为研究对象,采用激光重熔与微弧氧化复合工艺制备陶瓷膜层,基于ANSYS平台建立激光重熔过程温度场的数值模型,分析了不同工艺参数下温度场的演化规律,研究了该陶瓷膜层的显微组织与耐蚀性能,并基于基因遗传算法优化误差反向传播神经网络建立微弧氧化工艺参数与膜层厚度之间预测模型,主要工作如下:1.利用ANSYS有限元分析软件对激光重熔的温度场进行了数值模拟,利用参数化设计语言仿真了激光重熔的动态过程,分析了其温度场随时间的分布规律。2.采用SEM、XRD对比分析了激光重熔-微弧氧化复合工艺和单一微弧氧化工艺制备的陶瓷膜层的微观组织,结果表明:激光重熔-微弧氧化复合工艺的膜层截面微裂纹和孔洞都要少于单一微弧氧化膜层,同时,激光重熔-微弧氧化复合工艺膜层表面的孔隙率较单一微弧氧化膜层下降了约1/2,膜层的致密性得到提高。3.采用电化学工作站和浸泡实验对比分析了激光重熔-微弧氧化复合工艺和单一微弧氧化工艺制备的陶瓷膜层的耐蚀性能,结果表明:激光重熔后的微弧氧化膜层较单一微弧氧化膜层的自腐蚀电流密度降低了一个数量级;激光重熔后的微弧氧化膜层电化学阻抗谱的容抗弧半径较单一微弧氧化膜层明显增大;激光重熔后的微弧氧化膜层出现明显腐蚀的时间要远远长于单一微弧氧化膜层,这都表明激光重熔-微弧氧化复合工艺在提高镁合金耐蚀性能方面要优于单一微弧氧化工艺。4.根据在不同微弧氧化工艺参数(电流大小、脉冲宽度、氧化时间)下获得的膜层厚度测量数据,采用基因遗传算法(GA)优化误差反向传播(BP)神经网络建立微弧氧化工艺参数与膜层厚度之间的GA-BP预测模型。结果表明:在相同的训练样本和检验样本条件下,GA-BP模型具有更优的函数逼近能力,更好的泛化能力,其中,BP神经网络预测平均误差为8.62%,而GA-BP神经网络预测平均误差仅为1.65%。
[Abstract]:Magnesium alloy has many advantages such as high specific strength, low density, good electromagnetic shielding and shock absorption. As a lightweight material, magnesium alloy has been widely used in various industries, but its poor corrosion resistance seriously limits its application. In order to improve the corrosion resistance of magnesium alloys, surface treatment is one of the most simple and effective methods. In this paper, WE43 rare earth magnesium alloy was used as the research object, the ceramic film was prepared by laser remelting and micro-arc oxidation, and the numerical model of temperature field in laser remelting process was established based on ANSYS platform. The evolution law of temperature field under different technological parameters was analyzed, and the microstructure and corrosion resistance of the ceramic film were studied. The prediction model between the process parameters of micro-arc oxidation and the thickness of the film is established based on genetic algorithm optimization error back-propagation neural network. The main work is as follows: 1. The temperature field of laser remelting is numerically simulated by using ANSYS finite element analysis software. The dynamic process of laser remelting is simulated by parameterized design language, and the distribution of temperature field with time is analyzed. The microstructure of ceramic film prepared by laser remelting and micro-arc oxidation process and single micro-arc oxidation process was analyzed by means of SEM XRD. The results show that the microcracks and holes in the film section of the laser remelting / microarc oxidation composite process are less than that of the single micro-arc oxide film, and at the same time, The porosity of the film surface of laser remelting / micro-arc oxidation composite process is about 1 / 2 lower than that of the single micro-arc oxide film, and the density of the film is improved by .3. The corrosion resistance of ceramic film prepared by laser remelting and micro-arc oxidation and single micro-arc oxidation was analyzed by electrochemical workstation and immersion experiment. The results show that the self-corrosion current density of the micro-arc oxide film after laser remelting is one order of magnitude lower than that of the single micro-arc oxide film. The electrochemical impedance spectrum of the micro-arc oxide film after laser remelting is much larger than that of the single micro-arc oxide film, and the corrosion time of the micro-arc oxide film after laser remelting is much longer than that of the single micro-arc oxide film. All these indicate that the composite process of laser remelting and micro-arc oxidation is superior to single micro-arc oxidation process in improving corrosion resistance of magnesium alloy. The thickness of the film was measured under different parameters (current, pulse width, oxidation time). The genetic algorithm (GA) optimization error back-propagation (BP) neural network was used to establish the GA-BP prediction model between the process parameters of micro-arc oxidation and the thickness of the film. The results show that the GA-BP model has better function approximation ability and better generalization ability under the same training samples and test samples. The average prediction error of BP neural network is 8.62 and that of GA-BP neural network is only 1.65.
【学位授予单位】：江苏科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TG178

【参考文献】