摘要
以Cr、Co、Al、Sr、RE的含量和制备方法为输入层节点参数,以腐蚀电位为输出层节点参数,构建了神经网络分析模型,并对预测能力进行试验验证,同时测试了模型选出的最优耐蚀性镁基储氢合金的性能。结果表明,镁基储氢合金耐蚀性神经网络分析模型的预测精度较高,半连续感应熔炼后再机械球磨的分步法制备出的Mg2NiCr0.3Sr0.1RE0.1具有最优耐蚀性,且吸放氢性能与Mg2Ni相当,循环稳定性明显优于Mg2Ni,循环20次后放电容量衰减率从81.3%下降至39.8%。
The neural network analysis model was built by taking Cr content, Co content, Al content, Sr content, RE content and preparation method as input parameters, and corrosion potential as output parameter. Then the checkout test was given out and other properties of the alloy with best corrosion resistance, which is selected by the model. The results show that the neural network analysis model for the corrosion resistance of magnesium-based hydrogen storage alloy has high precision, and Mg2NiCr0.3Sr0.1RE0.1 alloy has optimum corrosion resistance, and the capability of hydrogen absorption and desorption is similar to Mg^Ni alloy, but the cycling stability is markedly superior. Compared with Mg2Ni alloy, the attenuation ratio of the discharge capacity for Mg2NiCr0.3Sr0.1RE0.1 alloy reduces from 81.3% to 39.8%.
出处
《热加工工艺》
CSCD
北大核心
2013年第18期33-35,共3页
Hot Working Technology
基金
黑龙江省教育厅人文社会科学项目(12522325)
黑龙江省教育科学规划课题(JJC1211088)
齐齐哈尔大学教育研究重点资助项目(2012017)
关键词
镁基储氢合金
耐腐蚀性
合金元素
制备方法
神经网络
magnesium-based hydrogen storage alloy
corrosion resistance
alloying element
preparation methods
neural network
