摘要
为有效缩短陶瓷材料设计中繁琐的试验过程,利用误差反向传播神经网络(BPNN)能够正确逼近非线性映射关系的优点,将其运用到多相复合结构陶瓷材料断裂韧性预测中,克服了陶瓷材料研究中单因素实验法不能正确反映断裂韧性与添加组分多因素之间复杂的非线性映射关系的弱点.对A l2O3/S iC/(W,Ti)C复合陶瓷材料断裂韧性的预测和试验验证表明,该方法可行有效,为快捷、经济地开发研制陶瓷材料提供了新的思路和有效手段.
In order to shorten the experimental procedure of ceramic materials design effectively, a fracture toughness predicting system of advanced ceramic composites based on BP neural network was developed, which can precisely predict the relationship between material composition and the fracture toughness through self-training with the present data, and can perfectly aid the ceramic materials design. This system has friendly interfaces, extensive application, good operating feasibility and reliability examined with the present Al2O3/ SiC/（ W, Ti） C ceramics.
出处
《材料科学与工艺》
EI
CAS
CSCD
北大核心
2005年第5期456-458,共3页
Materials Science and Technology
基金
国家自然科学基金资助项目(50405047)
山东省优秀中青年科学家科研奖励基金资助项目(2000-49)
关键词
BP神经网络
陶瓷材料
断裂韧性
预测
BP neural network
advanced ceramic composites
fracture toughness
prediction
