摘要
依据RBF人工神经网络构建原理与腐蚀过程的相似性,以铝合金外场大气腐蚀数据训练并构建了RBF类型的铝合金腐蚀预测人工神经网络模型,并赋予该RBF网络隐节点数据中心是腐蚀敏感区中心的物理意义.该模型以合金成分、环境因素、时间等为网络输入参量,以腐蚀增重为网络输出;由于RBF网络具有局部响应特性,该类腐蚀预测模型尤其适合训练具有区域集中特点的外场腐蚀数据;仿真结果表明该模型具有良好的预测精度.
RBF neural network was built and trained with outfield atmosphere corrosion data of aluminum alloys for corrosion forecast. Chemical composition of alloys, environment factors and time were designed as network input, while corrosion weight gain as network output. Simulated result showed that the network has good forecast accuracy.
出处
《腐蚀科学与防护技术》
CAS
CSCD
北大核心
2009年第3期245-246,共2页
Corrosion Science and Protection Technology
关键词
铝合金
大气腐蚀
预测模型
RBF神经网络
aluminum alloys
atmosphere corrosion
forecast model
RBF neural network
