摘要
采用等离子热喷涂技术对Q215钢结构分别喷涂Cu、Al、Cr,通过中性盐雾腐蚀试验取得钢结构试样的实际腐蚀速率测试数据,将其作为神经网络的训练样本和验证样本,建立了4×15×1拓扑结构的三层神经网络进行钢结构盐雾腐蚀性能的预测,并对网络模型的预测精度进行分析。结果表明,该三层神经网络可实现钢结构盐雾腐蚀性能的高精度预测。
The corrosion rate of steel construction Q215 with the coating of Cu, AI and Cr using plasma thermal spraying were measured by neutral salt spray corrosion test, and the sample data were used for training and certification of the neural network. The 4 ×15 ×1 three layers neural network model was established to predict the salt spray corrosion of steel construction, and the prediction precision of network model was discussed. The results show that 4×15×1 three layers neural network can be used for salt spray corrosion prediction of steel construction with high precision.
出处
《热加工工艺》
CSCD
北大核心
2013年第8期72-74,共3页
Hot Working Technology
关键词
神经网络
钢结构
盐雾腐蚀
预测
neural network
steel construction
salt spray corrosion
prediction
