摘要
通过对Q195钢丝在不同温度、时间下的退火处理,测试了退火前后的抗拉强度。采用BP神经网络建立了Q195钢丝连续退火后抗拉强度与初始抗拉强度、钢丝直径、保温时间和退火温度之间的预测模型,对钢丝连续退火后的抗拉强度进行预测。结果表明:BP网络预测最大相对误差为3.49%。该预测模型对于Q195钢丝连续退火抗拉强度的预测是有效的、可行的。
Q195 steel wire was annealed at different temperatures and times, and the tensile strength was tested before and after annealing. A predicting model based on BP neural network was constructed. The mapping relationship between tensile strength of Q195 steel wire continuous annealed and initial tensile strength, diameter, holding time, annealing temperature was established. The tensile strength of continuous annealed steel wire is predicted. The results show that the neurl network training error is less than 3.49%. The model is valid, and feasible to predict the tensile strength of Q195steel wire after continuous annealing.
出处
《热加工工艺》
CSCD
北大核心
2012年第12期163-165,共3页
Hot Working Technology
关键词
钢丝连续退火
BP神经网络
预测
wire continuous annealing
BP neural network
prediction