摘要
在长型材生产过程中,影响最终产品质量和力学性能的因素很多,采用计算机对特钢型材生产过程进行模拟,可以分析现场出现的问题,制定出解决办法,这对优化钢材生产工艺有着重要意义。提出了利用初始化学成分和生产工艺参数组成预测特钢型材产品力学性能的BP神经网络模型,通过生产过程中的实际数据的训练,实现了对型材力学性能的离线预测,预测模型适应性好,预测精度高。
There are many factors affect the final production quality and mechanical property in long section production process. The problems in production process for special steel can be discussed and measurement cab he given by computer simulationresults. All of these will have great significance for optimization of production process. In this paper, the BP nerve network models are provided to predict the mechanical properties of long section for special steel with initial chemical composition and technical parameters in production line. The off-line prediction of long section mechanical properties are realized by trainning actual production data. The results shows that the prediction model have good adapbility and high accuracy.
出处
《金属世界》
2006年第5期17-19,共3页
Metal World
关键词
特殊钢
力学性能
模型
预测
BP神经网络
Special steel
mechanical property
model
prediction
BP nerve network