摘要
介绍粘结性漏钢的形成过程,指出漏钢预报其实是一种对温度变化波形的模式识别问题。在此基础上设计出基于BP神经元网络的漏钢预报系统,并进行了神经元网络的训练和测试。测试结果表明该系统预报的准确率与预报速度等均优于只采用逻辑判断建立的漏钢预报系统。
This paper introduces sticker-type generating process.It points out that breakout prediction is one kind of pattern recognition of mould temperature changed by breakout.It presents the design of breakout prediction system based on BP neural network.It trains and tests the BP neural network .The result shows that the prediction accuracy rate and prediction speed of the system are superior to use logical judgment to establish logical breakout prediction.
出处
《重庆科技学院学报(自然科学版)》
CAS
2008年第1期66-68,共3页
Journal of Chongqing University of Science and Technology:Natural Sciences Edition
关键词
BP神经元网络
板坯连铸
粘结性漏钢
漏钢预报
BP neural network
slab continuous casting
sticker-type breakout
breakout prediction