摘要
由于拉坯阻力在时域上的非线性特征,漏钢现象产生的信息不完全以及连铸生产环境、工艺复杂等问题,利用灰色理论对传统的使用神经网络进行故障预测的模型进行了改进和优化;文章首先论证了建立灰色-神经网络模型预测拉坯阻力状态的实际需要和可行性,而后阐述了利用灰色理论和神经网络解决问题的方法,最后在论证的基础上利用编程仿真证明了模型建立的可行性和可靠性;文章根据实际现场数据以及生产工艺参数,结合生产故障典型特征,得出更加精确有效的故障诊断模型。
Because the problems that mould friction has some nonlinear characteristics in the time domain, breakouts mess^iges are not complete, and the complication of the circumstance and technology in continuous casting , this paper uses Grey theory and Neural Network theory to improve and optimize traditional breakouts prediction model. In the paper, firstly, proved reality needs and feasibility to predict the status of mould friction, then, it explain that how to solve these problems by using those two theories. In the end, based on proof, the mod- el is proved valid by the simulation and programming. Based on actual data and process parameters, combined with classic fault features, we can achieve more accurate and efficient fault diagnosis model.
出处
《计算机测量与控制》
CSCD
北大核心
2012年第3期602-605,共4页
Computer Measurement &Control
基金
天津市自然科学基金项目(09JCZDJC23900
10JCZDJC23100)
天津市科技支撑计划(10ZCECJD43080)
关键词
拉坯阻力
灰色理论
神经网络
漏钢预报
mould friction
grey theory
neural network
breakouts prediction
