基于灰色理论和神经网络的拉坯阻力状态预测模型研究  

Research Mould Friction Status Prediction Model Based on Grey Theory and Neural Network Theory

作　　者：赵辉[1] 任秀峰[1] 王红君[1] 岳有军[1] 

机构地区：[1]天津理工大学天津市复杂控制理论与应用重点实验室,天津300384

出　　处：《计算机测量与控制》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)

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.

关 键 词：拉坯阻力 灰色理论 神经网络 漏钢预报 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]