摘要
介绍了预报粘结性漏钢的基本方法,并对结晶器热电偶测得的大量温度数据进行预处理,再利用小波神经网络技术对经过预处理的检测数据进行训练,优化神经网络系统的结构和参数,识别出具有漏钢征兆的波形,提高了预报系统的精度和快速性;给出了用MATLAB实现的网络训练和测试的仿真结果,同时用VC开发了能识别结晶器内单偶、横向、纵向漏钢征兆温度波形的仿真系统。
The method of forecasting mould breakout is introduced.Then,the large data which measured by thermocouple technique are pretreated.Then these large data can be trained by wavelet neural network method in this paper.The results are employed to optimize the neural network structure and parameters.The wavelet neural network which has been trained can recognize the dynamic wave pattern of mould breakout.The speed and curacy of breakout prediction system are improved.Finally this paper giving the simulation results of training and testing the network using MATLAB.Meanwhile I developed the simulation system which can identify the wave pattern of temperature which be measured by single thermocouple,horizontal thermocouple net,and vertical thermocouple net with Visual C＋＋.
出处
《计算机测量与控制》
CSCD
2008年第3期407-410,426,共5页
Computer Measurement &Control
关键词
小波神经网络
漏钢预报
数据预处理
仿真系统
wavelet neural network
breakout prediction
data processing
simulation system
