用神经网络构建板坯缺陷预报系统

Constructing Slab Defect Prediction System with Neural Network

提出了一种新的基于聚类分析、APCA及RBF神经网络的模式识别方法,用于连铸生产线板坯缺陷的诊断和预报,有效地剔除了样本中的异常数据,对样本参数进行了优选。仿真结果证明了算法的优越性。

A new pattern recognizing method based on clustering analysis,APCA neural network and RBF neural network is put forward to diagnose and forecast the slab defect in successive casting product line.Abnormal data is effectively wiped off,parameters are selected.The superiority of this arithmetic is proved by simulate result.