摘要
热轧带钢厚差诊断专家系统是基于传感器型的在线实时专家系统, 数据采集与处理系统是该专家系统的重要组成部分。精轧机组信号由传感器采集后, 需进行预处理。数据预处理包括三次预处理, 一次预处理包括速度折算、温度趋势项消除, 使影响厚差的各因素变量周期和幅值不经常改变;二次预处理包括低通滤波和高通滤波; 三次预处理是对低通滤波和高通滤波后的信号分别进行频谱分析, 得到特征参数并进行模糊处理。处理后的数据便可以用于厚差诊断专家系统的推理, 以提高厚差专家系统的命中率。
Thickness Deviation Diagnosis Expert System (TDD ES) is the real time and on line system based on sensors, the data acquisition and processing are the main portions of TDD ES. Finisher signal needs to be pretreated after being collected by the sensor. The pretreatment include three steps, the first step is to be converted by speed and then by temperature, the second step includes low pass filtering and high pass filtering, the third step is to be analyzed by frequency spectrum for obtaining the characteristic parameters and then fuzzy processing them. The data after pretreatment can be used for the inference of TDD ES so that the target rates can be improved.
出处
《中国有色金属学报》
EI
CAS
CSCD
北大核心
1999年第4期873-877,共5页
The Chinese Journal of Nonferrous Metals
基金
教育部高等学校博士学科点专项科研基金!资助项目97014515
辽宁省企业技术创新示范工程资助!项目96213051
关键词
热轧
带钢
数据处理
厚差诊断
专家系统
hot rolled strip
data processing
thickness deviation diagnosis
expert system