서지주요정보
신경망을 이용한 후판 압연 공정의 판재 치수 제어시스템 설계 = Design of a plate dimension control system for hot rolling processes using artifitial neural network
서명 / 저자 신경망을 이용한 후판 압연 공정의 판재 치수 제어시스템 설계 = Design of a plate dimension control system for hot rolling processes using artifitial neural network / 이대엽.
저자명 이대엽 ; Lee, Dae-Yub
발행사항 [대전 : 한국과학기술원, 1999].
Online Access 원문보기 원문인쇄

소장정보

등록번호

8009531

소장위치/청구기호

학술문화관(문화관) 보존서고

MME 99048

휴대폰 전송

도서상태

이용가능

대출가능

반납예정일

초록정보

Hot rolling process holds a crucial position in modern manufacturing industry as the first process that is used to convert material into finished wrought product. Thus the effect of quality improvement in this process is more impactive than that of other processes. The quality of the products in the process has been mainly focused on the uniform thickness of the plate, and a lot of researches have been made for the control of the thickness in the process. In these days, the quality control of the rolling process is extended to dimensional control, that is both the thickness and the width control. The variation of the thickness and the width of the plate are made by the gap control of the horizontal and the vertical rolls respectively. To control both the width and thickness of the plate, well-combined motion of the horizontal and vertical rolls are needed, since the two-dimensional factors affect each other. In this paper, the plate dimensional control system for the hot rolling process which consists of AWC(Automatic Width Control), AGC(Automatic Gap Control) using neural network estimator and controller are proposed. In AWC system, vertical rollers compensate the lateral spread of a slab before it is rolled between two horizontal rolls. For this compensation when a slab enters vertical roller zone, the gap between vertical rollers have to be estimated considering the lateral spread by horizontal rollers. For that, the width estimator using a neural network is built and its performance is evaluated through a series of simulation on the AWC system. To produce the plate with estimated dimension by the vertical rollers, a neural network controller is proposed since the process has the nonlinear characteristics. The neural network controller has the combined structure of ART2 and RBF, which has the self-organizing and adaptive characteristics. The same structure of the controller is applied to the AGC. A series of simulation shows that neural network estimator and controllers have good performance and produce a plate with uniform thickness and width.

서지기타정보

서지기타정보
청구기호 {MME 99048
형태사항 viii, 144 p. : 삽도 ; 26 cm
언어 한국어
일반주기 저자명의 영문표기 : Dae-Yub Lee
지도교수의 한글표기 : 조형석
지도교수의 영문표기 : Hyung-Suk Cho
학위논문 학위논문(석사) - 한국과학기술원 : 기계공학과,
서지주기 참고문헌 : p. 141-144
주제 신경회로망
공정제어
후판압연공정
자동판폭제어
판재치수제어시스템
Artifitial neural network
Process control
Hot rolling process
Automatic width control
Plate dimension control system
QR CODE qr code