최적화 기법과 분산 컴퓨팅을 이용한 재료 성형공정의 역문제 에 관한 연구

다국어 초록 (Multilingual Abstract)

In this paper, an inverse problem of glass forming process is studied to determine a number of unknown<br/>
heat transfer coefficients which are imposed as boundary conditions. An analysis program for transient heat<br/>
conduction of axi-symmetric dimension is developed to simulate the forming and cooling process. The<br/>
analysis is repeated until it attains periodic state, which requires at least 30 cycles of iteration. Measurements<br/>
are made for the temperatures at several available time and positions of glass and moulds in operation. Heat<br/>
removal by the cooling water from the plunger is also recorded. An optimization problem is formulated to<br/>
determine heat transfer coefficients which minimize the difference between the measured data and analysis<br/>
results. Significant time savings are achieved in finite difference based sensitivity computation during the<br/>
optimization by employing distributed computing technique. The analysis results by the optimum heat transfer<br/>
coefficients are found to agree well with the measured data.

목차 (Table of Contents)

• Abstract
  
• 1. 서론
  
• 2. 프레스 성형공정 해석
  
• 3. 온도 측정 및 해석결과와의 비교
  
• 4. 최적화기법에 의한 열전달 계수 결정
  
• 5. 분산 컴퓨팅에 의한 최적화
  
• 6. 결론
  
• 후기
  
• 참고문헌