Browse > Article

Performance Analysis of Neural Network Compensation Algorithm of Multiaxis Thrust Measurement Stand  

Kim, Joung-Keun (국방과학연구소)
Publication Information
Journal of the Korean Society of Propulsion Engineers / v.11, no.4, 2007 , pp. 52-58 More about this Journal
Abstract
The irregular fuel surface was observed by the visualization of hybrid rocket combustion. Even though the test condition maintained oxidizer rich environment, the irregular dark fuel surface was formed as the result of incomplete combustion. In order to investigate the correlation of the characteristics of oxidizer flow and the irregular fuel surface, various flow conditions were imposed such as swirl flow, induced swirl flow by helical fuel configuration and straight flow. Test results revealed no correlation was found between oxidizer flow condition and irregular fuel surface. And this can be a commonly observed phenomena in the tests with different fuel/oxidizer combination. Thus, the irregular fuel surface can be a result of the interaction of blowing flow of vaporized fuel and the boundary layer of oxidizer flow. A further study will be required to confirm the assumption for the formation of irregular fuel surface.
Keywords
Thrust Vector Control; Multiaxis Thrust Measurement Stand; Crosstalk; Crosstalk Compensation Algorithm;
Citations & Related Records
연도 인용수 순위
  • Reference
1 A. Richasdson, "A Comparison Of Neural Network Training Algorithms", Kristina AIAA -97-7303
2 김중근,박익수, " 신경망을 적용한 다축시험대 분석기법 연구", 국방과학연구소, 2001
3 R.L. Davis and A.F. Domal, "Techniques for successful 6-component force measurement of Solid-Propellant Rocket Motor", Aro Inc
4 Peiman G. Maghami, " Design Of Neural Networks For Fast Convergence And Accuracy", AIAA 98-1780
5 Norgaard, M., Jorgensen, C. C., and Ross, J.C., " Neural Network Prediction of New Aircraf Design Coeficients", NASA TM 112197, May 1997
6 김중근,박익수,이규준,윤일선, "고체 추진기관 6분력 시험대의 모델링 기법", 한국추진공학회지, 제5권 제 3호, 2001, pp.26-33
7 김중근 외 5명, "대형 연소시험대 교정 및 연소시험 분석 프로그램 개발 보고서, 국방과학연구소 보고서", MSDC-521-980906
8 Slawomir W. Stepniewski, " Designing Compact Feedforward Neural Model With Small Training Data Set", AIAA 2000-0170