Investigation and back-propagation modeling of base pressure at sonic and supersonic Mach numbers

被引:46
作者
Afzal, Asif [1 ]
Khan, Sher Afghan [2 ]
Islam, Md Tariqul [3 ]
Jilte, R. D. [4 ]
Khan, Ambareen [5 ]
Soudagar, Manzoore Elahi M. [6 ]
机构
[1] Visvesvaraya Technol Univ, Dept Mech Engn, PA Coll Engn, Mangalore 574153, India
[2] IIUM Malaysia, Fac Engn, Dept Mech Engn, Kuala Lumpur 50728, Malaysia
[3] Univ Queensland, Sch Chem Engn, Brisbane, Qld 4072, Australia
[4] Lovely Profess Univ, Dept Mech Engn, Phagwara 144411, Punjab, India
[5] Univ Sains Malaysia, Sch Aerosp Engn, Nibong Tebal 14300, Penang, Malaysia
[6] Univ Malaya, Fac Engn, Dept Mech Engn, Kuala Lumpur 50603, Malaysia
关键词
SUDDENLY EXPANDED FLOWS; ACTIVE CONTROL; EXPANSION;
D O I
10.1063/5.0022015
中图分类号
O3 [力学];
学科分类号
08 ; 0801 ;
摘要
The experimental analysis of base pressure in a high-speed compressible flow is carried out. The flow is made to expand abruptly from the nozzle into an enlarged duct at fifteen sonic and supersonic Mach numbers. The analysis is made for variation in the nozzle pressure ratio (NPR), length to diameter ratio, and area ratio. The effect of active micro-jets on the base and wall pressure is assessed. The data visualization of the huge experimental data generated is performed using heat maps. For the first time, six back-propagation neural network models (BPMs) are developed based on input and output possibilities to predict the pressure in high-speed flows. The experimental analysis revealed that depending upon the type of expansion, the base pressure changes. A jet of air blown at the base using micro-jets is found to be effective in increasing the base pressure during the under-expansion regime, while the wall pressure remains unaffected. The data visualization provided an insight into the highest impact on the base pressure by the NPR. The six BPMs with two hidden layers having four neurons per layer are found to be most suitable for the regression analysis. BPM 5 and BPM 6 accurately predict the highly non-linear data of the base and wall pressure.
引用
收藏
页数:17
相关论文
共 36 条
[1]   Parallelization Strategies for Computational Fluid Dynamics Software: State of the Art Review [J].
Afzal, Asif ;
Ansari, Zahid ;
Faizabadi, Ahmed Rimaz ;
Ramis, M. K. .
ARCHIVES OF COMPUTATIONAL METHODS IN ENGINEERING, 2017, 24 (02) :337-363
[2]  
Ahmed M., 2012, Scientific Research and Essays, V7, P3099
[3]   BASE PRESSURE AND NOISE PRODUCED BY ABRUPT EXPANSION OF AIR IN A CYLINDRICAL DUCT [J].
ANDERSON, JS ;
WILLIAMS, TJ .
JOURNAL OF MECHANICAL ENGINEERING SCIENCE, 1968, 10 (03) :262-&
[4]   Fuzzy c-Least Medians clustering for discovery of web access patterns from web user sessions data [J].
Ansari, Zahid ;
Faizabadi, Ahmed Rimaz ;
Afzal, Asif .
INTELLIGENT DATA ANALYSIS, 2017, 21 (03) :553-575
[5]  
Baig M.A.A., 2016, IOSR J MECH CIV ENG, V16, P83, DOI [10.9790/1684-16053038389, DOI 10.9790/1684-16053038389]
[6]  
Bethesda, 2012, USRDS 2012 ANN DAT R, V2, P237, DOI DOI 10.1109/ICMLC.2010.74
[7]  
Eliade M., 2007, DILEMATECA, VII, P50
[8]   The impact of caesarean section on breastfeeding initiation, duration and difficulties in the first four months postpartum [J].
Hobbs, Amy J. ;
Mannion, Cynthia A. ;
McDonald, Sheila W. ;
Brockway, Meredith ;
Tough, Suzanne C. .
BMC PREGNANCY AND CHILDBIRTH, 2016, 16
[9]   Control of suddenly expanded flow [J].
Khan, S. A. ;
Rathakrishnan, E. .
AIRCRAFT ENGINEERING AND AEROSPACE TECHNOLOGY, 2006, 78 (04) :293-309
[10]  
Khan S.A., 2018, INT J MECH PROD ENG, V8, P1057, DOI [10.24247/ijmperdjun2018111, DOI 10.24247/IJMPERDJUN2018111]