A Numerical Study of Response Surface Based Shape Optimization Using Neural Networks

Lawrence Munashe Mavima; Tanaka Mukoko; Kudzanai Shinda Zhou

doi:doi:10.11648/j.ajmcm.20200504.14

| Peer-Reviewed

A Numerical Study of Response Surface Based Shape Optimization Using Neural Networks

Lawrence Munashe Mavima, Tanaka Mukoko, Kudzanai Shinda Zhou

Published in American Journal of Mathematical and Computer Modelling (Volume 5, Issue 4)

Received: 6 October 2020 Accepted: 24 October 2020 Published: 30 November 2020

Views: Downloads:

Download PDF

Share This Article

Twitter
Linked In
Facebook

Abstract

Industrial pipes that are used for fluid transport generally have to undergo many changes of shape to accommodate interfacing equipment related to plant operation, which results in flow maldistribution zones and higher pressure drops, and in turn leads to higher power consumption. In an attempt to redress this problem, ANSYS, a commercial Computational Fluid Dynamics (CFD) software, is used to perform numerical simulations based on a deterministic computational model of the internal fluid flow using the Reynolds Averaged Navier Stokes equations (RANS), a multi objective optimization study employing Response surface methodology and artificial neural networks. This numerical analysis has been performed on a galvanized steel duct for water recirculation. The focus of the paper is the study of the effect of a chosen set of several geometrical dimensions on the pressure drop and flow distribution inside the duct. Subsequently, a new set of designs with different geometrical parameters has been obtained to minimize the pressure drop and achieve a more uniform flow distribution by using artificial neural networks to generate a response surface and further employing Screening (Shifted-Hammersley sampling) as the optimization method that was used to select the best designs from amongst those that have been generated from the response surface.

Published in	American Journal of Mathematical and Computer Modelling (Volume 5, Issue 4)
DOI	10.11648/j.ajmcm.20200504.14
Page(s)	116-126
Creative Commons	This is an Open Access article, distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution and reproduction in any medium or format, provided the original work is properly cited.
Copyright	Copyright © The Author(s), 2020. Published by Science Publishing Group

Keywords

Artificial Neural Networks, Response Surface Methodology, Turbulence Modeling, Shifted Hammersley Sampling, Multi-objective Optimization

References

[1]	M. Dehghani, H. Ajam and S. Farahat, "Automated Diffuser Shape Optimization based on CFD Simulations and Surrogate Modeling," Journal of Applied Fluid Mechanics, vol. 9, no. 5, pp. 2527-2535, 2016.
[2]	N. N. Sonawane and P. P. P. Patil, "Design and Optimization of Aerodynamics Nozzle for maximum thrust using Multidisciplinary Design Optimization (MDO)- An overview," International Research Journal of Engineering and Technology, vol. 2, no. 8, pp. 1681-1683, 2015.
[3]	S. J. Winne and M. Chandrasekaran, "OPTIMIZATION OF NOZZLE: CONVERGENCE USING ANSYS WITH RSM, MOGA," Asian Research Publishing Network, vol. 10, no. 13, pp. 5486-5489, 2015.
[4]	J. Behin and N. Farhadian, "Response surface methodology and artificial neural network modeling of reactive red 33 decolorization by Ozone/UV in a bubble column reactor," Advances in Environmental Technology, pp. 33-44, 2016.
[5]	S. E. Hamed Safikhani, "Pareto based multi objective optimization of turbulent heat transfer flow in helically corrugated tubes," Applied Thermal Engineering, vol. 95, pp. 275-280, 2016.
[6]	B. Wolfram, L. Martin, B. Michael and W. Graf, "Neural Network Based Response Surface Methods – a Comparative Study," in 5. LS-DYNA Anwenderforum, Robustheit / Optimierung II, Ulm, 2006.
[7]	G. Y. C. W. L. Z. S. Yong Sun, "Artificial neural networks with response surface methodology for optimization of selective CO2 hydrogenation using K-promoted iron catalyst in a micro channel reactor," Journal of CO2 utilization, vol. 24, pp. 10-31, 2018.
[8]	W. H. J. H.-j. S. Y.-f. W. Zheng, "Study on response surface methodology with artificial neural networks application in aerodynamics optimization," in 2009 WRI World congress on computer science and information engineering, Los Angeles, 2009.
[9]	W. C. D., Turbulence Modeling for CFD, La Cañada: DCW Industries, 1998.
[10]	S. B. Pope, Turbulent Flows, Cambridge University Press, 2000.
[11]	E. Helgason and Krajnovi, "AERODYNAMIC SHAPE OPTIMIZATION OF A PIPE USING THE ADJOINT METHOD," in ASME 2012 International Mechanical Engineering Congress & Exposition, Houston, Texas, 2012.
[12]	A. Mukhtar, N. K. Ching and M. Z. Yusoff, “Optimal Design of Opening Ventilation Shaft by Kriging Metamodel Assisted Multi-objective Genetic Algorithm,” International Journal of Modeling and Optimization, vol. 7, no. 2, pp. 92-97, 2017.
[13]	S. E. Dreyfus, "Artificial Neural Networks, Back Propagation, and the Kelley-Bryson Gradient Procedure," Journal of Guidance, Control, and Dynamics, vol. 13, no. 5, p. 926–928, 1990.
[14]	E. Alpaydin, Introduction to Machine Learning, MIT Press, 2010.
[15]	J.-T. Oh and N. B. Chien, "Optimization Design by Coupling Computational Fluid," in Computational Fluid Dynamics - Basic Instruments and Applications in Science, IntechOpen, 2018, pp. 123-136.
[16]	I. Cayiroglua and R. Kilic, "Wing Aerodynamic Optimization by Using Genetic Algoritm and Ansys," ACTA PHYSICA POLONICA A, vol. 132, no. 3-II, Special issue of the 3rd International Conference on Computational and Experimental Science and Engineering (ICCESEN 2016), pp. 981-985, 2017.
[17]	G. Lei, "Adaptive random search in quasi-monte carlo methods for global optimization," computers and mathematics with applications, no. 43, pp. 747-754, 2001.
[18]	M. Karimi, G. Akdogan and S. M. Bradshaw, "EFFECTS OF DIFFERENT MESH SCHEMES AND TURBULENCE MODELS IN CFD MODELLING OF STIRRED TANKS," Physicochemical Problems of Mineral Processing, vol. 48, no. 2, p. 513−531, 2012.

Cite This Article

Plain Text BibTeX RIS

APA Style

Lawrence Munashe Mavima, Tanaka Mukoko, Kudzanai Shinda Zhou. (2020). A Numerical Study of Response Surface Based Shape Optimization Using Neural Networks. American Journal of Mathematical and Computer Modelling, 5(4), 116-126. https://doi.org/10.11648/j.ajmcm.20200504.14

Copy | Download

ACS Style

Lawrence Munashe Mavima; Tanaka Mukoko; Kudzanai Shinda Zhou. A Numerical Study of Response Surface Based Shape Optimization Using Neural Networks. Am. J. Math. Comput. Model. 2020, 5(4), 116-126. doi: 10.11648/j.ajmcm.20200504.14

Copy | Download

AMA Style

Lawrence Munashe Mavima, Tanaka Mukoko, Kudzanai Shinda Zhou. A Numerical Study of Response Surface Based Shape Optimization Using Neural Networks. Am J Math Comput Model. 2020;5(4):116-126. doi: 10.11648/j.ajmcm.20200504.14

Copy | Download

@article{10.11648/j.ajmcm.20200504.14,
  author = {Lawrence Munashe Mavima and Tanaka Mukoko and Kudzanai Shinda Zhou},
  title = {A Numerical Study of Response Surface Based Shape Optimization Using Neural Networks},
  journal = {American Journal of Mathematical and Computer Modelling},
  volume = {5},
  number = {4},
  pages = {116-126},
  doi = {10.11648/j.ajmcm.20200504.14},
  url = {https://doi.org/10.11648/j.ajmcm.20200504.14},
  eprint = {https://article.sciencepublishinggroup.com/pdf/10.11648.j.ajmcm.20200504.14},
  abstract = {Industrial pipes that are used for fluid transport generally have to undergo many changes of shape to accommodate interfacing equipment related to plant operation, which results in flow maldistribution zones and higher pressure drops, and in turn leads to higher power consumption. In an attempt to redress this problem, ANSYS, a commercial Computational Fluid Dynamics (CFD) software, is used to perform numerical simulations based on a deterministic computational model of the internal fluid flow using the Reynolds Averaged Navier Stokes equations (RANS), a multi objective optimization study employing Response surface methodology and artificial neural networks. This numerical analysis has been performed on a galvanized steel duct for water recirculation. The focus of the paper is the study of the effect of a chosen set of several geometrical dimensions on the pressure drop and flow distribution inside the duct. Subsequently, a new set of designs with different geometrical parameters has been obtained to minimize the pressure drop and achieve a more uniform flow distribution by using artificial neural networks to generate a response surface and further employing Screening (Shifted-Hammersley sampling) as the optimization method that was used to select the best designs from amongst those that have been generated from the response surface.},
 year = {2020}
}

Copy | Download

TY  - JOUR
T1  - A Numerical Study of Response Surface Based Shape Optimization Using Neural Networks
AU  - Lawrence Munashe Mavima
AU  - Tanaka Mukoko
AU  - Kudzanai Shinda Zhou
Y1  - 2020/11/30
PY  - 2020
N1  - https://doi.org/10.11648/j.ajmcm.20200504.14
DO  - 10.11648/j.ajmcm.20200504.14
T2  - American Journal of Mathematical and Computer Modelling
JF  - American Journal of Mathematical and Computer Modelling
JO  - American Journal of Mathematical and Computer Modelling
SP  - 116
EP  - 126
PB  - Science Publishing Group
SN  - 2578-8280
UR  - https://doi.org/10.11648/j.ajmcm.20200504.14
AB  - Industrial pipes that are used for fluid transport generally have to undergo many changes of shape to accommodate interfacing equipment related to plant operation, which results in flow maldistribution zones and higher pressure drops, and in turn leads to higher power consumption. In an attempt to redress this problem, ANSYS, a commercial Computational Fluid Dynamics (CFD) software, is used to perform numerical simulations based on a deterministic computational model of the internal fluid flow using the Reynolds Averaged Navier Stokes equations (RANS), a multi objective optimization study employing Response surface methodology and artificial neural networks. This numerical analysis has been performed on a galvanized steel duct for water recirculation. The focus of the paper is the study of the effect of a chosen set of several geometrical dimensions on the pressure drop and flow distribution inside the duct. Subsequently, a new set of designs with different geometrical parameters has been obtained to minimize the pressure drop and achieve a more uniform flow distribution by using artificial neural networks to generate a response surface and further employing Screening (Shifted-Hammersley sampling) as the optimization method that was used to select the best designs from amongst those that have been generated from the response surface.
VL  - 5
IS  - 4
ER  -

Copy | Download

Author Information

Lawrence Munashe Mavima

Department of Mechanical Engineering, Saad Dahlab University, Blida, Algeria
Tanaka Mukoko

Department of Management and Engineering, Link?ping University, Link?ping, Sweden
Kudzanai Shinda Zhou

Department of Mechanical and Production Engineering, Harare Polytechnic, Harare, Zimbabwe

Download PDF

Sections

Plain Text BibTeX RIS

APA Style

Lawrence Munashe Mavima, Tanaka Mukoko, Kudzanai Shinda Zhou. (2020). A Numerical Study of Response Surface Based Shape Optimization Using Neural Networks. American Journal of Mathematical and Computer Modelling, 5(4), 116-126. https://doi.org/10.11648/j.ajmcm.20200504.14

Copy | Download

ACS Style

Lawrence Munashe Mavima; Tanaka Mukoko; Kudzanai Shinda Zhou. A Numerical Study of Response Surface Based Shape Optimization Using Neural Networks. Am. J. Math. Comput. Model. 2020, 5(4), 116-126. doi: 10.11648/j.ajmcm.20200504.14

Copy | Download

AMA Style

Lawrence Munashe Mavima, Tanaka Mukoko, Kudzanai Shinda Zhou. A Numerical Study of Response Surface Based Shape Optimization Using Neural Networks. Am J Math Comput Model. 2020;5(4):116-126. doi: 10.11648/j.ajmcm.20200504.14

Copy | Download

@article{10.11648/j.ajmcm.20200504.14,
  author = {Lawrence Munashe Mavima and Tanaka Mukoko and Kudzanai Shinda Zhou},
  title = {A Numerical Study of Response Surface Based Shape Optimization Using Neural Networks},
  journal = {American Journal of Mathematical and Computer Modelling},
  volume = {5},
  number = {4},
  pages = {116-126},
  doi = {10.11648/j.ajmcm.20200504.14},
  url = {https://doi.org/10.11648/j.ajmcm.20200504.14},
  eprint = {https://article.sciencepublishinggroup.com/pdf/10.11648.j.ajmcm.20200504.14},
  abstract = {Industrial pipes that are used for fluid transport generally have to undergo many changes of shape to accommodate interfacing equipment related to plant operation, which results in flow maldistribution zones and higher pressure drops, and in turn leads to higher power consumption. In an attempt to redress this problem, ANSYS, a commercial Computational Fluid Dynamics (CFD) software, is used to perform numerical simulations based on a deterministic computational model of the internal fluid flow using the Reynolds Averaged Navier Stokes equations (RANS), a multi objective optimization study employing Response surface methodology and artificial neural networks. This numerical analysis has been performed on a galvanized steel duct for water recirculation. The focus of the paper is the study of the effect of a chosen set of several geometrical dimensions on the pressure drop and flow distribution inside the duct. Subsequently, a new set of designs with different geometrical parameters has been obtained to minimize the pressure drop and achieve a more uniform flow distribution by using artificial neural networks to generate a response surface and further employing Screening (Shifted-Hammersley sampling) as the optimization method that was used to select the best designs from amongst those that have been generated from the response surface.},
 year = {2020}
}

Copy | Download

TY  - JOUR
T1  - A Numerical Study of Response Surface Based Shape Optimization Using Neural Networks
AU  - Lawrence Munashe Mavima
AU  - Tanaka Mukoko
AU  - Kudzanai Shinda Zhou
Y1  - 2020/11/30
PY  - 2020
N1  - https://doi.org/10.11648/j.ajmcm.20200504.14
DO  - 10.11648/j.ajmcm.20200504.14
T2  - American Journal of Mathematical and Computer Modelling
JF  - American Journal of Mathematical and Computer Modelling
JO  - American Journal of Mathematical and Computer Modelling
SP  - 116
EP  - 126
PB  - Science Publishing Group
SN  - 2578-8280
UR  - https://doi.org/10.11648/j.ajmcm.20200504.14
AB  - Industrial pipes that are used for fluid transport generally have to undergo many changes of shape to accommodate interfacing equipment related to plant operation, which results in flow maldistribution zones and higher pressure drops, and in turn leads to higher power consumption. In an attempt to redress this problem, ANSYS, a commercial Computational Fluid Dynamics (CFD) software, is used to perform numerical simulations based on a deterministic computational model of the internal fluid flow using the Reynolds Averaged Navier Stokes equations (RANS), a multi objective optimization study employing Response surface methodology and artificial neural networks. This numerical analysis has been performed on a galvanized steel duct for water recirculation. The focus of the paper is the study of the effect of a chosen set of several geometrical dimensions on the pressure drop and flow distribution inside the duct. Subsequently, a new set of designs with different geometrical parameters has been obtained to minimize the pressure drop and achieve a more uniform flow distribution by using artificial neural networks to generate a response surface and further employing Screening (Shifted-Hammersley sampling) as the optimization method that was used to select the best designs from amongst those that have been generated from the response surface.
VL  - 5
IS  - 4
ER  -

Copy | Download