ANN evaluation of bearing strength on pin loaded composite plates in different environmental conditions

Ayla Tekin; Ayşe Öndürücü; Ümran Esendemir

doi:10.3139/120.111059

Article

ANN evaluation of bearing strength on pin loaded composite plates in different environmental conditions

Ayla Tekin , Ayşe Öndürücü and Ümran Esendemir

Published/Copyright: July 10, 2017

Published by

Become an author with De Gruyter Brill

Submit Manuscript Author Information

From the journal Materials Testing Volume 59 Issue 7-8

Abstract

The main aim of this study is to estimate the bearing strength on pin loaded composite plates using test data of different ambient conditions depending on time. Artificial neural network (ANN) tool is used for prediction purpose. Artificial neural network program is developed by MATLAB software. The composite plates were divided into nine groups, each one contains 60 specimens. These groups were kept in different environment conditions. To obtain the optimum geometrical dimensions of specimens, the distance from hole axis to the side of the test sample of the pin diameter ratio (E/D) and the sample width to pin diameter ratio (W/D) were changed systematically in the experimental samples. Data from fatigue test results obtained from the multi-layered, feedforward and backpropagation algorithm were used to train the artificial neural network model. In modeling of artificial neural network, the test conditions, the test periods, the day intervals, the distance from hole axis to the side of test sample of the pin diameter ratio (E/D) and the sample width to pin diameter ratio (W/D) were used as input parameters and the bearing strength data was used as output parameter. The values obtained from the artificial neural network training and testing were evaluated by applying statistical analyses that are widely used in ANN models. It is widely known that difficult experimental studies and complexity of the analytical expression could be solved by ANN models as it was seen in many studies. This study has also shown that artificial neural network is a convenient method for predicting the bearing strength on pin loaded composite plates.

Kurzfassung

Das Hauptziel der diesem Beitrag zugrunde liegenden Studie bestand darin, die Tragfestigkeit von bolzenbelasteten Kompositplatten abzuschätzen, in dem Testergebnisse unter verschiedenen zeitabhängigen Umgebungsbedingungen verwendet wurden. Zur Vorhersage wurde das Hilfsmittel des künstlichen neuronalen Netzes eingesetzt (Artificial Neural Network (ANN)). Das ANN-Programm wurde mit der MATLAB Software entwickelt. Die Kompositplatten wurden in neun Gruppen aufgeteilt, von denen jede 60 Proben enthielt. Diese Gruppen wurden verschiedenen Umgebungsbedingungen unterworfen. Um die optimalen geometrischen Bedingungen der Proben zu erhalten, wurden das Verhältnis des Abstandes von der Lochachse bis zur Seite der Probe und dem Bolzendurchmesser (E/D) und das Verhältnis zwischen der Probenbreite und dem Bolzendurchmesser (W/D) systematisch verändert. Dabei wurden die Daten aus den Ermüdungsversuchen verwendet, um das ANN zu trainieren. Bei der Modellierung des ANN wurden die Testbedingungen, die Testdauern, die Tagesintervalle, das Verhältnis des Abstandes von der Lochachse zum Probenrand und dem Bolzendurchmesser (E/D) und das Verhältnis zwischen Probenbreite und Bolzendurchmesser (W/D) als Inputparameter und die Tragfestigkeit als Outputparameter verwendet. Die Werte, die sich aus dem ANN-Training und der Prüfung ergaben, wurden mittels statistischer Analyse evaluiert, wie sie häufig in ANN-Modellen eingesetzt werden. Wie in vielen schwierigen experimentellen Studien mit komplexen analytischen Ausdrücken, so auch in dieser Untersuchung, zeigte sich, dass ANN ein geeignetes Verfahren ist, um die Tragfestigkeit von bolzenbelasteten Kompositplatten vorherzusagen.

*Correspondence Address, Associate Prof. Dr. Ayşe Öndürücü, Department of Mechanical Engineering, Faculty of Engineering, Süleyman Demirel University, 32260 Isparta, Turkey, E-mail: ayla.tekin@cbu.edu.tr, ayseondurucu@sdu.edu.tr, umranesendemir@sdu.edu.tr

Assist. Prof. Dr. Ayla Tekin, born in 1973, graduated with a BSc degree in Mechanical Engineering from Süleyman Demirel University, Isparta, Turkey in 1996. She obtained her MSc degree in Mechanical Engineering from Süleyman Demirel University in 1998. She received her PhD degree in Mechanical Engineering from Celal Bayar University, Manisa, Turkey, in 2007. Since 2007, she has been working as Assistant Professor at Manisa Celal Bayar University, Turkey. Her expertise areas are mechanical vibrations and artificial neural networks.

Assoc. Prof. Dr. Ayşe Öndürücü, born in 1974, graduated with a BSc degree in Mechanical Engineering from Denizli Engineering Faculty, Denizli, Turkey, in 1995. She obtained her MSc degree from the Mechanical Engineering Department at Pamukkale University, Denizli, Turkey in 1998. Then, she received her PhD degree in Mechanical Engineering in 2003 from Süleyman Demirel University, Isparta, Turkey. Since 2013, she has been working as Associate Professor in the same department for Süleyman Demirel University. Her expertise areas are solid mechanics, composite materials and finite element method.

Prof. Dr. Ümran Esendemir, born in 1971, graduated with a BSc degree in Mechanical Engineering from Akdeniz University, Isparta Engineering Faculty, Isparta, Turkey, in 1991. She obtained her M.S. and Ph.D. degrees from the University of Süleyman Demirel, Isparta, Turkey, in 1994 and 1999, respectively, both in Mechanical Engineering. She was appointed as Assistant Professor in 1994 and Associate Professor in 2011 at Süleyman Demirel University. Since 2017, she has been working as Professor at Süleyman Demirel University. Her research expertise includes composite materials mechanics and anisotropic elasticity.

References

1 E.Madenci, S.Shkarayev, B.Sergeev, D. W.Opliger, P.Shyprykevich: Analysis of composite laminates with multiple fasteners, International Journal of Solids and Structures35 (1998), No. 15, pp. 1793–181110.1016/S0020-7683(97)00152-2Search in Google Scholar

2 Y.Pekbey: Effect of preload moment and geometrical parameters to failure behaviour of pin-loaded laminated composite plates, Proc. of 8^th International Fracture Conference, Istanbul, Turkey (2007), pp. 554–562Search in Google Scholar

3 A.Koruvatan: Failure Analysis of the Laminated Composite Plates with Pin/Bolt Loaded Joints Manufactured under Various Cure Temperatures and Periods, PhD Thesis, Balıkesir University, Graduate School of Natural and Applied Sciences, Balıkesir, Turkey (2008)Search in Google Scholar

4 K.Turan, M.Kaman, M.Gür: Experimental and numerical failure analysis in the laminated composite plates with circular hole, Proc. of 5^th International Advanced Technologies Symposium (IATS’09), Karabük, Turkey (2009)Search in Google Scholar

5 C.Echavarría, P.Haller, A.Salenikovich: Analytical study of a pin-loaded hole in elastic orthotropic plates, Composite Structures79 (2007), No. 1, pp. 107–11210.1016/j.compstruct.2005.11.038Search in Google Scholar

6 A.Ataş: Woven Fiber Reinforced Pin Parallel Investigation of the Behavior under Static Load of Perforated Polyester Laminated Composite Plate, MSc Thesis, Balikesir University, Graduate School of Natural and Applied Sciences, Balıkesir, Turkey (2007)Search in Google Scholar

7 D.Özer, M.Özbay: Finite element analysis by the method of the elastic stresses in the composite plate with circular hole in the middle plane loaded, Gazi University Journal of the Faculty of Engineering and Architecture19 (2004), No. 1, pp. 51–57Search in Google Scholar

8 N.Allahverdi: Expert Systems, Artificial Intelligence Applications, Atlas Publishing, Istanbul, Turkey (2011)Search in Google Scholar

9 E.Öztemel: Artificial Neural Networks, Daisy Publishing, Istanbul, Turkey (2012)Search in Google Scholar

10 Z.Zhang, K.Friedrich: Artificial neural networks applied to polymer composites: A review, Composites Science and Technology63 (2003), pp. 2029–204410.1016/S0266-3538(03)00106-4Search in Google Scholar

11 L. H.Yam, Y. J.Yan, J. S.Jiang: Vibration-based damage detection for composite structures using wavelet transform and neural network identification, Composite Structures60 (2003), No. 4, pp. 403–41210.1016/S0263-8223(03)00023-0Search in Google Scholar

12 H.Kadi: Modeling the mechanical behavior of fiber-reinforced polymeric composite materials using artificial neural networks, Composite Structures73 (2006), No. 1, pp. 1–2310.1016/j.compstruct.2005.01.020Search in Google Scholar

13 R.Haj-Ali, H.Kim: Nonlinear constitutive models for FRP composites using artificial neural networks, Mechanics of Materials39 (2007), No. 12, pp. 1035–104210.1016/j.mechmat.2007.05.004Search in Google Scholar

14 V.Albuquerque, J.Tavares, L.Durao: Evaluation of delamination damage on composite plates using an artificial neural network for the radiographic image analysis, Journal of Composite Materials44 (2009), No. 9, pp. 1139–115910.1177/0021998309351244Search in Google Scholar

15 Ü.Esendemir, A. M.Cabıoğlu: Investigating bearing strength of pin-loaded composite plates in different environmental conditions, Journal of Reinforced Plastics & Composites32 (2013), No. 22, pp. 1685–169710.1177/0731684413500858Search in Google Scholar

16 A. M.Eker, M.Dikmen, S.Cambazoğlu, H. Ş. B.Düzgün, H.Akgün: Application of artificial neural network and logistic regression methods to landslide susceptibility mapping and comparison of the results for the Ulus district, Bartın, Journal of the Faculty of Engineering and Architecture of Gazi University27 (2012), No. 1, pp. 163–17310.1080/13658816.2014.953164Search in Google Scholar

17 Ç.Karataş, A.Sözen, E.Dülek: Modelling of residual stresses in shot peened material C-1020 by artificial neural network, Expert Systems with Applications36 (2009), No. 2, pp. 3514–352110.1016/j.eswa.2008.02.012Search in Google Scholar

18 A. M.Cabıoğlu: Investigating Bearing Strength on Pin Loaded Composite Plates in Different Temperature Environments, MSc Thesis, Süleyman Demirel University, Isparta, Turkey (2012)Search in Google Scholar

Published Online: 2017-07-10

Published in Print: 2017-07-14

You are currently not able to access this content.

Articles in the same Issue

https://doi.org/10.3139/120.111059