Artificial neural networks applied to epoxy composites reinforced with carbon and E-glass fibers: Analysis of the shear mechanical properties

Abstract Inspired by the biological nerve system, artificial neural networks (ANN) have been tools of artificial intelligence for data classification and pattern recognition, and can be used to simulate a wide variety of non-linear complex scientific systems. Artificial neural networks are being used in medical applications; image recognition and control of dynamic systems, but only recently have been considered for the prediction of the mechanical behavior of materials and particularly composites. In this work, ANNs were considered specifically to predict the shear stress–strain behavior from carbon fiber/epoxy and glass fiber/epoxy composites. A multilayered neural network perceptron (MLP) architecture was used, and the results showed that the application of the Levenberg–Marquardt learning algorithm leads to a high predictive quality to epoxy composites, i.e. nearly 80% of standard error of prediction was found to be ≥0.9. The initial tests considered a simple architecture 3-[3-3] 2 -1 resulting in low predictive quality. However, increasing the number of neurons in the hidden layers and the number of training instances resulted in an enhancement of the neural network predictive quality.

[1]  Ignacio J. Turias,et al.  Modelling the effective thermal conductivity of an unidirectional composite by the use of artificial neural networks , 2005 .

[2]  Bobby G. Sumpter,et al.  Computational neural networks and the rational design of polymeric materials: the next generation polycarbonates , 1998 .

[3]  Christos Andrea Antoniou,et al.  Experiments in speech recognition using a modular MLP architecture for acoustic modelling , 2003, Inf. Sci..

[4]  Michael T. Manry,et al.  A neural network training algorithm utilizing multiple sets of linear equations , 1996, Conference Record of The Thirtieth Asilomar Conference on Signals, Systems and Computers.

[5]  Yonghong Tan,et al.  Optimization techniques for the design of a neural predictive controller , 1996, Neurocomputing.

[6]  L. Ye,et al.  Functionalized composite structures for new generation airframes: a review , 2005 .

[7]  Ji Zhong,et al.  Acquiring the constitutive relationship for a thermal viscoplastic material using an artificial neural network , 1996 .

[8]  Daniel W. C. Ho,et al.  A new training and pruning algorithm based on node dependence and Jacobian rank deficiency , 2006, Neurocomputing.

[9]  Simon Haykin,et al.  Neural Networks: A Comprehensive Foundation , 1998 .

[10]  Mohammad Bagher Menhaj,et al.  Training feedforward networks with the Marquardt algorithm , 1994, IEEE Trans. Neural Networks.

[11]  David Enke,et al.  The use of data mining and neural networks for forecasting stock market returns , 2005, Expert Syst. Appl..

[12]  A. C. Okafor,et al.  Delamination Prediction in Composite Beams with Built-In Piezoelectric Devices Using Modal Analysis and Neural Network , 1996 .

[13]  Klaus Friedrich,et al.  Dynamic mechanical properties of PTFE based short carbon fibre reinforced composites: experiment and artificial neural network prediction , 2002 .

[14]  Hamid Garmestani,et al.  Prediction of nonlinear viscoelastic behavior of polymeric composites using an artificial neural network , 2006 .

[15]  Yousef Al-Assaf,et al.  Fatigue life prediction of unidirectional glass fiber/epoxy composite laminae using neural networks , 2001 .

[16]  P. Werbos,et al.  Beyond Regression : "New Tools for Prediction and Analysis in the Behavioral Sciences , 1974 .

[17]  J. Karger‐Kocsis,et al.  Artificial neural network predictions on erosive wear of polymers , 2003 .

[18]  Yousef Al-Assaf,et al.  Energy-based fatigue life prediction of fiberglass/epoxy composites using modular neural networks , 2002 .

[19]  N. Huber,et al.  A neural network tool for identifying the material parameters of a finite deformation viscoplasticity model with static recovery , 2001 .