Fatigue life prediction of sandwich composite materials under flexural tests using a Bayesian trained artificial neural network

The authors discuss the use of an artificial neural network (ANN) to estimate fatigue lifetime of a sandwich composite material structure subjected to cyclic three-point bending loads. A total of 27 samples (three different loading levels for nine samples each) were investigated to provide training, validation and testing data for a series of multi-layer perceptron ANNs. The networks were implemented using both conventional maximum likelihood and Bayesian evidence based training algorithms. It was found that the Bayesian evidence based approach provided a superior and smoother fit to the experimental data. Completely independent fatigue tests were conducted at intermediate levels of loading to evaluate the capacity of the fitted ANN model to interpolate to previously unseen regions of data. Excellent agreement was obtained between the model predicted outputs and the new experimental data. The capability and estimation of predication error when using the Bayesian technique is discussed along with the application of the model to generate classical S–N lifetime curves.

[1]  L. Nicolais,et al.  Effect of Stress Ratio on the Flexural Fatigue Behaviour of Continuous Strand Mat Reinforced Plastics , 1996 .

[2]  D. Mackay,et al.  Bayesian Neural Network Analysis of Fatigue Crack Growth Rate in Nickel Base Superalloys , 1996 .

[3]  W. Hwang,et al.  Cumulative Damage Models and Multi-Stress Fatigue Life Prediction , 1986 .

[4]  D. Zenkert,et al.  Handbook of Sandwich Construction , 1997 .

[5]  A. Bezazi,et al.  Modelling the flexural behaviour of sandwich composite materials under cyclic fatigue , 2004 .

[6]  Harald Zenner,et al.  Lifetime calculation under variable amplitude loading with the application of artificial neural networks , 2005 .

[7]  S. Tsai,et al.  Introduction to composite materials , 1980 .

[8]  Fernand Ellyin,et al.  Effect of stress ratio on the fatigue of unidirectional glass fibre/epoxy composite laminae , 1994 .

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

[10]  A. Bezazi,et al.  Investigation of Cross-Ply Laminates Behaviour in Three Point Bending Tests. Part II: Cyclic Fatigue Tests , 2003 .

[11]  David J. C. MacKay,et al.  Information Theory, Inference, and Learning Algorithms , 2004, IEEE Transactions on Information Theory.

[12]  T. Adam,et al.  Life–prediction for constant–stress fatigue in carbon–fibre composites , 1997, Philosophical Transactions of the Royal Society of London. Series A: Mathematical, Physical and Engineering Sciences.

[13]  G. Caprino,et al.  Fatigue lifetime of glass fabric/epoxy composites , 1999 .

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

[15]  Peter L. Bartlett,et al.  The Sample Complexity of Pattern Classification with Neural Networks: The Size of the Weights is More Important than the Size of the Network , 1998, IEEE Trans. Inf. Theory.

[16]  Zhongya Zhang,et al.  Artificial neural networks applied to polymer composites: a review , 2003 .

[17]  Keith Worden,et al.  Nonlinearity in Structural Dynamics , 2019 .

[18]  Christopher M. Bishop,et al.  Neural networks for pattern recognition , 1995 .

[19]  W. Hwang,et al.  Statistical study of strength and fatigue life of composite materials , 1987 .

[20]  Ian T. Nabney,et al.  Netlab: Algorithms for Pattern Recognition , 2002 .

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

[22]  J. Berthelot,et al.  Flexural Fatigue Behavior of Cross-Ply Laminates: An Experimental Approach , 2003 .

[23]  Darryl P Almond,et al.  The use of neural networks for the prediction of fatigue lives of composite materials , 1999 .

[24]  W. Hwang,et al.  Fatigue of Composites—Fatigue Modulus Concept and Life Prediction , 1986 .

[25]  Yousef Al-Assaf,et al.  Prediction of the fatigue life of unidirectional glass fiber/epoxy composite laminae using different neural network paradigms , 2002 .

[26]  Yousef Al-Assaf,et al.  FATIGUE LIFE PREDICTION OF COMPOSITE MATERIALS USING POLYNOMIAL CLASSIFIERS AND RECURRENT NEURAL NETWORKS , 2007 .

[27]  D. P. Almond,et al.  21 – A neural-network approach to fatigue-life prediction , 2003 .