Multivariable and multiaxial fatigue life assessment of composite materials using neural networks

In the present paper, multivariable and multiaxial fatigue life assessment of multidirectional composite laminates of polymeric-based composites was investigated using neural networks (NN) model. NN with non-linear auto-regressive exogenous inputs (NARX) structure was employed for the problem considered and the training algorithm of Levenberg-Marquardt with Bayesian regularization was chosen. The task of fatigue life assessment was accomplished in such a way that it was realized as one-step ahead prediction with respect to each stress level-S corresponding to stress ratio values-R. In addition, by sliding over one-step to one-step of the stress levels, the prediction dynamically covered all the corresponding spectrum loadings including multiaxial orientations examined. As a result, fatigue life assessment of the composite materials can be fashioned for a wide spectrum of loading in an efficient manner based upon solely the training data as the basis of the NARX regressor, thus developed multivariable and multiaxial fatigue analysis.