Adaptive neuro-fuzzy inference system in modelling fatigue life of multidirectional composite laminates

Adaptive neuro-fuzzy inference system (ANFIS) has been successfully used for the modelling of fatigue behaviour of a multidirectional composite laminate. The evaluation of the neuro-fuzzy model has been performed using a data base containing 257 valid fatigue data points. Coupons were cut at 0° on-axis and 15°, 30°, 45°, 60°, 75°, and 90° off-axis directions from an E-glass/polyester multidirectional laminate with a stacking sequence of [0/(±45)2/0]T. Constant amplitude fatigue tests at different tensile and compressive conditions were conducted for the determination of the 17 S–N curves. The modelling accuracy of this novel, in this field, computational technique is very high. For all cases studied, it has been proved that a portion of around 50% of the available data are adequate for accurate modelling of the fatigue behaviour of the material under consideration. The new technique is a stochastic process which leads to the derivation of a multi-slope S–N curve based on the available experimental data without the need for any assumptions. Employment of this technique can lead to a substantial decrease of the experimental cost for the determination of reliable fatigue design allowables.