Optimisation of the lateral buckling strength of corrugated composite material plate by neural networks method

Abstract

Enhancing the buckling strength of laminated composite materials can be achieved in numerous ways. One method involves corrugating the laminated composite material in one direction. Corrugation provides good buckling strength in the direction perpendicular to the corrugation but a low buckling strength in the same direction as the corrugation. This investigation used composite materials strips implanted in the direction of the laminate's corrugation to modify the ability to buckle without excessive weight on the laminate. Finite elements were applied to analyse the problem. In addition, to overcome the extensive computational requirements, a neural network (NN) system was utilised to model the study case and then optimise the structure. The NN was trained by the results of the finite elements. The parameters examined and their effects on buckling strength include the number of strips, number of layers of strips and dimension of strips. Results confirmed that the technique of strengthening the laminate using strips in the direction of corrugation waves is beneficial for increasing the critical buckling load. Specifically, the optimisation result presented an increase of 52 times in the buckling load strength versus approximately twice the increase in the mass of the plate. Using the NN to simulate and optimise the structure is a powerful approach that consumes less time than employing the finite element method.