| Summary: | The paper presents the optimisation of thin-walled composite structures on a representative sample of a thin-walled column made of carbon laminate with a channel section-type profile. The optimisation consisted of determining the configuration of laminate layers for which the tested structure has the greatest resistance to the loss of stability. The optimisation of the layer configuration was performed using two methods. The first method, divided into two stages to reduce the time, was to determine the optimum arrangement angle in each laminate layer using finite element methods (FEM). The second method employed artificial neural networks for predicting critical buckling force values and the creation of an optimisation tool. Artificial neural networks were combined into groups of networks, thereby improving the quality of the obtained results and simplifying the obtained neural networks. The results from computations were verified against the results obtained from the experiment. The optimisation was performed using ABAQUS<sup>®</sup> and STATISTICA<sup>®</sup> software.
|
|---|