Development of a computational predictive model for the nonlinear in-plane compressive response of sandwich panels with bio-foam

Abstract

In this study, sandwich structures with commercial-grade aluminium alloy skins and bio-inspired core (mycofoam) were fabricated and tested to obtain the axial compression response in terms of in-plane deformation measures and stress. The ensuing spectrum of response data from experimental tests were then fed into three different data driven models that include simple linear regression (SLR), artificial neural network (ANN) and adaptive neuro-fuzzy inference system (ANFIS). The performance of the models is compared in estimating the compressive response of sandwich panels with the mycofoam. To assess the performance of models, coefficient of determination (R2), root mean squared error (RMSE) and mean absolute error (MAE) are used. Eleven different training algorithms are tested in ANN and Bayesian Regularization backpropagation with 9 hidden neurons is found to be the optimum ANN structure. In ANFIS model, triangular-shaped membership function (MF) with 20 rules gives the highest performance among 8 different MFs. All three models are found to be capable in estimating the compressive response. ANFIS model has the highest performance, followed by ANN model then SLR model with R2, RMSE and MAE being 0.9999, 0.0818, 0.0415 for the training dataset; 0.9999, 0.1626, 0.0491 for the testing dataset and 0.9999, 0.0943, 0.0437 for the validation dataset, respectively.