Critical Buckling Load Evaluation of Functionally Graded Material Plate Using Gaussian Process Regression

Abstract

Functionally Graded Materials (FGM) is the advanced material that covers the advantages of both metal and ceramic which contain FGM. Due to the perfect combination, FGM plate is widely developed with the requirement for the practical application to avoid the difficulties from conventional approaches. Consequently, the study establishes the database from analytical model and uses it for developing a machine learning model in the desire of finding an alternative model to predict critical buckling load of the plate. Such a machine learning model is crucial for predicting critical buckling load of the FGM plate without complexity of analytical developments or finite element resources. The Gaussian Process Regression model has been developed based on the database containing 1000 labeled samples created from the analytical model. The Gaussian Process Regression models with and without optimization process are compared and the outstanding of the model with the optimization algorithm is revealed. The proposed model is validated on both train and test set with the R-square values larger than 0.99 and errors are significantly low indicating that the proposed model exhibits the required ability.