Prediction of Fracture Parameters of High Strength and Ultra-high Strength Concrete Beam using Gaussian Process Regression and Least Squares

This paper studies the applicability of Gaussian Process Regression (GPR) and Least Squares Support Vector Machines (LSSVM) to predict fracture parameters and failure load (P-max) of high strength and ultra-high strength concrete beams. Fracture characteristics include fracture energy (G(F)), critical stress intensity factor (K-IC) and critical crack tip opening displacement (CTODC) Mathematical models have been developed in the form of relation between several input variables such as beam dimensions, water cement ratio, compressive strength, split tensile strength, notch depth, modulus of elasticity and output fracture parameters. Four GPR and four LSSVM models have been developed using MATLAB software for training and prediction of fracture parameters. A total of 87 data sets (input-output pairs) are used, 61 of which are used to train the model and 26 are used to test the models. The data-sets used in this study are derived from experimental results. The developed models have also been compared with the Artificial Neural Networks (ANN), Support Vector Regression (SVR) and Multivariate Adaptive Regression Splines (MARS). From the overall study, it is observed that the concept of GPR and LSSVM can be successfully applied to predict fracture parameters of high strength and ultra high strength concrete.