Implementing Sobol's Global Sensitivity Analysis to SFRC's Flexural Strength Predictive Equation

Steel fibers are essential for SFRC since they strengthen the material’s resistance to bending and cracking stresses and guarantee its endurance. However, the uncertainty of input parameters such as concrete compositions and steel fibers causes the stochasticity of flexural strength. The study uses big data to analyze the influence of fibers and other concrete compositions on the flexural strength properties of SFRC. For this purpose, the study focuses on developing predictive models for SFRC flexural properties based on a comprehensive database comprising two hundred and seven experimental results recorded by seventeen researchers. Bayesian Model Averaging is employed to identify significant components that influence the overall flexural strength and to develop a predictive flexural strength model. Monte Carlo simulation generates big data by utilizing the probability distribution of input variables and the predictive flexural strength model. The study used Sobol’s global sensitivity analysis method to assess various input parameters’ sensitivity to SFRC flexural strength based on the generated database. The impact order of input variables on the flexural strength is identified, as determined by the Sobol’ Indice.