Effect of mix proportion on the uniaxial tensile properties of polypropylene fibre cementitious composites

Abstract

Numerous studies have proven that the incorporation of fibre-reinforced composites significantly increases the tensile strength of cement mortar. In this work, to further explore the influence laws of three factors, namely, the fly ash content, water-binder ratio, and sand-binder ratio, on the ultimate stress of polypropylene fibre cement-based composites (PP-FRCC) under uniaxial tension, a full-factorial design of experiments (DOE) was adopted to analyse the data. Each of the three selected factors had three levels and twelve repetitions. Analysis of variance (ANOVA) was used to determine the statistically significant effects among the overall means and whether there were interactions between the factors, establish the optimal mix proportion that best met the test requirements, and develop a prediction model for the tensile strength of PP fibre-reinforced, cement-based composites. The research results show that a change in the sand-binder ratio has a greater effect on the ultimate stress than does the fly ash content and that the fly ash content has a greater effect than does the water-binder ratio. However, all three factors have extremely significant effects on the ultimate stress. The interactions between the fly ash content and the water-binder ratio and between the fly ash content and the sand-binder ratio are extremely significant, whereas the interaction between the water-binder ratio and the sand-binder ratio is significantly influential. The sand-binder ratio is positively correlated with the ultimate stress of the composites, whereas the fly ash content and water-binder ratio are negatively correlated with the ultimate stress of the composites. The optimal tensile strength is obtained when the fly ash content is 0.35, the water-binder ratio is 0.27, and the sand-binder ratio is 0.84. The root mean square error (RMSE) of the radial basis function neural network (RBFNN) test set is 0.24197, and the coefficient of determination (R2) reaches 0.90, indicating that the model has strong explanatory power. This method is feasible, and its prediction accuracy meets engineering requirements, providing a new approach for investigating the strength of hybrid fibre-reinforced concrete in engineering.