This study predicts the mechanical properties of cement mortars. It focuses specifically on cement strength classes of 32.5, 42.5, and 52.5 MPa, respectively. To this end, a total of 270 specimens using 54 mix designs were constructed and subjected to freezing and thawing cycles (FTCs) at −18 and 18 °C. Then, mechanical properties such as compressive and flexural strength and porosity of the specimens after 0, 50, 100, 150, and 200 cycles were obtained. Afterward, genetic expression programming (GEP) was used to predict the obtained results and to represent nonlinear relationships between the mechanical properties, freezing and thawing cycle, and cement strength class. The experimental data was used for training and testing of two GEP approach models. The results show that the strength of mortar reduces with increase in FTC. On other hand, a close correlation was observed between the predicted and experimental values when cement strength class and number of cycle were considered as independent input parameters. This fact indicates the significance of the aforementioned parameters along with the other parameters such as water to cement ratio, sand to cement ratio, and weight of cement on the accuracy of the prediction.

Keywords