Effect of porosity on predicting compressive and flexural strength of cement mortar containing micro and nano-silica by multi-objective ANN modeling

[Display omitted] •Ms and Ns reduce porosity and increase the compressive and flexural strength.•Compressive and flexural strength could be predicted accurately by multi-objective ANN model.•Experimental and MOANN predicted results have a good convergence.•The performance of ANN models improved by considering porosity as an input parameter. This study describes the effect of various porosity values on the compressive and flexural strength of cement mortar hardened specimens containing micro silica (Ms) and nano silica (Ns) in the two phases of experiment and the prediction modeling. For this purpose, an extensive experimental program containing 32 different mix designs with 960 specimens, two different water to binder ratios (0.50 and 0.40) and various percentages of Ms and Ns contents (1.4, 2.8 and 4.2 wt%) by weight of cement has been performed. In order to investigate the effects of Ms and Ns contents on the microstructure of hardened cement mortar specimens and, consequently, on the mechanical properties of cement mortar, a Field Emission Scanning Electron Microscopy (FE-SEM) analysis was employed. Furthermore, proposing a multi-objective ANN model (MOANN model) along with the correct selecting of the input parameter are the two essential factors for achieving the mixture with the desired properties. In order to determine the sensitivity of the models in selecting the effective and correct input parameter for reliable and accurate predictions, a parallel examination was performed on the status of ANN and MOANN models, both with and without the presence of the porosity values as the input parameter. The experimental results indicated that the simultaneous presence of Ns and Ms contents in the cement mixture increases the compressive and flexural strength in addition to the decreasing the porosity values. The modeling results show that the MOANN-II model has a higher accuracy in prediction due to the simultaneous prediction of the compressive and flexural strength of cement mortar where the presence of the porosity as an effective and proper input parameter was considered. Finally, the proposed model was validated with the help of the collections' dataset of previous studies.