8. Conclusions
The following conclusions are derived based on the numerical and experimental studies conducted in this work. • The experimental results of the compaction of pervious concrete produced with spherical aggregates showed that the influence of the P/ A over the compaction curve is negligible. The application of the numerical models proposed here explained this observation, suggesting that the small variations induced by the P/A are eclipsed by the high variability of the results. • The experimental study of the compaction of pervious concrete in fresh state produced with crushed limestone in two different grading (5–12 mm and 9–20 mm) showed that lower pressure was necessary to compact the specimens with bigger size of particles. This phenomenon is related to the number of contacts between aggregates, which is inversely proportional to the size of particles. The same trend was also observed in the new models developed here to simulate the compaction of pervious concrete. • Although the Kelvin-Voigt rheological model is generally used to represent the interaction between aggregates surrounded by cement paste in fresh conventional concrete, it might lead to important inconsistencies when the contact between particles of pervious concrete is simulated. To overcome these inconsistencies, the constitutive law proposed and validated here should be used. • The new constitutive law reproduced the normal behavior of the contact between two identical spheres with a visco-elastic material in the middle (experiments conducted by Shyshko et al. [21]) with accuracy. The comparison of the numerical and experimental results for equivalent displacements showed a correlation coefficient (R2 ) of 0.975, 0.97, 0.936, 0.947 and 0.935 for the velocities of approximation of 200, 100, 50, 20 and 10 mm/min, respectively.
