Adopting an image analysis method to study the influence of segregation on the compressive strength of lightweight aggregate concretes

Aggregate segregation in concretes, especially in lightweight aggregate concretes (LWAC), is a pathology that can seriously impact on different material properties, both in terms of mechanical strength and durability. Some studies have already analysed the impact of this phenomenon in the compressive strength of LWAC , almost always adopting segregation quantification methods based on the comparison of variables such as weight or density measured in sections extracted from cylindrical samples. Although the methods used so far are simple and easy to apply, in certain circumstances they may not offer sufficient accuracy to measure the phenomenon. Moreover, the analysis of segregation in LWAC, which due to the lower density of aggregates in relation to the mortar matrix occurs in the opposite direction to the segregation of conventional concretes (from bottom to top), has not been widely studied so far. The analysis of the cross sections of cylindrical concrete samples through image analysis techniques has been increasingly used to estimate parameters related to the distribution of materials inside them. This study aims to measure the impact of segregation on the compressive strength of LWAC cylindrical specimens, employing an image analysis technique capable of quantifying the phenomenon with a more accurate segregation index. During the experimental campaign 22 scenarios were defined, combining different types of aggregate, vibration methods and vibration times, to achieve different levels of segregation. Among the main results it can be noted that the impact of the type of moulding (one layer vs. two layers) considerably affected the mechanical properties of concrete, with concretes vibrated in two layers showing a maximum reduction in compressive strength due to segregation of 9%, while concretes vibrated in only one layer showed reductions ranging from 63% to 118%. Analyzing the type of rupture of the samples, the study identifies the appearance of well-defined compression cones in samples where segregation is excessive, with 90% of the cases presenting this type of rupture and that the size of the compression cone increases as the effect of segregation is accentuated.This research was funded by the University of Alicante ((GRE13-03) and (VIGROB-256))