Enhancing the shrinkage resistance of roller-compacted lean concrete via post-set induced damage: experimental analysis and mesostructure-based shrinkage prediction

This study investigates an approach to improving the shrinkage resistance of roller-compacted lean concrete (RCLC) through the controlled induction of damage in the post-set specimens. The RCLC mix design was developed using the Compressible Packing Model (CPM) and expectation function method. Post-set damage was then introduced using a vibratory compactor. The mechanical properties and shrinkage performance were evaluated accordingly. X-ray computed tomography (CT) was utilised to characterise mesostructural defects, which were quantitatively analysed using grey relational entropy analysis to determine their correlation with shrinkage performance. Furthermore, grey prediction models were developed to forecast shrinkage based on mesostructural parameters. The results showed that post-set induced damage had a negligible impact on strength and modulus, with healing rates exceeding 90%. Notably, specimens damaged after 2 days of curing with a 40% damage degree exhibited the most significant shrinkage reduction, decreasing drying and thermal strains by 23.86% and 31.22%, respectively. Among the mesostructural parameters, the fractal dimension was identified as the most influential factor affecting shrinkage. The GM(1,2) grey model demonstrated high accuracy in predicting shrinkage reduction based on this parameter. These findings offer both theoretical insight and methodological basis for enhancing the shrinkage resistance of RCLC base layers through targeted post-set damage strategies.