Intelligent optimization of particle size distribution in unscreened recycled coarse aggregates using 3D surface analysis

Abstract

The efficient measurement and optimization of the particle size distribution (PSD) of recycled coarse aggregates (RCA) is critical to ensuring consistent quality in high-performance concrete production. Unlike primary aggregates, which typically demonstrate minimal variability over extended periods and require only occasional testing, RCA often exhibit substantial fluctuations in quality over short timeframes. This variability necessitates a precise, automated, and real-time quality assessment approach, which is lacking in conventional aggregate processing. In this study, a rapid, automated, and non-contact 3D surface analysis method is proposed to assess and optimize the PSD of unscreened RCA during continuous transport on a conveyor belt. A custom-designed conical feeder and splitter facilitate the formation of continuous, symmetric triangular RCA piles, ranging from 4.0 to 16.0 mm in size. Representative PSD measurements are obtained by analyzing a designated strip located at one-third of the pile's height. High-resolution 3D point cloud data are processed using a watershed segmentation algorithm that leverages gradient-based path tracing for efficient topographical mapping. This enables parallel data processing, thereby reducing computational time. The proposed method enables real-time and accurate PSD analysis at industrial throughput levels (≥50 tons per hour) without interrupting conveyor operation, achieving a Root Mean Square Error (RMSE) between 4.69 % and 6.09 %. Furthermore, an optimization strategy based on cumulative percentage retained curves enhances RCA quality and supports continuous process control. The integration of these techniques contributes to improved RCA management and promotes sustainable resource utilization and waste reduction in the construction sector.