In-situ density measurement using GPR considering random pore distribution of asphalt pavement

In-situ density measurement is vital for compaction quality control and assessment of asphalt pavement. This can be achieved using ground-penetrating radar (GPR) based on electromagnetic mixing-based density prediction models, which assume a constant dielectric property of the asphalt layer. This neglects the spatial distribution variations of particles within asphalt mixtures, leading to inaccurate dielectric constant and density predictions. This research proposed a modified density prediction model, incorporating a random pore distribution coefficient to consider material vertical heterogeneity for the porous asphalt (PA), stone mastic asphalt (SMA), and asphalt concrete (AC) mixture. A meso-scale electromagnetic (EM) modelling approach and non-linear least square curve fitting was utilized to optimize the proposed coefficient for each mix. Influences of air void content and antenna height were examined. Controlled laboratory tests confirmed the proposed model's superior precision, outperforming the conventional Al-Qadi-Lahouar-Leng model by 6.69% for PA, 2.70% for SMA, and 1.93% for AC mixtures, respectively.