Automated control of asphalt concrete mix distribution considering road base parameters

Abstract

Automatic systems of asphalt pavers maintain the prescribed elevations and cross slope relying on a stringline, a laser beam or a tracking ski, which leaves the thickness of the laid course as a quantity derived from the relief of the base. The bearing capacity of the finished structure, however, results from the combined behaviour of the asphalt package and the underlying courses, whose stiffness varies along the working section on a scale comparable with the useful range of the control itself. The aim of the study is to formulate and verify by calculation a law of asphalt concrete mix distribution in which the course thickness and the subsequent rolling regime are assigned according to the measured modulus of elasticity of the base. The calculation follows Odemark's method of equivalent thicknesses in Ullidtz's formulation for a two-layer asphalt package resting on a base of variable stiffness, supplemented by an author's model of compaction kinetics and an estimate of the temperature window for rolling; three distribution regimes are compared on a model section 1000 m long divided into twenty calculation segments with base moduli of 82,3-172,4 MPa. With a constant thickness of 70 mm and eight roller passes, the required overall modulus of 200 MPa is not reached on seven segments out of twenty, and the residual air void content exceeds the upper limit of 7,0% on three of them. Adaptive assignment of thickness within 50-90 mm and of the number of passes within 8-10 reduces the number of deficient segments to two, removes the void content excess completely and lowers the coefficient of variation of the overall modulus from 16,43 to 8,86%, while mix consumption falls by 7,14%. A lower applicability threshold of thickness compensation is established: at a base modulus below 86,6 MPa the maximum thickness permitted by the standard no longer brings the structure to the required modulus. Overestimating the measured base modulus by 15% raises the number of deficient segments from two to ten, whereas underestimation by the same amount does not worsen the outcome. The course thickness acts as a range-limited compensator of base heterogeneity, and the reliability of the initial stiffness estimate enters the final result asymmetrically, calling for a conservative bias of that estimate.