ABSTRACT
A five km long road section, a bypass in the vicinity of Gislaved in southern Sweden, was opened for traffic in 2013. The pavement design was based on performance specifications, primarily for rutting and roughness. There were also safety aspects like specified friction criteria. The structural rutting was related to a subgrade strain criterion, and a classic fatigue criterion for asphalt concrete. In addition to quality control/quality assurance, the construction was also monitored by Falling Weight Deflectometer (FWD) testing on the subgrade, subbase, and bound base layers. After the construction, FWD testing was performed on a yearly basis. Surface characteristics data were also sampled at intervals according to the contract. For the first five years the pavement deterioration proved to be normal and followed the theoretical expected values for a mechanistic design process. Then, some sections showed an accelerated deterioration, with cracks appearing in the wheel paths. Thorough investigations followed to disclose the reason for the premature failures. Possible reasons sought were unexpected overloads, climate change, and underground water movements. While coring the road, it was observed that the cooling water would not drain through the unbound base material. Thus, the top millimeter or so consisted of very fine material with low permeability, almost like clay. Obviously, traffic had acted on the material as to crush and wear it down into smaller fractions. The further mineral investigation also showed a high mica content. The present paper discusses how this process could increase the strain on the pavement, but also opens to debate whether more mechanistic procedures should be added to the Mechanistic-Empirical (ME) model. The mica content and other specifications are usually based on empirical data. With higher legal loads and load factors on trucks, the empirical data may not suffice for predicting pavement performance.
