The written papers from the conference were published in the language in which they were submitted to TAC.
Estimating segregation potential (SP) is one of the challenging problems in pavement design in cold regions. This parameter can be quantified directly in the laboratory, in-situ measurements or indirectly using the simplified literature models. Several models have been proposed in the recent years, mainly based on laboratory measurements, to assess this parameter. Nevertheless, despite these efforts, these models still present some shortcomings and discrepancies with the in-situ measurements. A research project was therefore carried out at the soils and aggregates laboratory of the Quebec Ministry of Transportation and Sustainable Mobility to analyze the laboratory measurements of SP and compare them with estimates using the simplified literature models. The analyses revealed that the main parameters controlling the SP are the grain size distribution, the specific surface of the 400 μm particles and the mean particle diameter of the fine fraction (<80 μm). Although these parameters are included in some simplified SP estimation literature models, the comparison with laboratory measurements shows that these models present significant inaccuracy and often overestimate of the SP. Moreover, the back-calculation of the SP using the in-situ measurement of the frost heave confirmed that the literature models overestimate the SP and that the model proposed in this study seems to be more representative.