The goal of the French project GELCRO(1)(2)(4) is to develop a model which will allow us to forecast the surface state of the road before a snowfall and the evolution of a snow layer deposited on the pavement. This system will improve efficiency of road winter maintenance service.
The model named Gelcro is built from two one-dimensional coupled numerical models :
- Crocus is a physically-based model originally developed for operational avalanche risk forecasting. It simulates snow cover evolution.
- Gel1D calculates frost propagation in the pavement.
Physical properties of the interface between snow and pavement have to be determined and then introduced into the model.
For this study, an experimental site was chosen owing to its important snowfall frequency and the variety of meteorological situations. This site is equipped with six experimental pavements representative of present French techniques. These pavements are fitted out with different sensors (snowheight and temperatures sensors, video camera) which allow us to observe pavement snowing up and to characterize snow evolution for different meteorological situations. This site is also equipped with a complete weather station including specific snowcover measurements. So a data base is built to determine snow/road interface properties and to validate the model Gelcro. Manual measurements are also performed while or just before each snowfall and composite samples of snow and pavement are taken. To preserve snow microstructure on the pavement during sampling, the pore space is filled with diethyl-phtalate (melting point -5°C) and the specimen is frozen. An original technique has been developed in cold laboratory (at -20°C) to observe snow/pavement interfaces in these samples. This technique includes : dry-cutting simultaneous snow and pavement to obtain a vertical section and treating the surface to enhance contrast between snow, phtalate and pavement for photomicrography. So we observe at the microscale a section plane of the snow/pavement bond structure. Images are then used in a two-dimensional thermal conduction model in order to calculate the interface thermal resistance. This thermal resistance, related to snow and pavement features, will be introduced in the model Gelcro.
Direct microscale observation of a snow-road interface makes possible the design of a realistic modeling of both thermal bridges and capillary effects at the scale of the interface