Investigation of the structural response of pavements in cold region using instrumented test site data

Sammanfattning: The structural behaviour of pavement structures is known to be affected by the traffic-related loading and by the ambient factors to which the structure is subjected. A new mechanistic-empirical (M-E) pavement design method is under development in Sweden with the main purpose to adequately predict pavement structural response and performance. An M-E design method for a flexible pavement means application of the principles of engineering mechanics to evaluate the response of pavement structure to traffic loading and much improved design methods to carry out distress prediction or how performance changes with time. This would ensure a fundamental understanding of how the pavement structure responds to a certain action or loading conditions. The mechanistic-empirical approach is more flexible as it is able to adapt to new situations such as new pavement materials and loading situations. It is important to take into account the real loading and climatic conditions and predict the resulting changes in material properties and structural behaviour at the time of loading as well as in the long-term. New models are therefore required for the further development of pavement design method, and it needs to be validated through reliable data obtained through realistic measurements.In this licentiate thesis, the effects of the environmental factors and loading by heavy vehicles in pavements are investigated. The results of the study are based on environmental data from multiple locations in Sweden and on measurements from two instrumented road sections located near the village of Långträsk in the northern part of Sweden. Both roads consist of thin flexible pavements, the behaviour of which is highly dependent on the variation of the temperature of the asphalt layer, the moisture content in the unbound granular layers, and frost depth conditions.The licentiate report consists of three scientific publications. Paper 1 presents a country-specific case study in which the frost penetration depth in various Swedish roads is predicted by a statistically derived empirical model that uses the air freezing index calculated from the air temperature as an input. The model correlation is based on meteorological data from 44 meteorological stations and pavement cross-sectional temperature distribution data from 49 road weather information system (RWIS) stations over all five climatic zones throughout all of Sweden.Paper 2 focuses on the response of an instrumented test section subjected to loading by falling weight deflectometer (FWD) and heavy vehicles. The mechanical response instrumentation consisted of asphalt strain gauges (ASG), strain measuring units (εMU), and soil pressure cells (SPC) installed at different locations in the structure. The layer stiffness values were obtained via backcalculation based on the FWD surface deflections bowls. The recorded values of the mechanical response were compared against calculated values by multilayer elastic theory (MLET) based software. Three different heavy vehicles weighing from ~64 tons to ~74 tons were compared in terms of damage caused to the pavement structure. It was found that if the number of axles was increased and dual tyres were used, longer heavier vehicles were not more destructive to the pavement structure than shorter vehicles with fewer axles and higher axial load and tyre pressure.In paper 3, the effect of the seasonal variation of the environmental factors on the behaviour of an instrumented road test section was investigated. The same loading configuration described in paper 2 was used on four different measurement campaigns in different seasons over the span of 1 year. The environmental variables were monitored throughout the year by asphalt thermocouples, a frost rod, and time-domain reflectometer (TDR) probes. The mechanical response sensors and the environmental sensors were found to be a reliable data collection method throughout the entire year. By comparing the recorded response values to the MLET calculated values, it was shown that it is possible to model the mechanical behaviour of pavement structures using linear-elastic MLET if the temperature variations in the asphalt layer and the moisture variations in the granular layers are taken into account.