Example of the Pollutant Transport Modelling in the Pavement and Embankment

The transport model of pollutant leaching from the secondary road construction material was developed by the Environmental Research Group from the University of New Hampshire, USA (Apul et al., 2003). Water flow in a Minnesota highway embankment was modelled in one dimension for several rain events and calibrated to the field condition (Fig. 11.20). The test facility consists of 40 and 152m-long

hot mix asphalt and Portland cement concrete test sections with varying structural designs. Each test section is instrumented to monitor strength and hydraulic proper­ties. The hydraulic properties of the embankment were predicted from water content measurements made in the embankment, a Portland cement concrete pavement with an asphalt shoulder. The hypothetical leaching of Cadmium from coal fly ash was probabilistically simulated in a scenario where the top 0.50 m of the embankment was replaced by coal fly ash. The groundwater table was set at 1.9 m below ground level (b. g.l.), which is within the range (1.3-4.6m b. g.l.) observed at test site. An entire year’s precipitation data repeated 10 times was input as the variable flux boundary condition. The molecular diffusion coefficient of Cadmium in free water

was input in the model as a constant (6.2 x 10-5 m2/day) and tortuosity factor was calculated within the finite element code, HYDRUS2D, as a function of the water content.

The probability distributions of unsaturated hydraulic properties of the em­bankment were determined from parameter posterior probabilities obtained from embankment infiltration simulations. The probability distributions were used to fit to the four parameters of the van Genuchten SWCC model (see Chapter 2, Section 2.7.3). Weighted moment equations were applied to calculate the means and standard deviations for the normal distributions. Saturated permeability and saturated water content were assigned joint log-normal distributions. To account for the variability of partition coefficient, kd, uniform distributions were assigned. In the study kd was considered as lumped parameter. The temporal and spatial variability of kd that would be expected in the field was incorporated in the modelling approach by probabilistically varying kd values of the subgrade and the coal fly ash for each simulation.

The average percentage of initial available mass leached after 10 years, as ob­served 0.01 m below ash, is presented, idealised, in Fig. 11.21, for a point a short way into the ash layer (point marked on Fig. 11.20). No significant Cadmium fluxes were observed 0.25 m below the coal fly ash or at the groundwater table depth. After 10 years, the fraction of initial available mass leached was 5 x 10-6 percent at 0.25 m below the coal fly ash, and 0% at the groundwater table depth (at the 90th percentile of uncertainty). The cumulative release at the 90th percentile of uncertainty and the appropriate probability distributions, was 2.65 x 10-3 mg Cd/kg ash after 10 years. The mean of the release estimate was 1.15 x 10-3 mg Cd/kg ash. Further details

are available from Apul et al. (2005) but it is clear that infinitesimal leaching occurs in real pavement/earthworks sections arranged in a similar way to the construction studied in Minnesota.

11.5 Conclusions

The partial differential equations that govern solid mechanics, water transfers, heat transfers and pollutant transfers have been restated. The specificities of the finite element method when dedicated to such non-linear phenomena and their coupling have been summarised. Then a number of numerical simulations have been pre­sented. They cover moisture transfers, freezing, mechanical strains and permanent deformations. It appears that, for the most part, realistic numerical modelling is today available, at least for advanced research teams. But progress is still needed, for example to couple changes of moisture level and changes of the mechanical behaviour.