Heat tracing of embankment dam leakage: Laboratory experiments and 2D numerical modelling

Recently, the use of heat as a tracer to evaluate the process of leakage in embankment dams has attracted wide attention. A more accurate flow-heat coupling model of embankment dams could help us to better understand the patterns of water flow and heat transfer in embankment dams and provide a scientific basis for the seepage prevention and repair of these dams. In this paper, combined with thermal conductivity empirical models (TCEMs), the saturated-unsaturated flow-heat coupling model of embankment dams was established. Through laboratory sand tank experiments of concentrated leakage in embankment dams, the accuracy of the flow-heat coupling model under 10 types of TCEMs were tested and compared. The results show that the performance of the flow-heat coupling model varies under different types of TCEMs, and the Chung and Horton (1987) model shows better simulation effects, with a coefficient of determination (R-2), root mean square error (RMSE) and relative error (Re) ranging from 0.916 to 0.980, 0.266-0.467. and 1.370-2.442%, respectively. Therefore, this model could better reflect the dynamic temperature variations in embankment dams. Finally, the flow-heat coupling model was improved by modifying the COMSOL built-in equation, i.e. built-in COMSOL model was replaced by the Chung and Horton (1987) model, which further improved the accuracy of the flow-heat coupling model in the numerical simulation of seepage heat monitoring. Based on the improved model, the concentrated leakage of embankment dams under dynamic water levels was simulated numerically. Under the condition of a dynamic water level, the flow velocity and pressure at the leakage passage are positively correlated with the water level change, and the temperature field also shows the same change trend.

Automated summary

This paper establishes a saturated-unsaturated flow-heat coupling model of embankment dams and compares its accuracy under different thermal conductivity empirical models (TCEMs). The Chung and Horton (1987) model shows better simulation effects, and the flow-heat coupling model is further improved by modifying the COMSOL built-in equation. The improved model is used to numerically simulate the concentrated leakage of embankment dams under dynamic water levels, showing positive correlation between flow velocity, pressure, and water level change.