Water leakage through the karst terrain under the Višegrad dam on the Drina river was observed already in the first years of its exploitation. Reducing losses and preventing further erosion and collapse of the dam required urgent remediation. The goal of this capital project, financed by the World Bank, was to fill the underground cavities with granulated stone and concrete.
To support the design phase of the project, we have developed a software platform able to estimate the configuration of underground network of fissures based on various in-situ measurements. During the remediation process, it was evaluating the effects of the implemented technological processes in near real-time, providing decision-makers with a powerful tool for planning future activities.
The platform relies on numerical models able to calculate piezometric levels, flow rates, and tracer concentrations for an assumed system configuration under given boundary conditions. The data obtained from the measurement system were periodically in an automated manner assimilated into numerical model, giving the model states that best comply with the reality.
This complex multi-model consists of three models: (1) the hydraulic model of karst groundwater flow; (2) the model of solute transport and (3) the model of granular material transport and sedimentation. Parameters of the modeled conduits were estimated using multi-objective evolutionary algorithms on distributed computing environment. Driven by the platform, the remediation project reduced the leakage by five times.
