Algorithms of Flood Water Flow Control in Small River Floodplain Channels

The problem of counteracting anthropogenic degradation of modern river floodplains through the construction of dams in river channels is an urgent topic of research worldwide. This paper presents a comparative analysis of the accuracy and efficiency of three algorithms for controlling flood water flows in small channels in the northern part of the Volga-Akhtuba floodplain in order to increase the flooded target area. A two-level dynamic programming algorithm, an algorithm for optimizing flows in a network, and the original two-level and one-level heuristic donor-acceptor algorithms are investigated. Water dynamics calculations are based on shallow water equations. The calculations use the original EcoGIS-Simulation software package for numerical modeling of surface water dynamics with a digital elevation model of the northern part of the Volga-Akhtuba floodplain. The computational core of EcoGIS-Simulation is based on the numerical algorithm CSPH-TVD, which combines Lagrangian (Smoothed-ParticleHydrodynamics) and Eulerian (TotalVariationDiminishing) approaches. Parallel computations are performed on NVIDIATesla GPUs (graphics processing units). The conducted study showed in general low accuracy and efficiency of the investigated algorithms for water flow control. The relative error of all algorithms significantly exceeded the relative error of geoinformation and hydrodynamic modeling. The reason for this is the insufficiently high adequacy of the models for controlling unsteady water flows in a channel system with unstable structure. The heuristic one-level algorithm “donors-acceptors”, using the simplest control model, showed higher accuracy and efficiency compared to other algorithms.