Dam Break Hydraulics in Natural Rivers

Abstract

An investigation is carried out for numerical computations of dam break hydraulics in Natural Rivers with Explicit Finite Difference (EFD) Schemes. Although computational Dam Break Hydraulics is a topic of interest over more than 100 years, numerical simulations of dam break flow in relatively simple channels are found more often compared to real river flood simulations. Natural River channel with wide floodplains make the computation cumbersome as Natural River channels are highly non-prismatic with significant variations in River bed slope and friction. In this study, first order Diffusive Scheme, second order modified two-step Predictor Corrector scheme and Total Variations Diminishing (TVD) McCormack Predictor Corrector using Venn Leer Flux limiter are analyzed by solving unsteady flow equations in conservative and non conservative forms for simulating a hypothetical dam break situation in a Himalayan River in India . The height of the dam is 245 metres. Upstream length of the channel is 40,000 m and downstream is 64,000 m, the elevation of the bed of the channel changes from 545 m to 126.95 m, change in the channel width ranges from 300 m to 5650 m and Manning’s roughness coefficient varies from 0.03 to 0.035 depending on the channel characteristic. The numerical solutions of the EFD are relatively tested for their performances giving special emphasis to some parameters such as the flood depth at different sections and inundated area at different time periods after the failure of the dam and travel time of the flood waves; which are most important for the practicing engineers for efficient flood management. The computational aspects of the numerical models i.e. the implementation efforts of the schemes, run time required are compared. The applicability of the complex numerical corrections such as TVD in the numerical model for refinement in the solutions of dam break hydraulics in Natural Rivers is also examined. The stability and accuracy of the numerical solutions for both conservative and non conservative formulations are analyzed, and it is observed that numerical simulation of flows for complex real River topography in conservative forms with simple EFD schemes are advantageous from practical point of view compared to the higher order Explicit or Implicit Finite Difference or Finite Element schemes which increase the run time significantly and make the implementation complex.