Journal of Hydraulic Engineering

The turbulent‐flow field above dunes is predicted with equations describing the transport of the kinetic energy of turbulence (k) its rate of dissipation (ε), and algebraic relations derived from a second‐moment turbulence closure (Algebraic Stress Model). The resulting set of expressions jointly with the identified boundary conditions are solved with a computer code based on a finite‐difference solution that uses the PISO algorithm to handle the coupling between the continuity and momentum equations to obtain mean velocity and turbulent‐stress profiles in the flow field, turbulent kinetic energy, and pressure and shear stress distributions over the dune surface. Estimates of flow resistance are obtained by integrating the pressure and shear stress distributions acting on the dune surface. Predictions compared well with detailed data of experiments on turbulent open‐channel flow over dunes and with experimental data of total flow resistance.

The exact linear solutions for the sediment transport and bed form evolution for one‐dimensional sediment‐water two‐phase motion have been obtained using the St. Venant shallow‐water equations with the assumption of quasisteady flow. These solutions are applicable to alluvial channels of infinite length, initially at equilibrium followed by an arbitrary forcing function of either sediment transport or bed elevation imposed as an upstream boundary condition. The solutions have been used to predict aggradation in a channel due to constant overloading. Comparison of the results with available experimental data and with the solution obtained from a parabolic model is satisfactory. The present theory is significant conceptually since it provides valuable insight into the physical phenomenon as well as into the mathematical behavior of the solutions.

Hydrographs and flood‐frequency distributions are often required for hydrologic and hydraulic analysis of small ungauged basins. Due to the lack of runoff data, a numerical model to relate rainfall with runoff cannot be used without appropriate calibration. Thus, it would be extremely useful to obtain critical elements of hydrograph and flood‐peak distribution for small ungauged basins from rainfall and basin characteristics. In this study, simple expressions are developed to express the time of concentration, time to peak, and peak flows as functions of basin and storm characteristics for small basins, based on results obtained from a specially derived kinematic wave rainfall‐runoff model. Following the approach presented in several classic papers by Eagleson, the distribution of flood peaks is derived. These flood‐peak distributions are based on the assumptions that the arrival of storm events can be modeled by a Poisson process, and that the relationship (as derived from this study) expressing peak flow as a function of rainfall and basin characteristics is valid. The derived flood‐frequency distributions are applied to several contrived basins with various storm and basin characteristics. The sensitivities of these parameters on flood‐frequency distribution are also investigated.

The primary objective of this study is to improve understanding of the mechanism causing scour in unidirectional current. Experiments have shown how local scour develops around submarine pipelines in noncohesive sediments. The study shows that piping is the dominant cause of the initiation of scour. Piping and the stagnation eddy combine to undermine the pipeline, and mark the onset of scour. The critical hydraulic gradient associated with the initiation of scour is equal to the flotation gradient of the bed sediment. The pressure drop between the stagnation pressure upstream and wake pressure downstream of the pipe induces this hydraulic gradient. When a pipe is just embedded, the onset of scour does not occur if the ratio of the flow depth to pipe diameter exceeds 3.5. Similarly, the onset of scour does not occur for half‐buried pipes. The reduction in pressure gradient across the pipeline for these flow/pipe combinations accounts for the lack of scour. The onset of scour can be prevented by placing an impermeable membrane underneath the pipeline.

Friction factors of armored river beds at the discharge that forms the armor coat are investigated. The analysis shows that the friction factor is rather independent of grain‐size distribution of the material forming the bed and the maximum grain size of this material. The major controlling parameter is the slope of the energy grade line. These unexpected results are explained by the coupling of the friction factor with the incipient motion problem and the rearrangement of the grains of the coarsest fraction in the armor coat. The difference between rounded and angular material is discussed. Field data collected in irrigation canals support the results obtained from laboratory experiments.

A numerical model for simulation of temporal bed evolution in channel bends with fixed walls is constructed and verified with laboratory experiments. The model is based on governing equations of flow and sediment motion, with some simplifying assumptions. Numerical experiments are conducted, using the model to investigate influences of various factors on bed evolution. It is found that the process of bed evolution in channel bends with steady discharge and uniform sediment can be approximated by a complementary exponential decay function of a time parameter. A regression equation is established for the coefficient as a function of flow condition, bend geometry, and sediment size. These relations, the numerical model, or both, may be applied to evaluate evolution of bed topography in river bends. Numerical results are also verified with laboratory experiments.

Velocity and turbulence distributions are measured in a 270° curved open channel of half trapezoidal section with extra roughness applied to the outside bank slope. The purpose of this investigation is to provide some understanding of flow mechanics that may be useful in the design of riprap protection. Profile‐averaged velocity and bed shear velocity distributions are calculated. A significant variation of bed shear up the slope and a much higher stress near the toe of the slope are found for the roughened outside bank slope. It is found that a reasonable correlation of the velocity and bed shear stress on the bank is given by a form of the Colebrook equation. Lateral momentum transfer by secondary flow and turbulent stress are also evaluated and found to be significant in the bank vicinity.

Available least squares parameter estimation techniques for Muskingum flood routing are critically examined and compared. Explicit expressions for the Muskingum routing parameters are provided for two of those procedures. The optimality of the described methods is studied and their consistency with the use of the Muskingum method for forecasting purposes is discussed.

Flow on stepped spillways can be divided into nappe and skimming flow regimes. This technical note presents an approximate method of predicting the characteristics of skimming flow on stepped spillways. Expressing the turbulent shear stress between the skimming stream and the recirculating fluid, trapped by the stream between the steps, in terms of a fluid friction coefficient and dynamic pressure (in terms of the mean velocity), a flow equation has been developed. Using the model studies of Sorensen on the new Monksville dam, this friction coefficient was found to be equal to about 0.18. An estimate has been made of the energy loss caused by the steps and it was found that this loss is significant. It was also found that the skimming flow sets in when the critical depth is greater than about 0.8 times the height of the steps.

In the course of modeling hydrodynamic phenomena in a limited area domain, one must specify open boundary conditions. Various forms of the Sommerfeld radiation condition have been used on the boundary of the domain, in the past, to permit internal disturbances to pass unhindered through the boundary. This article shows that if one tries to compute the phase speed associated with the Sommerfeld condition, inevitable numerical errors in this computation adversely affect the interior solutions. A recursive filter that results in smooth, accurate phase speeds and therefore better interior solutions is proposed and its utility demonstrated.