Abstract

Flow separation is a common phenomenon in decelerated subcritical flows as in open channel expansions. A highly distorted velocity and shear stress distribution due to flow separation can lead to a continuous reduction of energy and trigger an adverse pressure gradient resulting in flow separation. This causes loss of energy and hydraulic efficiency of the systems. An experimental investigation was conducted with the use of a gradual rising hump on the bed of an expansion in a rectangular open channel. Besides the hump, split vanes in the flow field were also used to reduce the expansion angle and in turn reduce the adverse effect of flow separation. These modifications resulted in a relatively more uniform velocity and shear stress distribution in the transition and in the channel downstream of the expansion. A laboratory model of rectangular open channel transition expanding was constructed with Plexiglas plates. It facilitated the measurement of the flow velocity and turbulence characteristics with the aid of Laser Doppler Anemometer (LDA). The total divergent angle of the transition was 19.78 degrees. Velocities were measured along the x, y and z directions, positioning the LDA from both the bottom and the side of the channel. Two humps with gradual linear rises of 12.5 mm and 25 mm were used. A second device included the use of a single vane and a three vane splitter plates system formed with thin Plexiglas plates. Mainly velocity distributions, with and without humps and the splitter vanes were the results sought. The variations of energy and momentum coefficients were analyzed to find the effectiveness of the devices used in the transition to control flow separation. As a small addition to the study, the use of computational fluid dynamics (CFD) to predict the flow characteristics of open channel was also undertaken. Due to their lower time demand and lower cost, these numerical methods are preferred to experimental methods after they are properly validated. In the present study, the CFD solution is validated by experimental results. A limited number of CFD simulations were completed using the commercial Software ANSYS-CFX. In particular, mean velocity distributions for the rectangular open channel transitions were used for model validation. To this end, the three-dimensional Reynolds-Averaged Navier-Stokes (RANS) equations and the two equations k-[varepsilon] models were used. The validation of the model using test data was reasonable.