Experimental and Numerical Investigation of the Precessing Helical Vortex in a Conical Diffuser, With Rotor–Stator Interaction

Abstract

The flow unsteadiness generated in a swirl apparatus is investigated experimentally and numerically. The swirl apparatus has two parts: a swirl generator and a test section. The swirl generator which includes two blade rows, one stationary and one rotating, is designed such that the emanating flow at free runner rotational speed resembles that of a Francis hydroturbine operated at partial discharge. The test section consists of a conical diffuser similar to the draft tube cone of a Francis turbine. Several swirling flow regimes are produced, and the laser Doppler anemometry (LDA) measurements are performed along three survey axes in the test section for different runner rotational speeds (400–920 rpm), with a constant flow rate, 30 l/s. The measured mean velocity components and its fluctuating parts are used to validate the results of unsteady numerical simulations, conducted using the foam-extend-3.0 CFD code. Furthermore, phase-averaged pressure measured at two positions in the draft tube is compared with those of numerical simulations. A dynamic mesh is used together with the sliding general grid interfaces (GGIs) to mimic the effect of the rotating runner. The delayed detached-eddy simulation method, conjugated with the Spalart–Allmaras turbulence model (DDES–SA), is applied to achieve a deep insight about the ability of this advanced modeling technique and the physics of the flow. The RNG k−ε model is also used to represent state-of-the-art of industrial turbulence modeling. Both models predict the mean velocity reasonably well while DDES–SA presents more realistic flow features at the highest and lowest rotational speeds. The highest level of turbulence occurs at the highest and lowest rotational speeds which DDES–SA is able to predict well in the conical diffuser. The special shape of the blade plays more prominent role at lower rotational speeds and creates coherent structures with opposite sign of vorticity. The vortex rope is captured by both turbulence models while DDES–SA presents more realistic one at higher rotational speeds.