Numerical and experimental study of ducted turbines in bi-directional tidal flows

Abstract

Installing a turbine in a duct generally will increase the flow through the rotor. If there is flow confinement due to limited water depth, the presence of lateral boundaries, or the presence of adjacent turbines, a further flow enhancement through the rotor will be achieved. It appears there are no or limited studies on ducted turbines in a tidal farm. In recent times, a duct equipped with a flange (referred to as a wind-lens) has attracted extensive attention. Unlike the traditional early designs of ducted turbines that rely on increasing the area ratio between the inlet and duct exit diameters, it depends on creating a low backpressure region through a strong vortex formation behind its broad flange and hence, a substantial increase in the wind speed. Although wind-lens has demonstrated significant power augmentation, a vast majority of investigations on wind-lens designs are mainly based on several existing typical shapes. Therefore, the wind-lens design is still far from optimal. In addition, the power performance of ducted turbines in a tidal farm, as well as the interaction effects between adjacent ducted turbines, have not been investigated in detail before. The present study carried out a thorough analysis of the hydrodynamics of different flanged duct shapes (symmetric and asymmetric) for both uni-directional flow and bi-directional flows and hence proposed an optimum wind-lens design. The analysis was performed using data obtained from numerical and experimental modelling of flanged duct tidal turbines which were isolated or within an array. The numerical assessment employed an actuator disk representation of the rotor that provides an axial resistance to the flow, and thus models the extraction of linear momentum. And based on the proposed optimum flanged duct type (i.e., duct Type C) obtained from the numerical simulation, physical experiments were carried out in an open-channel recirculating water flume with a 1:200th scaled flanged duct tidal stream model representing a turbine of diameter 20m in a water depth of 44.4m. Raw data sets were obtained using a Prony brake and a side-looking ADV. Analysis was presented in terms of power performance, wake velocity ratios and turbulence intensities. Conclusions were drawn from the comparisons of the power performance and interaction effects between flanged duct tidal turbines or the bare (non-duct) turbine in isolation or an array. The numerical results showed that a flanged duct turbine with an inlet-arc flap and a curved flange can achieve a maximum value of CP∗ (duct area-based power coefficient) close to the Betz-Joukowsky limit (BJL) for flow from either direction while the symmetric duct turbine showed a maximum CP∗ of about 60% of the BJL. The effect of flow confinement on CP∗ showed CP∗ increases with blockage ratio, ϵ and the ratio of the peak power coefficient of ducted turbine over the peak power coefficient of bare turbine varies slightly with the blockage ratio. The experimental results showed that ducted turbines generate higher power per rotor unit area instead of duct area. The power coefficient obtained for the flanged duct turbines was within the range of a 200 kW OpenHydro demonstrator device i.e., about 0.3 with a 2.5m/s rated flow speed. Results were qualitatively similar to the numerical results. The results further revealed that a second turbine axially installed 10 rotor diameters from an upstream turbine underperformed for each of the individual devices. A reduction of about 20% for the bare turbine and about 25% for the ducted turbines compared to the corresponding upstream turbine. The ducted turbines still performed better than the bare turbine in terms of the peak power coefficient. The power reduction implies that the performance of the second turbine mainly depends on the axial distance between the turbines for the in-line configuration considered. Therefore, power reduction in this region can be improved if the turbines are placed apart such that they are minimally affected by the wakes from the turbines directly upstream. For all devices, the experimental investigation showed that the velocity of incoming flow decreases while the corresponding streamwise turbulence intensity increases as the flow passes through the turbine. The decreasing range of wake velocities and the increasing range of turbulence intensities were both the largest for the ducted turbines and the lowest for the bare turbine. However, all devices showed monotonically increasing and decreasing trends for the variations of wake velocities and turbulence intensities with increasing axial distance. Across the axial distance downstream, the bare turbine absorbed lesser energy from flow and exerted smaller effects on its wake than the ducted turbines. For all devices, the wake gradually recovered up until the recovery of flow was detected at about 13 and 20 rotor diameters downstream for bare and ducted turbines respectively. The turbulence intensities at those respective axial distances were slightly smaller than that at the inflow boundary. This finding is comparable to previous studies that downstream spacing could be in the order of 15-20 rotor diameters. The velocity ratio profile in the bare rotor wake compared well with published data from a recirculating water flume/channel and the spread in data was identified to be due to the differences in the range of turbulence intensities, blockage ratio (thus, variations in the thrust) and turbine model. For the two in-line turbines, the downstream bare turbine recovered faster than the downstream ducted turbines. At 24 rotor diameters downstream and beyond, the velocity ratio of the downstream bare turbine was very close to that of the isolated one (correspondingly at 15 rotor diameters downstream and beyond). At these distances, the second downstream turbine operated as if it was in isolation. For the ducted turbines, unlike in the isolated case, the wake flow ratio persisted beyond 30 rotor diameters downstream for the two ducted turbine arrangements with the turbulence intensity slightly less than the upstream conditions. However, the peak power coefficients of the downstream ducted turbines were still higher than that of the downstream bare turbine. Therefore, the deployment of additional ducted turbines in a tidal farm may have an advantage. Although, the precise economic estimates of energy cost are unavailable at the present state of development. The originality of this study is that it is the first to experimentally explore duct performance and far wake characteristics of flanged duct turbines with the in-line arrangement, and with the optimal shape of the duct determined by numerical simulations. Thus, this study has provided a comprehensive set of experimental data to the research community that will guide the course of future marine current turbine research and for validating theoretical and numerical methods. It is envisioned that the efficient and cost-effective deployment of full-scale tidal farms solely depends on the progress in finding optimum duct shape (or size), blade design, and turbine array layout.