Takafumi Nishino

Tidal power generation - A review of hydrodynamic modelling

Extracting power from the tide is a potential avenue for renewable energy production but is also a significant engineering challenge. This challenge has many different aspects but the basic problem is the hydrodynamic problem of converting the movement of the ocean into mechanical power. This paper presents a review of some of the hydrodynamic modelling techniques which can be used to model tidal barrages and tidal turbines. The analysis of these is broken down into different length scales, ranging from a single device, to an array of devices, and up to regional scales. As well as discussing modelling techniques some of the hydrodynamic problems, such as resource assessment and efficiency of power generation, are discussed.

Numerical Study of Wind-Tunnel Sidewall Effects on Circulation Control Airfoil Flows

Two- and three-dimensional numerical simulations are performed of the flow around a circulation control airfoil (using a Coanda jet blowing over a rounded trailing edge) placed in a rectangular wind-tunnel test section. The airfoil model spans the entire tunnel and the span-to-chord ratio of the model is 3.26. The objective of this numerical study, in which we solve the compressible Reynolds-averaged Navier―Stokes equations in a time-resolved manner (but the solutions eventually converge to steady states), is to investigate the physical mechanisms of wind-tunnel sidewall effects on the flow, especially in the midspan region. The three-dimensional simulations predict that the Coanda jet flow is quasi-two-dimensional until the flow separates from the trailing edge of the airfoil; however, the spanwise ends of this Coanda jet sheet then three-dimensionally roll up on the side walls of the wind tunnel to form two large streamwise vortices downstream. Careful comparisons between the two- and three-dimensional simulations reveal that the wind-tunnel stream goes below the airfoil more in the three-dimensional cases than in the two-dimensional cases due to the presence of these two streamwise vortices downstream. This results in smaller lift and larger drag being produced at the midspan of the airfoil in the three-dimensional cases than in the two-dimensional cases.

Local blockage effect for wind turbines

This paper presents a combined theoretical and CFD study on the fluid-mechanical limit of power extraction by a closely-spaced lateral array of wind turbines. The idea of this study originates in recent studies on the array optimisation of tidal/marine turbines, for which the power coefficient of each turbine is known to increase significantly if the lateral spacing between turbines, or the local blockage, is optimised. The present study, using 3D Reynolds- averaged Navier-Stokes (RANS) simulations of a boundary-layer flow over a closely-spaced lateral array of up to 9 actuator discs, suggests that a similar—albeit less significant—power increase due to the effect of local blockage can be achieved even for wind turbines. A possible theoretical approach to estimating this power increase is also discussed.

Effect of Jet Nozzle Lip Momentum Loss on Circulation Control Airfoil Performance

*† Large-eddy simulations of flow around a circulation control airfoil (using a Coanda jet blowing over its trailing surface) are performed to investigate the influence of jet-nozzle-lip thickness on airfoil performance. The airfoil geometry is only slightly changed from our previous LES study [Nishino et al., Physics of Fluids, Vol. 22, 2010, 125105] to study three different nozzle-lip thickness cases; the geometry inside the nozzle is maintained the same. The results show that the jet profile across the nozzle exit is insensitive to the nozzle-lip thickness; however, the jet flow downstream of the nozzle exit decelerates more rapidly and thus the circulation around the airfoil decreases as the nozzle-lip thickness increases. It is subsequently shown that this effect is mostly cancelled out by adjusting the jet blowing rate in such a way that the difference of momentum loss arising from the nozzle lip is taken into account, demonstrating that the performance of a Coanda jet on a circulation control airfoil is determined not only by the jet momentum at the nozzle exit but also by the momentum loss behind the nozzle lip. These results suggest that it may be useful to define a new jet momentum coefficient that takes account of the momentum loss due to the nozzle lip, which can be roughly estimated once the velocity of the flow above the nozzle lip is known.

Low-Order Modelling of Blade-Induced Turbulence for RANS Actuator Disk Computations of Wind and Tidal Turbines

Modelling of turbine blade-induced turbulence (BIT) is discussed within the framework of three-dimensional Reynolds-averaged Navier-Stokes (RANS) actuator disk computations. We first propose a generic (baseline) BIT model, which is applied only to the actuator disk surface, does not include any model coefficients (other than those used in the original RANS turbulence model) and is expected to be valid in the limiting case where BIT is fully isotropic and in energy equilibrium. The baseline model is then combined with correction functions applied to the region behind the disk to account for the effect of rotor tip vortices causing a mismatch of Reynolds shear stress between short- and long-time averaged flow fields. Results are compared with wake measurements of a two-bladed wind turbine model of Medici and Alfredsson [Wind Energy, Vol. 9, 2006, pp. 219-236] to demonstrate the capability of the new model.

The Efficiency of Tidal Fences: A Brief Review and Further Discussion on the Effect of Wake Mixing

Recent discoveries on the limiting efficiency of tidal fences are reviewed, followed by a new theoretical investigation into the effect of wake mixing on the efficiency of ‘full’ tidal fences (i.e. turbines arrayed regularly across an entire channel span). The new model is based on the momentum and energy balance equations but includes several unclosed terms, which depend on the actual (three-dimensional) characteristics of turbine near-wake mixing and therefore need to be modelled empirically. The new model agrees well with three-dimensional actuator disk simulations when those unclosed terms are assessed based on the simulations themselves, suggesting that this low-order model could serve as a basis to analyse how various physical factors (such as the design of turbines) affect the limiting efficiency of tidal fences via changes in those terms describing the characteristics of turbine near-wake mixing. Also discussed is the effect of wake mixing on the efficiency of ‘partial’ tidal fences.Copyright

URANS simulations of separated flow with stall cells over an NREL S826 airfoil

A series of wind tunnel measurements and oil flow visualization was recently carried out at the Technical University of Denmark in order to investigate flow characteristics over a 14% thick NREL S826 airfoil at low Reynolds numbers. This paper aims at presenting numerical simulations of the same airfoil using unsteady Reynolds-averaged Navier-Stokes (URANS) approach. Results of the simulations are demonstrated in terms of mean flow velocity, lift and drag, as well as pressure distribution, and validated against available experimental data. The simulations are carried out with a wide computational domain (with a span-to-chord ratio of 5) and it is illustrated that the URANS approach is capable of predicting 3D spanwise structures, known as stall cells.

Flow through a very porous obstacle in a shallow channel

A theoretical model, informed by numerical simulations based on the shallow water equations, is developed to predict the flow passing through and around a uniform porous obstacle in a shallow channel, where background friction is important. This problem is relevant to a number of practical situations, including flow through aquatic vegetation, the performance of arrays of turbines in tidal channels and hydrodynamic forces on offshore structures. To demonstrate this relevance, the theoretical model is used to (i) reinterpret core flow velocities in existing laboratory-based data for an array of emergent cylinders in shallow water emulating aquatic vegetation and (ii) reassess the optimum arrangement of tidal turbines to generate power in a tidal channel. Comparison with laboratory-based data indicates a maximum obstacle resistance (or minimum porosity) for which the present theoretical model is valid. When the obstacle resistance is above this threshold the shallow water equations do not provide an adequate representation of the flow, and the theoretical model over-predicts the core flow passing through the obstacle. The second application of the model confirms that natural bed resistance increases the power extraction potential for a partial tidal fence in a shallow channel and alters the optimum arrangement of turbines within the fence.

Effect of the number of blades and solidity on the performance of a vertical axis wind turbine

Two, three and four bladed -shape Vertical Axis Wind Turbines are simulated using a free-wake vortex model. Two versions of the three and four bladed turbines are considered, one having the same chord length as the two-bladed turbine and the other having the same solidity as the two-bladed turbine. Results of the two-bladed turbine are validated against published experimental data of power coefficient and instantaneous torque. The effect of solidity on the power coefficient is presented and the instantaneous torque, thrust and lateral force of the two-, three- and four-bladed turbines are compared for the same solidity. It is found that increasing the number of blades from two to three significantly reduces the torque, thrust and lateral force ripples. Adding a fourth blade further reduces the ripples except for the torque at low tip speed ratio. This work aims to help choosing the number of blades during the design phase of a vertical axis wind turbine.

Development of a stochastic computational fluid dynamics approach for offshore wind farms

In this paper, a method for stochastic analysis of an offshore wind farm using computational fluid dynamics (CFD) is proposed. An existing offshore wind farm is modelled using a steady-state CFD solver at several deterministic input ranges and an approximation model is trained on the CFD results. The approximation model is then used in a Monte-Carlo analysis to build joint probability distributions for values of interest within the wind farm. The results are compared with real measurements obtained from the existing wind farm to quantify the accuracy of the predictions. It is shown that this method works well for the relatively simple problem considered in this study and has potential to be used in more complex situations where an existing analytical method is either insufficient or unable to make a good prediction.