Optimising design of large tidal energy arrays in channels: layout and turbine tuning for maximum power capture using large eddy simulations with adaptive mesh

by Timothy Andrew Divett

Institution: University of Otago
Year: 0
Keywords: Tidal power; tidal energy; tidal arrays; turbine arrays; adaptive mesh modelling; LES modelling; turbine tuning; tuned drag elements; maximum power capture; optimising array layout
Record ID: 1312687
Full text PDF: http://hdl.handle.net/10523/4995


Tidal energy has the potential to generate sustainable electricity from tidal flows using arrays of turbines. This thesis is the first exploration of the layout of large arrays in channels to explicitly model turbine-scale turbulence and channel-scale flow reduction due to power extraction. As such this project bridges the gap in scale between previous hydrodynamic modelling studies of coastal ocean-scale or small arrays. Optimising array layout becomes a balance of maximising total power capture and power per turbine while minimising impact on natural flow. The flow around each array has been modelled using Gerris — an adaptive mesh, Large Eddy Simulation, open-source flow solver — to solve the 2-D incompressible Navier-Stokes Equations. Idealised model components have been developed, reducing computational cost to enable simulation of many hundreds of array layouts, each with a unique tuning curve. The channel-scale flow is driven by a tidal head balanced by friction. This differs from previous 2-D models of small arrays that have applied a constant amplitude velocity, neglecting velocity reduction due to power capture. For all arrays, total power capture is near-linearly related to velocity reduction from the natural flow. Deviations from linearity of up to 15% are possible by packing turbines into one side of the channel. For the largest arrays, velocity decreases towards 60% of the velocity without turbines and the head loss across the array tends towards the driving head. The optimum turbine tuning for maximum power capture changes with layout and can be up to 6× higher in small arrays than in larger arrays. The number of rows, n, and turbines per row, N, were varied within arrays consisting of: uniform rows, with a range of global blockages simulated by varying N ; packing turbines densely into one side of the channel and staggering turbines in alternate rows. In uniform rows, there is an optimum global blockage for maximum power per turbine. As n increases that global blockage decreases near-linearly and that maximum power per turbine decreases. The total power capture, however, continues to increase as global blockage increases beyond that optimum and also as n increases. Packing turbines into one side of the channel increases power per turbine and total power capture, but only up to an optimum packing density and only if global blockage is below the optimum. This optimum packing density decreases with increasing rows. When global blockage is higher than the optimum, increasing packing density does not further increase the total power capture. Staggering turbines in alternate rows with low packing density increases power capture by up to 30% compared to the previous regularly arranged rows. At packing densities above 0.6 there is negligible difference in power capture from staggering, due to smaller gaps between turbine wakes. Optimum array design for a specific development is likely to be driven by a combination of the array-scale effects found in this thesis, economic factors and micro-siting…