Performance quantification of marine current energy converters in constrained flow fields

Abstract

The Marine Current Energy Converter (MCEC) industry is currently at a stage where early farms are being designed. These farms will provide an indication of the commercial viability of the technology to future investors. Only limited full-scale device data is available in order to inform designers of performance, due to only small numbers of o shore deployments. Smaller-scale laboratory testing can provide greater understanding of device performance and loading. Understanding device performance and loading are important as they will impact on projects finances by affecting energy yield and allowing extreme loading to be designed for to avoid unplanned maintenance. The loading of MCEC devices in constrained flow fields is currently not well understood at either full or laboratory-scale.

The aim of this study is to investigate the influence of constrained flow fields on MCEC device loading and wake development. The results will inform: the limits of current analytical MCEC performance models, the calibration of empirical wake models, and the limits of existing computational fluid dynamic array models. The findings will therefore aid the development of design tools used in MCEC device modelling. This work presents the results from a series of small-scale laboratory static porous disk experiments in which the changes in loading and wake structure of a single device are characterised with variation in channel properties such as: area blockage ratio, Froude number and channel aspect ratio. Dual-array devices are also studied with the effects of device spacing and channel geometry upon wake formation and device loading characterised.

Thrust coefficient is shown to be a function of Froude number, area blockage ratio and channel aspect ratio; with each non-dimensional parameter contributing significantly to differences in loading. Despite differences in thrust coefficient with changes in channel aspect ratio it is shown that one-dimensional linear momentum actuator disk theory models are capable of accounting for these differences. Wake velocity profile, wake expansion, transition point from near to far wake and wake recovery are shown to be different in the vertical and horizontal planes in some channels geometries. Variation in channel geometry, i.e. the proximity of bounding surfaces, is shown to have an effect on these parameters indicating that it may be possible to calibrate semi-empirical wake models to predict wake development in either flows of constrained geometry or within array deployments. Device spacing in a dual-array is also shown to affect wake structure and device loading. The relationships between these two parameters are shown to be a function of device diameter/water depth ratio.

More information

Identifiers

ORCID for Bradley Keogh: orcid.org/0000-0003-2960-0918

ORCID for Abubakr Bahaj: orcid.org/0000-0002-0043-6045

Catalogue record

Export record