Abstract
Renewable energy is seen by many as the key to unlocking a future without fossil fuels for electricity generation. Tidal energy represents an untapped predictable energy source, the commercial exploitation of this resource is possible with decreased costs. Data are a key to further the understanding of design limitations and operations of a tidal turbine in order to reduce costs but data are often expensive to obtain. The subject of this thesis is the O2, a 2 MW floating tidal turbine developed and installed by Orbital Marine Power. Four research studies in this thesis demonstrate how the implementation of data-based methods can reduce costs for the tidal sector. These studies pertain to: the streamlining of legacy data processing methods; reprocessing onboard, non-wave specific, data streams to calculate environmental wave statistics; validation of an existing computational model using operational field data; and the application of new sensors and techniques for incident flow measurement for tidal turbine power performance assessments. A range of legacy data processing methods are streamlined to enable quick execution time or better access to data for the engineering team. Fast data processes are shown to reduce task person-hours by a factor of 10, while better access to data aided operational safety. A number of methods for calculating wave statistics are presented using onboard data streams. Wave heights accurate to ±0.5 m were achieved using REDACTED. A loads model validation campaign demonstrated the accuracy of the drive train modelling software used by Orbital Marine Power in predicting loads at two different operational set points. Limitations around fatigue loading for a given wave height were shown as well as yield implications of setting controller based load limits for operation. Limitations in the current standard for incident resource measurement when applied to floating tidal energy were explored and outlined. The application of a horizontal rotor height ADCP enabled the turbine-relative profiling of incoming flow that could be used as an alternative to established incident flow measurement techniques. Exploiting data sets for more than one use and applying sensors with a multi-use intention gives a low cost means of gaining more information about a system. This is important for a sector looking to reduce costs.