Measured dynamic loading and power performance of tidal turbines in realistic flow conditions

Tidal turbines (akin to underwater wind turbines) are at a stage of development where full-scale prototype devices have been deployed and tested in the seas surrounding the UK and other locations in the world. The next step is to deploy farms or arrays of multiple devices to demonstrate operability, cost reduction and the ability to generate electricity at a larger scale.

In order to do this device developers and funders of the technology must have confidence and assurances that these arrays of tidal turbines will perform as predicted; but how do you predict something that you have never done before? Computer-based numerical models can simulate things before they are constructed or installed but without data to validate these models how can anyone know if they will be accurate?

This project addresses such a question "How can we aid industry to best validate dynamic (real time) array models for:(a) optimised array design and layout, (b) prediction of dynamic loadings and fatigue effects (rotors and blades) through inflow turbulence and device-device interactions and (c) reliability or planning for O&M
considering the lack of publically-available data?

Our answer is to:

1. To provide real, time -series data of loadings and power performance experienced by tidal turbines under realistic inflow conditions and when devices interact with one another (in array type configurations). At present there is little if any data on this (mostly average values or short time-scale experimental runs) which cannot sufficiently validate models. We will provide scale test data with all parameters required to use data sets for validation purposes
2. To provide measured time-series data for larger multiple-row arrays than has been previously conducted
3. To quantify so-called "steady loads" and measure changes in performance over time through long-term testing of a scale device(s) in the sea

And to make project data available directly to relevant marine energy stakeholders in a very timely manner

In the UK alone the Crown Estate has recently announced 6 new zones and 5 specific project sites for wave and tidal energy devices. Considering just the two most advanced tidal sites the short term installed capacity of 17MW is likely to incur capital costs of £120m so work to better prove array operability, power capture and reliability is timely in order to both increase understanding and confidence in the technology whilst helping to reduce costs. We believe device interaction will be avoided at this stage as it cannot be accurately quantified, at some point developers will not be able to avoid this and therefore reliable models must be developed.

Only the first-movers in the tidal turbine sector have real performance data collected over significant period of time but the quality/quantity of data is unknown and data for single device at a single site is limited as site conditions are generally unique. Certain variables are not accounted for such as varying inflow turbulence, device-generated (interaction) turbulence and developers at present might not measure blade root bending. We will.

Newer incoming developers also have to acquire this data for their own tools or purchase others which inhibit their competitiveness. We can level this part of the playing field allowing them a more rapid route to validating in-house tools allowing them to focus on development of device designs. For independent companies offering generic tools this data will allow improved validation which can only serve to increase the confidence of funding bodies and private investors have in this emerging but rapidly growing industry.

To facilitate and accelerate impacts of this research the project will upload authoring reports and time-series data sets to the internet (via an FTP portal or similar)and make this data available to all relevant UK stakeholders in the marine energy sector. We will not effectively embargo data by first publishing through academic publications (which can come later).