Load monitoring concept for tidal turbines based on strain measurements at the support arm of the vessel

Abstract

Tidal turbines show predictable energy output. Thus, this form of energy generation has potential to supplement established, but less predictable, renewable energy systems like PV and wind energy. However, the levelized cost of electricity (LCOE) of tidal energy is roughly ten times higher than those of solar and wind energy. Therefore, the LCOE must be reduced to ascertain economic competitiveness. A high share of the LCOE consists of the maintenance costs. Thus, decreasing the maintenance costs is an effective way to reduce the LCOE.

The current lack of monitoring systems necessitates a time-based maintenance schedule. Compared to predictive maintenance systems, this results in premature and sometimes unnecessary service. The application of a load monitoring system and corresponding maintenance plans allows to maintain the turbines based on the loads experienced, reducing not needed service and therefore maintenance costs.

A novel concept for a load monitoring system for tidal turbines based on strain measurements has been developed. The considered tidal turbine system consists of a support arm structure which is attached to a floating vessel. The strains are measured with strain gauges attached to the bespoken support arm structure. To reduce effects of the environmental conditions, strains are measured on the arm above sea level and not at the rotor. Measuring the strains on the support arm distant from the rotor necessitates a robust transfer function to determine the rotor loads.

In this paper, the development and optimization of the transfer function is presented. FEM simulations are used to determine the strains on the arm for varying static rotor loads. The simulation results are used to derive a transfer function from arm strain to rotor loads. The transfer function with strain measurements on the support arm in the field as an input comprises a real-time capable load monitoring system. To develop a cheap but reliable system, the number of measuring points of the system needs to be minimized, while still achieving a robust transfer function. The influence of the strain gauge’s measuring range is also considered for the robustness of the transfer function.

An optimization algorithm is used to find the minimal amount of measuring points for a 1 % relative error between applied and measured load. The layout is optimized regarding the condition of the transfer function to mispositioning and misalignment  measuring points.