Stochastic Analysis of Marine Renewable Energy Devices: Wave, Wind, and Hybrid Systems

Abstract

Renewable energy sources play an essential role in addressing the rising global energy demand. Among the renewable sources, offshore locations have a considerable potential for energy harvesting, from the many sources including offshore wind, ocean current, wave, salinity gradient, thermal, biomass, tidal stream, and tidal range. Offshore wind turbines are currently the most promising technology to harvest marine renewable energies (MREs). Since offshore wind sites often possess high wave energy resources, and offshore wind turbines and wave energy converters face similar challenges in terms of relatively high LCOE, there is a possibility of integrating both into a single hybrid platform to reduce costs associated with installation, infrastructure, and maintenance. The wind industry has converged to the three-bladed horizontal axis wind turbine, while the supporting platform may be either bottom-fixed or floating. These wind turbine platforms were initially adopted from the extensive Oil and Gas industry experience. However, some modifications are necessary to attend the requirements for wind turbine applications and reduce costs to become economically competitive with other renewable sources. Unlike the wind industry, wave energy devices have a variety of absorption mechanisms, and several wave energy converters (WECs) have been proposed. To date, the majority of WECs are in a lower level of maturity with few devices at the stage of producing energy, and consequently further developments are needed. Under these circumstances, hybrid platforms may contribute to the development and feasibility of wind and wave technologies due to potential cost reductions. In addition, different hybrid combinations may provide additional benefits, such as improving the platform response. However, the lack of experience with WECs and hybrid platforms still represents a strong barrier. In this general context, reduced-order models that describe the behaviour of MREs are fundamental at the early stages of design, where the changes in the design concepts are more flexible. These numerical models can be used to explore and optimise all promising conceptual designs efficiently and help to understand the fundamental behaviour of such devices. However, the models must accurately represent the physics; otherwise, the optimisation results can lead to sub-optimal designs or even unfeasible concepts. In this regard, numerical models must be suitable for automatic optimisation procedures and have flexible applications while producing reliable results with low computational costs. The model’s efficiency is particularly important for MREs because the devices are subjected to an uncertain environment, containing stochastic loads from wind, wave, current, earthquake, and ice loading. As a result, several environmental conditions need to be assessed to investigate the performance of new concepts since MREs operate for very long periods of time. Based on the above, the primary contribution of this thesis is to create a fast and reliable numerical method to be used as a framework for the stochastic analysis of wave, wind, and hybrid energy devices. Since the environment can be modelled using spectral and statistical properties, the numerical model uses the frequency-domain (FD) approach. Important sources of nonlinearities are estimated using random vibration techniques, such as statistical linearisation and quadratization approaches, which provide a reliable prediction of the first and second-order motion of the system’s nonlinear dynamics. Examples of application and analysis are provided through the thesis to a range of WECs (point absorbers, oscillating wave surge converter, and oscillating water column), floating offshore platforms (spar and semisubmersible), and a hybrid platform (semi-submersible floating offshore wind turbine with either submerged or floating point absorbers). The results demonstrated a reliable estimation of the response in terms of power spectral density, probability distribution, mean response, and mean power produced. The results were compared against their respective nonlinear time-domain (TD) simulations and an established numerical code for floating offshore wind turbines (OpenFAST). The main advantage of the proposed model is the low computational effort and storage requirement, being approximately three to four orders of magnitude faster than TD simulation for the statistical linearisation and two orders faster for the statistical quadratisation. The model also contributes to understanding the fundamental knowledge of such energy devices, such as their nonlinear effects and coupling between devices, and how to combine different devices. For instance, point absorbers can absorb part of the wave energy while reducing the motion of a floating wind turbine, which can extend the hybrid platform’s lifetime. Therefore, this thesis provides a foundation for later engineering-side design to optimise and analyse MREs, further developing MREs technologies.