Uncertainty in hydrodynamic model test experiments of wave energy converters

Renewable energy is key to solving the great challenges of climate change, energy security, and pollution. Ocean wave energy is a nascent renewable energy with a vast technical potential, but a relatively high Levelised Cost of Energy (LCOE) due to limited commercial Wave Energy Converter (WEC) technologies. This suggests that WEC development methods and guidelines are still maturing. For early-stage WEC development, which requires hydrodynamic model test experiments, a recognised gap is experimental uncertainty. This gap is striking because uncertainty can significantly influence experimental results. Despite extensive knowledge on experimental uncertainty in similar fields such as shipping and marine structures, applying this knowledge in WEC experiments is inadequate and carries risk because WECs uniquely maximise motions, use a Power Take-Off (PTO) system, and often have complex geometry and moorings that influence motions ‚- all of which can introduce significant experimental uncertainty in the results of power performance and hydrodynamic loads. To better understand the causes and effects of uncertainty in hydrodynamic model test experiments of WECs, we reviewed technical guidelines and literature to identify major uncertainties needing investigation, then conducted a series of model test experiments designed to investigate these uncertainties. This research presents an overview of and experimental investigations into uncertainty in hydrodynamic model test experiments of WECs, focusing on a Oscillating Water Column (OWC) WEC. The set of experiments, representative of Technology Readiness Levels 1-4, assess power performance and hydrodynamic loads in regular and irregular waves. The case study OWC WEC is based on Australian company Wave Swell Energy's (WSE) Uniwave technology, a bottom-fixed device with a unidirectional airflow PTO. Through reviewing the literature of WEC model test guidelines, advances, and uncertainties, we found that despite substantial progress in developing best practices, they remain dispersed across many documents, are inconsistent in some parameters and procedures, and have gaps or are inadequate in several areas. WEC-specific guidance was found to be lacking in: the modelling of moorings, PTOs, and arrays and clusters; identifying and modelling survival conditions; installation and tow-out tests; specific tests for calibrating and validating numerical models; methods for extrapolating model-scale results; full-scale validation; and, most important to this research, understanding of and methods to account for measurement uncertainty, scale effects, and laboratory effects. Hence, these uncertainties became the focus of subsequent experimental investigations: (1) a 1:30 scale model test of the OWC WEC in the Australian Maritime College (AMC) Model Test Basin, which, building on knowledge obtained from experimental work, was conducted to better understand measurement uncertainty and develop new WEC-specific uncertainty analysis (UA) methods; (2) a series of model tests at three scales (1:40, 1:30, 1:20) of the OWC WEC in the AMC MTB, conducted to identify, quantify, and evaluate parameters causing scale effects; and (3) reproducing this model test at 1:30 scale in a similar shallow water wave basin, the Queen's University Belfast Coastal Wave Basin, to identify, quantify, and evaluate parameters causing laboratory effects. In experiment (1), we develop a comprehensive UA methodology and apply it to the OWC WEC experiment, demonstrating how and when UA can be used throughout an experimental program. In doing so, we outline UA principles, identify parameters causing measurement uncertainty, and develop new WEC-specific methods for General Uncertainty Analysis (GUA), evaluating Type A and Type B uncertainty, and the Monte Carlo Method (MCM) to propagate uncertainty. We found that GUA is indispensable in experimental planning and design because it assures relevant and high-quality results are obtained, that a new Type A uncertainty evaluation method reduces the number of required repeat runs thereby saving time and cost, and that the MCM effectively and efficiently propagates uncertainty for the complex OWC WEC experiment. We also give detailed examples of evaluating Type B uncertainties. Results from the experiment show the expanded uncertainty averaged ¬±16% for capture width ratio (C\\(_W\\) ) and ¬±6% for loads, with Type B uncertainty tending to be slightly larger than Type A uncertainty, and uncertainty in irregular wave results slightly smaller than regular waves. Key causes of uncertainty in C\\(_W\\) were measurements used to derive the lower level measurands of incident wave power and OWC power, and the PTO modelling. We conclude that UA is required in WEC model tests because it assures and quantifies the quality of experimental results. Specific recommendations are also offered to update guidelines on UA for WECs. Importantly, experiment (1) generated the knowledge required to determine whether experimental results across model scales in experiment (2) or between laboratories in experiment (3) agreed or disagreed. In (2), we found moderate to major differences on average in power and loads results across scales (10-30%+). Significant scale effects in the results were evaluated to be mainly caused by deviating nonlinear incident wave profiles and deviating quadratic PTO damping due to maintaining the similitude condition of orifice-OWC chamber ratio. In (3), we found moderate to major differences on average in power results between laboratories. Significant laboratory effects in the results were evaluated to be mainly caused by differences in the test environment (wavemaker and nonlinear wave transformations), the model (deployment position and the PTO influenced by different test environment ambient conditions), and instrumentation (loads measurements). Other instrumentation, water properties, air compressibility, and human factors were evaluated to have a negligible to minor contribution to the differences in results in both (2) and (3). These results clearly showed that, despite conducting the model test experiments to a high standard according to international guidelines, model scale and the laboratory can significantly influence experimental results. Therefore, scale and laboratory effects cannot be neglected when carrying out these experiments. The primary conclusion from this research is that experimental uncertainty is an inseparable part of and can significantly influence the results of hydrodynamic model test experiments of WECs. Therefore, stakeholders in early-stage WEC development should be aware of this; should invest resources, time, and money into accounting for newly identified parameters causing measurement uncertainty, scale effects, and laboratory effects; and should consider implementing the outcomes of this research into future guidelines on WEC model tests. This research contributes to the efforts in developing an international standardised set of robust, consistent, and validated guidelines on WEC development, needed to assure the reliability of model test results, reduce technical and financial risks, enable better data-driven decisions, and ultimately reduce wave energy's LCOE.