Accurate predictions of the annual energy yield from wave energy converters are essential to the development of the wave industry. The current method based on power matrices uses only a small part of the data available from sea state estimations and it is consequently prone to inaccuracies. The research presented in this work investigates the issue of energy yield prediction and questions the power matrix method. This is accomplished by quantifying the influence of several directional sea states parameters on the performances of wave energy converters. The approach taken was to test several wave energy converters in the Edinburgh Curved tank with a large set of sea states. The selected wave energy converters are a fix OWC, a set of two OWCs acting as a weak directional device and the desalination duck model. Uni-modal and bi-modal sea states were used. For the uni-modal sea states, parameters related to the wave system shape were considered. For the bi-modal sea states, the relative position of the wave system peaks was investigated and the uni-modality index was introduced to quantify the degree to which sea states could be considered bi-modal. For all sea states, the significant wave height was kept constant. The experimental work required good spectral estimates. The MLM and MMLM were adapted to deterministic waves to improve their stability and accuracy. A routine to isolate wave systems was also developed in order to estimate parameters with respect to each wave systems. For uni-modal spectra, parametric models of the observed performances of the devices could be devised. The frequency spreading and its interaction with the energy period proved to be as important as the energy period itself, which suggests that the frequency spreading should be used for energy production prediction. For bi-modal spectra, evidence of the duck sensitivity to directionality was found while the OWCs were not affected.

No description available.

No description available.

Removable Epoxy Foam (REF) and other rigid foams experience severe changes in structure and properties when exposed to high heat. As thermal energy breaks network bonds in the foam many species are formed, including large polymer-like network fragments and smaller solvent-like molecules. During this process a liquid phase may form. The vapor-liquid equilibrium (VLE) behavior of the polymer solutions formed during initial decomposition can be highly non-ideal. In this research VLE behavior of high-temperature polymer solutions was studied and a procedure was developed for predicting that behavior during decomposition of rigid foams. A high-temperature VLE facility was built and validated, and equilibrium pressures were measured at temperatures between 75 and 250ºC for six polymer/solvent systems: two polymers – polyethylene glycol and polystyrene – with each of three solvents – benzene, furan, and 4-isopropylphenol. Calculations from eighteen polymer solution models were compared with experimental results to determine which model best described the VLE behavior. These models included six existing activity coefficient models used alone, as well as in combinations with the Peng-Robinson equation of state (EOS) through the Wong-Sandler mixing rules. Because several of the models required values for polymer volumes, a comparison of the GCVOL and GCMCM group-contribution volume estimation methods was performed. GCMCM was found to give lower overall deviations from literature polymer volume data. The models involving an equation of state required EOS parameter values for the pure polymers. A new method for determining these parameters was proposed. Models using parameters from the new method gave better agreement with equilibrium pressure data than models using parameters from the recommended method in the literature. While agreement with equilibrium pressure data was similar for several models, some models predicted a liquid phase split under certain conditions. Data were not available to verify the presence of two liquid phases, but are needed to make an appropriate recommendation of the best model. If liquid phase splitting does not occur, it is recommended that the UNIFAC-ZM activity coefficient model be used alone. If phase splitting behavior is observed, it is recommended that the UNIFAC-FV activity coefficient model be used in combination with the Peng-Robinson EOS.