Energy extraction from ocean waves by heaving and flexing mechanical systems

Thesis (Ph.D.)--Boston UniversityThis thesis is concerned with devices that convert ocean wave energy to electrical energy, and specifically with devices that are composed of a sequence of linked cylinders that nominally float half-submerged on the surface of the ocean. An example is the Pelamis Wave Energy Converter, which is manufactured by Pelamis Wave Power, a company headquartered in Scotland. The thesis is broader in scope, and the results should be applicable to a wide variety of devices with the same general features. Arguments are given in support of using a constant frequency wave excitation model, given that ocean waves in coastal regions are relatively narrow-band and that the analytical predictions are not too sensitive to frequency. Arguments are also given to the effect that a deep water approximation is justified for incident ocean waves under the circumstances of typical deployment. Analysis in the thesis also supports the modeling of the electromechanical converters at the joints between linked cylinders as mechanical dashpots, where the rate of energy converted is the power dissipated by the dashpots. An analytical model for the dynamics of the structural response to incident waves is developed using basic principles of mechanics and fluid dynamics. Principal results include the prediction of the heaving of the individual links as a function of the amplitude of the incident wave, the geometrical parameters of the cylindrical elements, the number of elements, the effective dashpot constant, and the frequency (the only devices considered in this thesis are those without intrinsic resonances built into them). It is shown that there is an optimal choice for the dashpot constant so that the power conversion is a maximum. A measure of the effectiveness of power conversion is the capture width, which is the ratio of power converted to incident wave power per unit crest length. It is shown that, when the number of links is moderately large, a model of a continuous limp beam is a good approximation, and the model yields relatively simple analytical predictions, which include the prediction that the greatest possible capture width is of the order of 2(ka)L, where k is the wavenumber, a is the cylinder radius, and L is the total length of the device