Return to search

Optimising the output of an articulated WEC through the use of a biologically-inspired and artificially evolved neural network

This thesis investigates the implementation of a practical control strategy for a simple wave energy converter in regular and irregular waves. Inspired by the neurophysiology of the lamprey, this work looks at the limitations of using this neural structure as a method for control of an articulated wave energy converter. Initially starting with a very simple mechanical model of a single heaving buoy, evolutionary techniques are employed to evolve a single lamprey segment that will be capable of acting as a controller. Without prediction of the incident wave, wide-bandwidth latching controllers, are evolved that show significant improvements over optimal damping in increasingly complex waveforms. More complex mechanical configurations are also investigated, expanding the simple heaving buoy into two and three interconnected buoys with power being developed through their relative motion. Neural latching controllers are evolved for these different configurations and it is shown that an effective neural latching controller cannot be evolved for more than two interconnected buoys. This thesis investigates the processes of producing a viable implementation of a latching strategy using neural oscillators as a non-linear feedback loop. It covers their performance in regular and irregular waves and demonstrates the limitations of latching control when applied to an articulated system.

Identiferoai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:659741
Date January 2006
CreatorsMundon, Timothy R.
PublisherUniversity of Edinburgh
Source SetsEthos UK
Detected LanguageEnglish
TypeElectronic Thesis or Dissertation
Sourcehttp://hdl.handle.net/1842/12703

Page generated in 0.0024 seconds