Particle swarm based model exploitation for parameter estimation of wave realizations

Ocean wave energy farms are composed of several wave energy converter devices. The objective of each converter is to capture the potential and kinetic energy in rolling ocean waves and convert them into electricity. An important task in optimizing power delivery from ocean wave energy farms is short term prediction of incoming ocean waves. Accurate predictions enable predictive wave energy converter control, energy storage control, and condition monitoring. In a novel approach we estimate the parameters of regular ocean waves by exploiting the domain with particle swarm optimization. We estimate initial parameters, and use those as guides in the swarm to estimate better ones. Our approach yields highly accurate estimates of ocean wave parameters that are very close to the Cramer-Rao lower bound on estimating these parameters. Our approach of domain exploitation with particle swarm optimization is also superior to stand alone global optimization methods such as simulated annealing and genetic algorithm.