Stochastic analysis and control for reliability improvement of sustainable energy systems

Future off-shore installations require network long-term planning and studying the impact of wind generation on the entire power system. Reliability assessment of energy systems integrated with renewable sources, and in particular wind farms, has been studied by several researchers recently. To deal with uncertainties in the energy sources, a number of methods have been proposed to model and predict their behavior. Different from the existing methods which are mainly discrete-time approaches, we propose in this paper an efficient method for optimal online prediction founded upon stochastic analysis of continuous-time Markov chains. The key feature of the chain is identified and the optimal online prediction problem reduces to a quiet simpler one by means of the Markov property. Performance indicators are stated and the method is illustrated on real data. The guarantee of reliability is still as objective underlying the second part of this paper which addresses the problem of optimal control of wind turbines. A power tracking controller, for some given set point reference, which is to operate in two switching regimes, is to be designed using the dynamic programming principle. The controller derivation is based on (i) a nonlinear electromechanically coupled model for variable-speed wind turbines with doubly fed induction generators, and (ii) the wind speed stochastic model provided by the first part of the paper.