| This thesis proposes a methodology for planning, scheduling and on-line control of an energy storage system for the integration of wind energy. Using the case study of a remote wind-diesel system, the different time frames of the design and implementation process are detailed. First, a long-term planning approach for rating of the power and energy capacities of the ESS is presented, based on stochastic optimization. The formulation is then adapted into a hourly scheduling approach and results are compared with the expected cost of energy and energy requirements resulting from the planning study. The optimization results are used as training data for an artificial neural network, in an effort to generate an on-line control that captures inherent rules, using artificial intelligence. The ESS is realized as a two-level ESS and a general control structure for on-line operation of multi-level ESS is proposed and adapted for the wind-diesel system, as the first level in a hierarchical control. The system is evaluated in simulation and selected results are validated using a hardware-in-the-loop representation of the system, demonstrating that the proposed controller is realizable. |
