Improving the network infeed accuracy of non-dispatchable generators with energy storage devices

The power output of generators based on renewable energy sources is often difficult to predict due to the non-deterministic behaviour of the energy source. Particularly in the case of wind turbines this leads to unpredicted line loading and requires balancing energy, at relatively high costs, depending on market structures. Consequently, the income from the production from such non-dispatchable generators can be significantly reduced by the penalty costs incurred. This paper investigates the potential of operating an energy storage device in parallel with the non-dispatchable generator in order to compensate the inaccuracies of the forecasted infeed and to avoid infeed deviations. A time series based simulation methodology is discussed, suitable for any type of non-dispatchable generator. The methodology contains a procedure for simulating different forecast errors, applying an exponentially weighted moving average approach. Analysis procedures and system performance indices are introduced for the evaluation of the configuration's performance. The applicability is shown in two case studies, using measurement data from a wind turbine and from a photovoltaic system. Both case studies show that the suggested configuration considerably improves the reliability or dependability of the network infeed, in turn reducing the demand for balancing energy and back-up generation. The relation between forecast error magnitude and required energy capacity is identified and the coherence of the time series analysis is discussed. (C) 2008 Elsevier B.V. All rights reserved.