Item

Abstract

Demand Side Response management encourages elec- tricity demand reduction during peak hours. One avenue for achieving this is through Demand Side Units (DSUs). These are large electricity consumers who can afford to reduce their demand on the electricity grid when required. Issues with DSUs revolve around verification that the correct demand reduction takes place, with limited monitoring capabilities from the electrical grid operator Eir- Grid. This issue is studied here with the current methods thoroughly analysed and new methods proposed. In this report six different forecasting methods are presented, and their accuracy is compared using two different error metrics. Due to inherent stochasticity in demand it is found that there is no one fore- casting method which is unequivocally best, but the ‘Keep it simple’ weekly and the temperature dependent models are identified as the most promising models to pursue. Initial investigations suggest that a ‘proxy day’ mechanism may be preferable to the current method of verifying that the correct demand reduction takes place.