Résumé
Energy demand aggregators are new actors in the energy scenario: they gather a group of energy consumers and implement a demand- response paradigm. When the energy provider needs to reduce the current energy demand on the grid, it can pay the energy demand aggregator to reduce the load by turning off some of its consumers loads or postponing their activation. Currently this operation involves only greedy energy consumers like industrial plants. In this paper we want to study the potential of aggregating a large number of small energy consumers like home users as it may happen in smart grids. In particular we want to address the feasibility of such approach by considering which scale the aggregator should reach in order to be able to control a significant power load. The challenge of our study derives from residential users' demand being much less predictable than that of industrial plants. For this reason we resort to queuing theory to study analytically the problem and quantify the trade-off between load control and tolerable service delays.