How to improve households' participation in Demand Response programs: the promising contribution of predictive control
Résumé
An increasing attention has been paid to Residential Demand Response programs from utilities and aggregators of dispersed capacities have emerged looking for demand response (DR) products which can be characterized in terms of power reduction and time duration. With the incoming results from some of the numerous smart grid demonstrators, it has been observed that households might not respond to price signals or at least not as much as expected. For programs aiming to reduce peak demand, the choice of the control strategy is the key issue for increasing households' involvement. For electrical space heating, in order to be fully efficient, the control strategy should depend on the building's characteristics and take into consideration household's behavior or habits. If the control strategy implemented uses only standard bound limits for indoor temperature as the comfort constraint, households facing discomfort situations might withdraw from DR programs or stop responding to it. This case is more likely to occur for low mass envelope buildings which represent the largest part of the French building stock and which have the highest load curtailment potential in terms of power demand. We suggest to compare different control strategies to evaluate the potential for load curtailment under comfort constraints for several types of buildings (those buildings have different level of insulation) modeled using a simplified thermal model for building and different types of electrical heating systems. Using a day as the time horizon and considering that the maximum peak period is two hours, we identified four control strategies as suitable for load curtailment programs, including a predictive controller with comfort constraints. The strategies are compared for each type of building in terms of load curtailment potential (characterized by power and duration of the curtailment) and consumed energy during the day. We also compare them accordingly to the discomfort indicators we defined. While those strategies lead to comparable results for well insulated buildings, the more sophisticated strategies can be interesting for poorly insulated buildings as they allow a diminution of discomfort effects and an increase in households' participation and interest in DR programs.