Optimization-based estimation of power capacity profiles for activity-based residential loads.

This paper proposes a framework to determine capacity profiles in smart buildings. In this scheme the users choose a level of power capacity to account for their stochastic demand while paying the corresponding electricity prices through a flexible time-and-level-of-use pricing policy. We formulate a two-stage stochastic optimization model that minimizes the total cost of booking a power capacity level and meeting the energy demand for the planning horizon. We present two approaches to select the scenarios for the stochastic optimization. In the first approach, we assume that the probability distributions of the start times of the loads are known, and the scenarios are generated using those distributions. In the second approach, we assume that only historical consumption data is available and we propose a new algorithm to build the scenarios using this data. Our simulation experiments validate the performance of both approaches and report cost savings of up to 16%. [ABSTRACT FROM AUTHOR]