Multi-time Scale Stochastic Model Predictive Control for Cooperative Distributed Energy Scheduling of Smart Homes

Abstract

In this chapter, a multi-time scale stochastic model predictive control (MPC) approach is applied to solve the cooperative distributed energy scheduling problem of smart homes. In this problem, a variety of energy resources are considered for each smart home, and every smart home has a capability of power transaction with the retailer through the electrical grid as well as with the other connected smart homes. The challenges of the problem include modeling the technical and economic constraints of the energy resources and addressing the variability and uncertainty issues of renewables’ power that change the optimization problem to a stochastic, dynamic (time-varying), and mixed-integer nonlinear programming (MINLP) problem. To deal with the variability and uncertainty issues, a multi-time scale stochastic MPC is applied. Applying the multi-time scale approach in the stochastic MPC is able to simultaneously consider the precise resolution for the problem variables and the vast vision for the optimization time horizon. In addition, linear programming (LP) and genetic algorithm (GA) are combined (GA-LP) and applied in the problem as an effective and fast optimization technique. The numerical studies about the small and large systems demonstrate the competences of the proposed approach.KeywordsCooperative distributed energy schedulingMulti-time scale model predictive control (MPC)Photovoltaic (PV) panelsSmart homesStochastic optimization