An optimal day-ahead dispatch strategy for deferrable loads

Deferrable loads can actively participate in the power system operation as a schedulable resource. The loads are modulated by shifting their demand characteristics along the time horizon, i.e., orchestrating the starting instants. Through flattening peaks and valleys of the demand curve, generating units are released from the burden of frequent output adjustments and a higher operating efficiency can be achieved. An optimal day-ahead dispatch strategy is proposed for deferrable loads. The next-day demand characteristics of the deferrable loads can be derived using historical data or behavior analysis. The starting instants of the loads are to be delicately delayed or advanced within the 24-h period. By such modulation, the aggregation of the controlled loads aims to closely track the dispatch objective that is assigned by the control center. An optimal model is constructed to describe the dispatch problem, which can be converted into a mixed integer quadratic programming (MIQP) one. The coefficients of the MIQP problem are computed using MATLAB. The solver SBB of GAMS is applied to solving the MIQP problem. Simulations are conducted to verify the effectiveness of the proposed optimal dispatch strategy. The aggregation of the controlled loads can closely track the dispatch objective.