Multiple Time Resolution Stochastic Scheduling for Systems with High Renewable Penetration

Abstract

This paper presents a stochastic unified unit commitment and economic dispatch model for short-term operations scheduling of power systems with high renewable penetration. The model is a unique stochastic real-time tool that performs economic dispatch with up to 36-hour look-ahead capability, providing dispatch instructions for the first time interval and advisory commitment and dispatch schedules for the remaining scheduling horizon. Variable time resolution is used to contain computational requirements, i.e., finer time resolution is used during the first hours and coarser time resolution during the last hours of the scheduling horizon. The mathematical formulation comprises a MILP problem and it is presented as a two-stage scenario-based stochastic, multiple time resolution unit commitment model. The proposed stochastic model is tested against its deterministic counterpart via an annual simulation of Greek power system short-term operations using real 2013 data. Performance indices concerning dispatch costs, energy balances, and unit cycling are calculated and discussed.