Decision-driven uncertainty management in multi-energy virtual power plant: A robust multi-market scheduling framework

Abstract

This paper studies the multi-market robust scheduling for a multi-energy virtual power plant (MEVPP) under decision-driven demand response uncertainty. A multi-market robust scheduling framework is proposed to jointly optimize the operation of MEVPP across electricity, reserve, gas, and carbon markets. To address uncertainties arising from wind power fluctuations and various load variations, a two-stage robust optimization model is developed to effectively manage both decision-independent uncertainties (DIUs) and decision-dependent uncertainties (DDUs). Specifically, the flexible load fluctuations induced by electricity market price decisions are modeled as DDUs. The presence of DDUs dynamically alters the boundaries of the uncertainty set, making traditional robust optimization algorithms challenging to converge within a finite number of iterations and even causing oscillations. To efficiently solve this issue, a parametric column-and-constraint generation (C&CG) procedure is designed for the two-stage robust optimization model. Simulation results demonstrate that the proposed model not only effectively mitigates the impacts of DIUs and DDUs, but also reduces carbon emissions by 36.6% and cuts costs by 2.82%, meeting the requirements of low-carbon economic development.