Stochastic Unit Commitment and Self-scheduling: A Review Considering CO₂ Emission Modeling

Abstract

Short term power system operation has traditionally been planned using optimization models that represent a cost minimization problem referred to as the unit commitment problem. Computational advances observed from the 1970s have made possible to solve increasingly complex and large scale formulations of this model, which usually are mixed-integer problems. More recently, the restructuring of many electricity markets in several countries has introduced competition among generators, and increased the interest in incorporating uncertainty into the unit commitment problem. Typical sources of uncertainty can be future demand, fuel prices, and unit availability. However, under competition, electricity spot price has also become an important uncertain variable, and has forced electric power generation companies to focus on operation strategies that may lead to the maximization of benefits. Short-term operation planning is not exempted from this shift and this has lead to the self-scheduling problem, where individual companies want to determine the unit commitment that would maximize their benefits. In this article, we review recent advances for the stochastic unit commitment and self-scheduling problems. After reviewing several recent modeling approaches we consider the relevance of environmental considerations, particularly CO₂ emissions, and review relevant propositions for modeling and incorporating emission constraints into the unit commitment problem.