A STOCHASTIC PROGRAMMING APPROACH TO QUANTIFYING EFFECTS OF CONTINGENCIES ON LOCATIONAL MARGINAL PRICES

摘要

This paper focuses on quantifying effects of contingencies and network parameter changes on variations of Locational Marginal Prices (LMPs) in restructured power markets. The process of planning for contingencies typically results in the use of more expensive facilities (spinning and/or non-spinning reserves) and/or generation curtailment/load shedding, leading to different prices and energy availability at the various parts of the network. In this paper we explore a stochastic programming (SP) model for incorporating contingency and parameter uncertainties, and maximizing the expected social welfare. SP is compared with the conventional security-constrained deterministic programming (DP) model that includes a pre-specified list of outages (and disregards parametric uncertainties). We provide a probabilistic interpretation of DP model that may incorporate post-fault recourse activities, like load shedding and generation adjustment (reduction or utilizing the reserves).SP model with random uncertainties (created by a Monte Carlo sampling technique) is solved as a two-stage (recourse function (RF) based) problem, using an interior-point log-barrier method coupled with Benders decomposition. The model is verified on the IEEE medium-scale example (with 300 nodes/411 branches).