Assessing Robustness of Risk-Constrained Operating Strategies for Power Systems with Renewables by Contamination-Based Technique

Abstract

Risk-constrained stochastic programming (SP) is an effective tool to cope with the increasing uncertainty renewable energy resources (RESs) bring in look-ahead dispatch (LAD). However, since the solutions of SP are dependent on characteristics of uncertain RESs, inaccurate predication of RESs significantly influences the LAD strategies. This paper proposes a contamination-based technique (CBT) to evaluate the robustness of a dispatch strategy against inaccurate RESs predication. Without loss of generality, the proposed CBT is used in a two-stage CVaR-constrained stochastic program, where the first stage optimizes a strategy based on the original inaccurate predication and the second stage corrects this strategy by contaminating the first-stage inaccurate predication with updated information. The robustness of a strategy against the inaccurate predication is quantified by this method and, furthermore, the sensitivities of some critical parameters, such as penetration level of RESs and system flexibility, are analyzed. Case study validates the feasibility of proposed method.