Data-driven Convexification for Frequency Nadir Constraint of Unit Commitment

Abstract

The increasing penetration of the renewable energy sources brings new challenges to the frequency security of power systems. In order to guarantee the system frequency security, frequency constraints are incorporated into unit commitment (UC) models. Due to the non-convex form of the frequency nadir constraint which makes the frequency constrained UC (FCUC) intractable, this letter proposes a revised support vector machine (SVM) based system parameter separating plane method to convexify it. Based on this data-driven convexification method, we obtain a tractable FCUC model which is formulated as a mixed-integer quadratic programming (MIQP) problem. Case studies indicate that the proposed method can obtain less conservative solution than the existing methods with higher efficiency.