Suppressing ineffective control actions in optimal power flow problems

Abstract

Many utilities are still reluctant in adopting optimal power flow (OPF) tools for decision‐making in operation. This paper scrutinizes this issue from the perspective of whether all control actions proposed by an OPF are truly effective to an operator. To this end, the paper focuses on suppressing ineffective control actions in OPF problems. This goal is aligned with the meaning of optimization in practice, that is improvement of operation performance of slightly noisy or imperfectly known real world models. The paper proposes a conceptually different new approach, which computes automatically the number of effective control actions that do not worsen the ideal model OPF objective by more than an operator‐specified tolerance. The proposed approach relies on a three‐step methodology that solves different OPF problems, in which smooth continuous approximation functions are used to convert the benchmark mixed integer nonlinear programming (MINLP) problems into nonlinear programming (NLP) problems. The proposed approach is compared with two other alternatives for the OPF problem of thermal congestion management using three test systems of 60, 118, and 2746 buses, respectively. The results show that, among the competing approaches, the solutions of the proposed continuous approximation lead to the best trade‐off between sub‐optimality and computation speed.

On the robustness of machine-learnt proxies for security constrained optimal power flow solvers

“…IV. TEST PROBLEM SETUP Our test system is a modified version [25] of the Swedish transmission network, i.e., the Nordic32 electric grid from [26] (refer to Fig. A1 and Fig.…”

Topology-aware reinforcement learning for tertiary voltage control

“…At first, three distinct variants 4 (namely Standard, Condenser and Reduced) were crafted from the case60nordic test case [24] as detailed in Figure 5, where all voltage limits are set to V = 0.9 p.u. and V = 1.1 p.u.…”