Toward economic NMPC for multistage AC optimal power flow

Abstract

Summary Recently, there has been a considerable progress on the analysis of stability and performance properties of so‐called economic nonlinear model predictive control (nmpc) schemes, ie, schemes employing stage costs that are not directly related to distance measures of precomputed setpoints. At the same time, with respect to the energy transition, the use of nmpc schemes is proposed and investigated in a plethora of papers in different contexts. For example, receding‐horizon approaches to generator dispatch problems, which is also known as multistage optimal power flow (opf) in power systems, naturally lead to economic nmpc schemes based on nonconvex discrete‐time optimal control problems (ocp). This paper investigates the transfer of analysis results available for general economic nmpc schemes to receding‐horizon multistage opf. We propose a blueprint formulation of multistage opf including ac power flow equations. Based on this formulation, we present results on the dissipativity and recursive feasibility properties of the underlying ocp. Finally, we draw upon simulations using a 5‐bus system and a 118‐bus system to illustrate our findings.

A distributed optimal power management system for microgrids with plug&play capabilities

“…However, gathering data from all over the microgrid is not practical for large microgrids. Several centralized power management schemes for microgrids have been proposed in the literature, for example, References 3‐11.…”

“…Note that the proposed distributed scheme addresses all elements mentioned in Problem 1, while none of the above‐cited work can address them all together. More specifically, the centralized schemes in References 3‐11 are not scalable, do not provide a plug&play capability, and require a large communication overhead; the decentralized schemes in References 12‐17 do not provide an optimal solution; and the distributed schemes in References 18‐23 have a high communications cost compared to the proposed scheme. It should be noted that in the proposed distributed scheme each subsystem communicates only with a sensor, while in the above‐cited distributed schemes every subsystem communicates with (some of) its neighbors.…”

Editorial Model Predictive Control for Energy Systems: Economic and Distributed Approaches

“…The second group of articles considers the application of MPC to problems arising in electrical power systems such as multiperiod power flow problems, smart grids, and induction motors …”

Computationally efficient methods for solving optimal power flow problems by exploiting supplementary information

“…From the viewpoint of MPC, the main difference between these two class of problem is the time-scale or horizon-time, where the time-scale of energy management problem is about 15 minutes to 1 hour while the time-scale of the stabilization problem is less than few seconds [200]. Hierarchical MPC approaches are also presented to handle both time-scales and solve a comprehensive control and management system [201], e.g.…”

“…While most of the literature on multi-period OPF problem benefits from convex approximations or relaxations, the authors in [214] implement the MPC for solving the multi-period ACOPF problem considering renewable resources. Going beyond [214], the authors in [200] proposes an economic nonlinear MPC scheme to solve the OPF problem including storage desires besides the renewable resources.…”