The evolving landscape of electric power networks, influenced by the integration of distributed energy resources require the development of novel power system monitoring and control architectures. This paper develops ...
详细信息
ISBN:
(纸本)9798350381849;9798350381832
The evolving landscape of electric power networks, influenced by the integration of distributed energy resources require the development of novel power system monitoring and control architectures. This paper develops algorithm to monitor and detect anomalies of different parts of a power system that cannot be measured directly, by applying neighboring measurements and a dynamic probing technique in a distributed fashion. Additionally, the proposed method accurately assesses the severity of the anomaly. A decision-making algorithm is introduced to effectively penalize anomalous agents, ensuring vigilant oversight of the entire power system's functioning. Simulation results show the efficacy of algorithms in distributed anomaly detection and mitigation.
This paper presents distributed multi-agent deep reinforcement learning (MADRL) approach for optimizing powerflow management in networked microgrids within distribution systems. In contrast to centralized training me...
详细信息
ISBN:
(纸本)9798350381849;9798350381832
This paper presents distributed multi-agent deep reinforcement learning (MADRL) approach for optimizing powerflow management in networked microgrids within distribution systems. In contrast to centralized training methods, our proposed approach leverages the multi-agent trust region policy optimization (MATRPO) algorithm to learn distributed policies that minimize operational costs while considering powerflow constraints in distribution networks. We model the cooperation among networked MGs as a partially observable Markov game and learn the distributed policies via peer-to-peer communication to address the challenges associated with centralized training. This approach offers scalability and efficiency benefits. The proposed approach is evaluated on a 24.9 kV distributed network with interconnected 4.16 kV MGs based on modified IEEE-34 and IEEE-13 bus systems. By comparing our approach with state-of-the-art MADRL methods, we demonstrate its effectiveness in enabling cooperative optimization of networked MGs, showcasing its applicability for managing distributed energy systems.
The recent interest in distributed control strategies is motivated by the fact that the future electric power grid is expected to accommodate significantly more distributed resources. This future distributed infrastru...
详细信息
ISBN:
(纸本)9781467380409
The recent interest in distributed control strategies is motivated by the fact that the future electric power grid is expected to accommodate significantly more distributed resources. This future distributed infrastructure requires more autonomous control schemes for its scattered components, e. g., powerflow control devices as considered in this paper. To this end, this paper presents a distributed control scheme for solving the DC optimalpowerflow problem for an electricity network in which transmission lines are potentially instrumented with distributed Flexible AC Transmission Systems (D-FACTS). Our approach constitutes a distributed iterative procedure which aims to solve the first order optimality conditions of the corresponding optimization problem. In this iterative algorithm, the responsibility of each bus includes updating a few local variables and exchanging updated variables with a few neighboring buses. Specifically, the updates at each bus include local innovation terms that are derived from the first order optimality conditions and preserve the coupling between the neighboring Lagrange multipliers and enforce the supply/demand balance. Due to the non-convexity of the primal problem, global optimality is not guaranteed, however, our results show that the distributed procedure converges to a solution of the first order optimality conditions yielding a local optimum. The performance of this algorithm is evaluated using the IEEE 14-bus and the IEEE-118 bus test systems.
In the future multi-terminal DC (MTDC) distribution grids, with high penetration of converter-interfaced distributed energy resources (DER), the system-level control is expected to be realised through a fully distribu...
详细信息
In the future multi-terminal DC (MTDC) distribution grids, with high penetration of converter-interfaced distributed energy resources (DER), the system-level control is expected to be realised through a fully distributed optimal power flow (OPF) algorithm that determines the voltage and power set-points for the converters. The OPF problem is divided to nodal OPF sub-problems, which are solved by distributed control units (DCUs) operating as a networked control system (NCS). This paper presents a fully-distributed OPF algorithm with an incorporated strategy, to survive failures in the NCS, such as permanent changes in the NCS topology due to the crash of a DCU or the loss of a communication link. By changing the constraints of the OPF sub-problems of the affected DCUs upon the detection of the failure in the NCS, the fully-distributed algorithm becomes tolerant to permanent changes in the NCS topology and converge to a feasible OPF solution. Simulation results in different IEEE networks and failure scenarios demonstrate the fast convergence, irrespective of the number and location of the failures in the NCS or the failure time-point during the execution of the algorithm.
This paper explores the feasibility of fully-distributed architectures for the electric power industry. In such architectures, the various electric power ecosystems interact with each other to achieve their own operat...
详细信息
ISBN:
(纸本)9781457721588
This paper explores the feasibility of fully-distributed architectures for the electric power industry. In such architectures, the various electric power ecosystems interact with each other to achieve their own operational and economic objectives without the need for real-time centralized coordination or optimization. This paper uses an efficient distributed optimal power flow algorithm which is applicable to generic power ecosystem optimization, and it describes the ecosystem interaction protocols for operation and trading. The proposed algorithm is illustrated in detail for a small case, and performance results are provided for a large, realistic-size case. The results provided demonstrate the feasibility of such a distributed architecture that can ultimately scale from interconnections to homes.
暂无评论