• Energy Research

Publisher Copyright: IEEE With the ever-escalating scale of urban distribution networks (UDNs), the traditional model-based reconfiguration methods are becoming inadequate for smart system control. On the contrary, the data-driven deep reinforcement learning method can facilitate the swift decision-making but the large action space would adversely affect the learning performance of its agents. Consequently, this paper presents a novel multi-agent deep reinforcement learning method for the reconfiguration of UDNs by introducing the concept of 'switch contribution'. First, a quantification method is proposed based on the mathematical UDN reconfiguration model. The contributions of controllable switches are effective quantified. By excluding the controllable switches with low contributions during network reconfiguration, the dimensionality of action space can be significantly reduced. Then, an improved QMIX algorithm is introduced to improve the policy of multiple agents by assigning the weights. Besides, a novel two-stage learning structure based on a reward-sharing mechanism is presented to further decompose tasks and enhance the learning efficiency of multiple agents. In the first stage, agents control the switches with higher contributions while switches with lower contributions will be controlled in the second stage. During the two-stage process, the proposed reward-sharing mechanism could guarantee a reliable UND reconfiguration and the convergence of our learning method. Finally, numerical results based on a practical 297-node system are performed to validate our method's effectiveness. Peer reviewed

Reliability enhancement is a major motivation for integrating electric and gas energy distribution systems. Compared with reliability assessment at the stage of planning, which emphasizes calculating the system annual average indices, the coupling of gas and electric energy flows contributes more to the operational reliability of the integrated system, especially during peak-load periods. In this paper, a method is proposed that assesses system outage risk in each time period by modeling the consequence of the worst N-1 fault in the integrated electric-gas system. In order to accurately model the load transfer process, a novel dynamic load restoration model is applied to minimize load loss of the system, where the fluctuation of multiple energy flows during fault clearing periods is considered. The effectiveness of the proposed approach is validated by performing case studies.

Transient stability assessment is playing a vital role in modern power systems. For this purpose, machine learning techniques have been widely employed to find critical conditions and recognize transient behaviors based on massive data analysis. However, an ever increasing volume of data generated from power systems poses a number of challenges to traditional machine learning techniques, which are computationally intensive running on standalone computers. This paper presents a MapReduce based high performance neural network to enable fast stability assessment of power systems. Hadoop, which is an open‐source implementation of the MapReduce model, is first employed to parallelize the neural network. The parallel neural network is further enhanced with HaLoop to reduce the computation overhead incurred in the iteration process of the neural network. In addition, ensemble techniques are employed to accommodate the accuracy loss of the parallelized neural network in classification. The parallelized neural network is evaluated with both the IEEE 68‐node system and a real power system from the aspects of computation speedup and stability assessment.

The uncertain nature of renewable energy resources (RERs) and fast demand response lead to recurring voltage violations in the power systems, which causes frequent transformer tap shifting and capacitor switching. Therefore, this paper resorts to the static var compensators (SVCs) to manage the bus voltages based on the multi-agent deep reinforcement learning (MA-DRL) algorithm. The proposed scheme includes several system agents and SVC agents to collaboratively adjust the injected reactive power to restrict the bus voltages within the normal range. All the agents are trained centrally and executed separately, which requires minimum communication cost. The IEEE 14-bus system, IEEE 300-bus system, and China 157-node urban power grid are used to verify the effectiveness of the proposed method.

The fast urbanization process in developing countries creates an electricity load surge that poses serious congestion problems to the urban power grid (UPG). The current practice of system operators is to reconfigure the topological structure of the 110-kV high voltage distribution network (HVDN) in an attempt to adjust the load distribution and eliminate the UPG congestion. However, HVDN reconfiguration is a large scale, non-convex and non-linear optimization problem that is intractable to be solved in an efficient manner. In order to reduce the computing cost and overcome the difficulty in the solving process, new partition rules are proposed to divide the HVDN into multiple smaller regions. The topological structure of each region can be optimized either independently or interdependently based on the condition of the UPG. The solving process is modeled as a bi-level iterative line switching problem. At the upper level, the operational boundary (OB) of each regional HVDN is calculated based on a set of convex ACOPF constraints. At the lower level, the line switching scheme of each regional HVDN is derived based on the corresponding mixed integer second-order cone programming (MISOCP) model considering the OB. Then, the line switching result is examined by the AC power flow to identify whether there is any violation or not. This iterative process continues until all the violations are corrected in the UPG. The proposed method was validated by a modified IEEE 58-node test system and a practical 157-node UPG system. The simulation results showed that the proposed method is able to achieve a higher computational efficiency than the traditional method so it is promising for online applications.

With rapid development of the electric vehicle (EV) industry, charging infrastructures are built fast. However, the unreasonable deployments with increasing EVs contribute to a long queuing time for charging demand of EVs, especially in the peak hours. How to navigate a specific EV to economically satisfy its charging demand, while relieve the traffic burden, is an urgent problem. To address that, a price incentive-based charging navigation strategy for EVs is proposed. Unlike previous charging navigation studies that mainly focus on the EVs-transportation-power systems modeling, it considers the spatial-temporal influence of EVs’ charging decision, especially the simultaneous charging requests. Specifically, the charging navigation framework with the collaborative working mode of EV-charging station-information exchange center-intelligent transportation system is established first. Following this, spatiotemporal distribution of the charging demand is obtained through the origin–destination analysis. After this, an event-driven dynamic queue model is constructed. It contributes to the modeling of the charging strategy, together with the proposed reservation opportunity cost mechanism. Finally, the simulation results indicate that the presented charging navigation strategy can not only reduce the EV's charging cost but also improve the utilization rate of charging facilities, which verify its effectiveness.