Networked Microgrids for Enhancing Grid Resiliency
By Zhaoyu Wang and Kaveh Dehghanpour
This article provides a summary of using networked microgrids to enhance power system resiliency. It will discuss two types of applications: the first one is for a power system with existing networked MGs, and the second one is dynamically forming networked microgrids. The article will also discuss using networked microgrids for pre-event preparation.
Microgrids (MGs) are small-scale power distribution systems integrating renewable energy, which can be operated in grid-connected or islanded modes. One important feature of microgrids is “self-adequacy,” i.e., loads within a microgrid can be supported by its local distributed generators continuously, which enables the MG to be disconnected from its upstream macro-grids during extreme events or contingencies. This salient feature significantly enhances grid resiliency by providing non-interruptible electricity. Connecting multiple MGs to form a distribution system with networked MGs can maximize the resiliency benefits. During emergencies, these MGs can support each other to ensure continuous power supply; while in normal operations, they can exchange electricity to achieve efficient and economical operation. Two major types of applications exist for enhancing grid resiliency using networked MGs: the first one is for power systems with existing networked MGs (i.e., MGs with pre-defined boundaries), and the second one is dynamically forming networked MGs by sectionalizing distribution grids through reconfiguration (i.e., networked MGs with non-stationary boundaries).
Coordinating existing networked MGs in distribution grids is an indispensable task for ensuring both the economic efficiency and system resiliency. This is especially critical for a stand-alone distribution grid that cannot receive support from transmission systems due to extreme weather events. Our vision is that these MGs can exchange power with each other or with upstream grids for economic benefits in normal operations, while supporting each other during outages. The challenge is to ensure real-time balancing of supply and demand within individual MGs and throughout the entire system, while maintaining voltage and line flow constraints. This challenge is further complicated by inherent uncertainties of distributed renewable resources and customer loads as well as the interdependency among MGs. This problem can be formulated as a centralized optimization model assuming there exists one decision maker that can control all MGs. However, these MGs may have different owners with various objectives. Therefore, another approach is to model this problem using game theory and find the equilibrium for all MGs. The centralized methods are easy to formulate and have the potential to achieve global optimality, but they may suffer from computational scalability and privacy issues. The networked MG coordination can also be modeled as a decentralized optimization problem where each MG makes their own decisions with minimum information exchange. In a recent research, we have proposed a reinforcement learning-based approach to coordinate networked MGs under incomplete information. This approach allows the networked MGs to adapt to evolving system conditions by observing the impacts of their control actions on the grid with minimal reliance on a centralized coordinator.
As suggested by IEEE Standard 1547.4 splitting distribution grids into multiple self-adequate MGs can enhance operation and reliability. The MG formation can facilitate system restoration during and after extreme weather events through fault containment, critical load protection, and expedient load pick-up. However, ensuring steady-state supply sufficiency within newly formed MGs while maintaining the system-wide voltage/frequency stability during the sectionalization is challenging. Deficient MG formation can result in unnecessary isolation of whole grid zones and excessive load shedding. Recently, we developed a comprehensive self-healing strategy to optimally and smoothly sectionalize distribution grids into autonomous MGs through intelligent system reconfiguration, which provides continuous power to the maximum number of customers during extreme events, while at the same time isolating faulty areas of the grid. Our solution leverages stochastic programming techniques to enable each formed MG to adequately supply its internal load and mitigate the uncertainties of local distributed generation resources within the dynamic MG boundaries.
In addition to post-event response applications, networked MGs can be leveraged in the pre-event stage. This entails pre-dispatching existing networked MGs or pre-forming networked MGs before extreme events strike based on weather forecasts, to better protect critical loads and provide sufficient safety margins against probable future events. Further, this MG-enabled preparation can facilitate post-event restoration as well. The challenge in the preparation problem is its inherent stochasticity, as the event intensity, damage and outage locations, and the post-event state of the grid are unknown beforehand. A poorly-designed preparation plan will not only fail to enhance resilience but can also lead to costly and conservative operation. Research shows that proactive management of MGs prior to extreme events can effectively minimize grid vulnerability and reduce the chances of blackout during the event. According to our recent study, multi-stage probabilistic mathematical models show significant promise in mitigating the uncertainties of disaster preparation without resorting to over-conservative decisions, while also radically reducing the costs of post-event restoration and loss of load.
Current research demonstrates the irrefutable benefits of self-adequate networked MGs in alleviating the potential disastrous impacts of extreme events on energy distribution sector; these benefits include, maintaining continuous supply of power to customers to the extent that is not possible in conventional grids, fast and economical service restoration after events, and enabling compartmentalized and distributed decision making for addressing specific needs of customers located in different areas of the grid using sensor data. However, these benefits come at the expense of additional investment costs in communication infrastructure, local control and data management facilities, and customer acceptance, which are obstacles against large-scale adoption of MGs. Future research should facilitate a better understanding of the significance of resilience at distribution sector for both customers and utilities, and convey the necessity and economic feasibility of networked MGs in resolving complex challenges. This can pave the way for promoting networked MGs as building blocks of future resilient energy distribution systems.
Zhaoyu Wang (S’13-M’15) is the Harpole-Pentair Assistant Professor with Iowa State University. He received the B.S. and M.S. degrees in electrical engineering from Shanghai Jiaotong University in 2009 and 2012, respectively, and the M.S. and Ph.D. degrees in electrical and computer engineering from Georgia Institute of Technology in 2012 and 2015, respectively. He was a Research Aid at Argonne National Laboratory in 2013 and an Electrical Engineer Intern at Corning Inc. in 2014. His research interests include power distribution systems and microgrids, particularly on their data analytics and optimization. He is the Principal Investigator for a multitude of projects focused on these topics and funded by the National Science Foundation, the Department of Energy, National Laboratories, PSERC, and Iowa Energy Center. Dr. Wang is the Secretary of IEEE Power and Energy Society (PES) Award Subcommittee, Co-Vice Chair of PES Distribution System Operation and Planning Subcommittee, and Vice Chair of PES Task Force on Advances in Natural Disaster Mitigation Methods. He is an editor of IEEE Transactions on Power Systems, IEEE Transactions on Smart Grid, IEEE PES Letters and IEEE Open Access Journal of Power and Energy, and an associate editor of IET Smart Grid.
Kaveh Dehghanpour received his B.Sc. and M.S. from University of Tehran in electrical and computer engineering, in 2011 and 2013, respectively. He received his Ph.D. in electrical engineering from Montana State University in 2017. He is currently a Postdoctoral Research Associate at Iowa State University. His research interests include machine learning and data mining for monitoring and control of active energy distribution grids.