Elsevier

Applied Energy

Volume 210, 15 January 2018, Pages 1251-1265
Applied Energy

Enabling resilient distributed power sharing in networked microgrids through software defined networking

https://doi.org/10.1016/j.apenergy.2017.06.006Get rights and content

Highlights

  • An SDN-enabled control and communication architecture is established for NMGs.

  • A resilient distributed power sharing control strategy is devised for NMGs.

  • Novel event-triggered communication is deployed through an SDN architecture.

  • A cyber-physical HIL testbed is built to validate NMGs’ control and communication strategies.

Abstract

Networked Microgrids (NMGs) offer a new, more resilient alternative to traditional individual Microgrids (MGs). Even though networking existing microgrids presents clear advantages, the scalable and resilient communication and control infrastructure necessary for supporting this innovation does not yet exist. This paper addresses this challenge by developing a Software-Defined Networking (SDN) enabled architecture that can achieve fast power support among microgrids, transforming isolated local microgrids into integrated NMGs capable of achieving the desired resiliency, elasticity and efficiency. Equipped with a novel event-triggered communication scheme, the SDN-based architecture enables distributed power sharing among microgrids in both the transient period and the steady state, a capability that is unattainable using existing technologies. Extensive experiments on a cyber-physical Hardware-in-the-Loop (HIL) NMGs testbed have validated the effectiveness and efficiency of the SDN-enabled distributed power sharing method.

Introduction

A microgrid normally refers to a localized autonomous distribution network designed to supply electrical and heat loads for a local community (e.g., a university campus [1], a commercial building [2] or a residential area [3]). It can be connected with the main grid (grid connected mode) or isolated during main grid emergencies (islanded mode). Because microgrids offer the following benefits, they have attracted increased interest in the last few years: they enable integration and coordination of renewable energy resources; they enhance the resilience of electrical system for customers; and they reduce economic and emission costs [4]. These benefits are particularly important given the rapid development of power electronics technologies as well as primary, secondary and tertiary control techniques in recent years [5].

The swift growth of microgrid research and development are leading to increased penetration of microgrids [6]. For instance, in urban areas where populations and critical loads are concentrated, microgrids are being increasingly deployed. A smart city (or smart and connected communities) zone is expected to have many microgrids operated by various stakeholders. It is therefore natural to ask whether coordinated networked microgrids can offer a more resilient system than individual microgrids. Indeed, our preliminary research [7] shows that, when local microgrids are networked, this not only enables faster distribution grid recovery during a main grid blackout but also significantly improves the system’s day-to-day reliability. In fact, the U.S. Department of Energy anticipates that researching and developing of networked microgrids will usher in the next wave of smart grid technology. This innovative approach promises to meaningfully modernize the nation’s grid system in response to issues such as climate change and the need for greater grid resilience [8].

Though networking existing microgrids offers a number of advantages, there is one major challenge that has not been addressed: a scalable and resilient communication and control infrastructure does not yet exist. Furthermore, given the standardized control architecture of individual microgrids (e.g., droop control, secondary control), it is highly desirable to establish a network-level control architecture that does not significantly modify the communication and control layers in individual microgrids. This paper aims to address these challenges by developing an SDN-enabled architecture that can readily network microgrids at the cyber layer in a simple and economically efficient way, transforming isolated local microgrids into integrated smart microgrids capable of achieving the desired resiliency, elasticity and efficiency. In particular, this novel method will enable a provably correct and previously unattainable distributed power sharing among microgrids in both the transient period and the steady state.

Related work: Networked microgrids, or coupling microgrids, can be defined as a cluster of microgrids interconnected in close electrical or spatial proximity with coordinated energy management and interactive support and exchange. Recently, the feasibility of coupling microgrids through common AC buses [9], utility feeders [10] and DC links [11] has been discussed. Ref. [12] presents a power dispatch strategy for maintaining islanded microgrids’ power balances through microgrid generation reallocation triggered by power deficiency events. Ref. [13] presents the use of networked microgrids to improve the self-healing of the distribution network under power outages, where microgrids are designed to pick up external loads with minimum switch operations. Further, an economic dispatch strategy for networked microgrids is developed [14], where the surplus capacities in individual microgrids are aggregated to fulfill the power requirements. Apart from facilitating power system restoration, networked microgrids can also participate in global frequency regulation by providing extra frequency control reserves [15]. The interaction between the distribution network operator and local microgrids has also been investigated [16], [17]. These studies focus on the longer-term coordination of interconnected microgrids at a time scale of minutes, hours or longer. In the real world, however, microgrids usually have low inertia and intermittent renewable generation. Thus, it is critically important to ensure fast power sharing while maintaining transient stability in networked microgrids. In order to adequately control such a complex system, a high-speed, resilient cyber infrastructure is indispensable, but this remains an open challenge.

In networked microgrids, one of the most important functions is to share power demands among the networked Distributed Generators (DGs). Power sharing in a single microgrid is achieved in tandem with voltage and frequency recovery either in a centralized or a distributed way [18], [19]. The latter has been attracting more attentions in recent years due to the potential benefits of avoiding the single point of failure and reducing communication overhead [20], [21]. In [22], a distributed control requiring only local communication is presented, which is capable of achieving proportional active power sharing and frequency restoration. This paper also identifies the conflict between voltage control and reactive power sharing for DG units with a droop-based primary control. An alternative approach for fast voltage recovery without considering reactive power sharing is developed in [23]. Among various distributed power sharing schemes, the Average Consensus Algorithm (ACA) is a popular choice for solving the problem in a fully distributed fashion. ACA, however, can compromise network resilience by requiring continuous intensive data transmissions which may cause bandwidth shortage, congestion, and processor overuse. Moreover, there is a lack of distributed power sharing schemes for networked microgrids in the existing literature.

Our contributions: To enable resilient networked microgrids and close the aforementioned gaps, we are introducing a novel SDN-based cyber architecture with a distributed event-triggered communication scheme. The authors have pioneered the use of SDN in enabling resilient microgrids [24] by devising a novel SDN-based cyber architecture for individual microgrids and developing SDN functions such as delay management, automatic failover, and traffic prioritization. The unprecedented flexibility and dynamic programmability of SDN [25], [26], [27] supports on-the-fly network updates and enables the interoperability of local microgrids. Therefore, the SDN-based architecture in [24] is further expanded to enable networked microgrids. We also integrate the event-triggered communication in the SDN-based communication architecture such that a microgrid only shares information with its neighbors when the specific states exceed predefined thresholds. Recent research into networked control systems has mathematically proven the effectiveness of the event-triggered communication in enabling more efficient and robust ACAs [28], [29], [30]. This paper makes three main contributions:

  • It devises a layered cyber and control architecture that supports the plug-and-play of networked microgrids. The local layer includes the primary and secondary controllers within individual microgrids while the global layer is responsible for the dynamic power sharing among different microgrids. This new architecture requires little modification in individual microgrids and enables seconds-level fast power support among microgrids.

  • It develops the SDN-enabled event-triggered communication scheme in the global layer. Power deficiency and its recovery are defined as “events” which are detected locally in the DGs and sent to the SDN controller. Once it receives these requests, the SDN controller will use an electrical distance matrix to find the closely coupled microgrids cluster and update the corresponding communication network. In this way, global data transmission is only required during triggered periods, which significantly reduces communication costs and enhances the system’s resiliency.

  • It builds a cyber-physical HIL testbed that validates the effectiveness and efficiency of the new architecture and the distributed power sharing method for networked microgrids. The new testbed will serve as a powerful instrument for developing advanced analytics and controllers for future networked microgrid research.

The remainder of the paper is organized as follows: Section 2 presents the distributed power sharing control for networked microgrids. Section 3 elaborates the SDN-enabled event-triggered communication methods. Section 4 introduces a cyber-physical HIL NMGs testbed, and Section 5 summarizes the experimental results that validate the effectiveness of the distributed power-sharing scheme for NMGs. Finally, Section 6 concludes the whole paper.

Section snippets

Distributed power sharing for networked microgrids

It is assumed that, in islanded NMGs, each MG is equipped with only local controllers (LCs) on inverter-interfaced DGs. Droop controllers are used as the primary control for automatically adjusting its power output under demand changes. To achieve local power sharing, the distributed-averaging proportional-integral (DAPI) control [22] is applied to these LCs as a secondary control due to its flexibility and scalability. In this section, a droop control and a local power-sharing algorithm are

SDN-enabled event-triggered communication

This section outlines how SDN-enabled and event-triggered communication is designed for global layer power sharing among selected microgrids with close electrical distances. Studies have shown that large communication latency in microgrid operation with a centralized secondary control can cause undesired control deviations and even stability concerns [35], [36]. Also, the performance of the ACA with communication delays is discussed in [37], showing that the ACA only converges when the

Hardware-in-the-loop testing environment

This section will review the design and establishment of a cyber-physical HIL testbed to provide a realistic testing environment. We begin by introducing the high-level design of the cyber-physical HIL testbed, and we then specify the hardware components, NMGs models, and development of the SDN network.

Experimental results

The NMGs’ communication and distributed control strategy is tested on our cyber-physical HIL testbed. Different communication conditions are examined for single-event cases. The results show that the combination of global power sharing control, K-NN microgrid clustering, and SDN-based event-triggered communication has the best performance considering the communication cost and system response. Multiple-event cases then demonstrate the robustness of the SDN controller in dealing with a series of

Conclusion

This paper pioneers the use of SDN in NMGs by leveraging the programmability and flexibility of the SDN architecture to enable highly resilient NMGs. A layered power-sharing scheme is developed for NMGs, supported by SDN-based event-triggered communication. The method is fully distributed and only requires an additional global power-sharing block on the local controller of the leader DGs. To further minimize the communication cost, a K-NN microgrids set is selected using electrical distance

Acknowledgement

This material is based upon work supported by the National Science Foundation under Grant Nos. 1647209 and 1611095.

We would like to thank Emma Burris-Janssen for proofreading the manuscript.

References (40)

  • J.W. Simpson-Porco et al.

    Synchronization and power sharing for droop-controlled inverters in islanded microgrids

    Automatica

    (2013)
  • P.A. Bliman et al.

    Average consensus problems in networks of agents with delayed communications

    Automatica

    (2008)
  • C. Marnay et al.

    Optimal technology selection and operation of commercial-building microgrids

    IEEE Trans Power Syst

    (2008)
  • J.M. Guerrero et al.

    Hierarchical control of droop-controlled AC and DC microgrids – a general approach toward standardization

    IEEE Trans Ind Electron

    (2011)
  • Z. Bie et al.

    Reliability evaluation of active distribution systems including microgrids

    IEEE Trans Power Syst

    (2012)
  • Cheung K. DOE perspective on microgrids. IEEE appl power electron conf expo; 2015....
  • F. Shahnia et al.

    Autonomous operation of multiple interconnected microgrids with self-healing capability

    IEEE Power Energy Soc. Gen. Meet.

    (2013)
  • R.H. Lasseter

    Smart distribution: coupled microgrids

    Proc IEEE

    (2011)
  • M.J. Hossain et al.

    Design of robust distributed control for interconnected microgrids

    IEEE Trans Smart Grid

    (2016)
  • J. Li et al.

    Distribution system restoration with microgrids using spanning tree search

    IEEE Trans Power Syst

    (2014)
  • Cited by (0)

    View full text