You are here


Title: TAG: A Timing Adaptive Grouping Protocol for Smart Grid Communications.
Name(s): Cai, Ziyuan, author
Yu, Ming, professor co-directing dissertation
Li, Hui, 1970-, professor co-directing dissertation
Meyer-Baese, Anke, university representative
Steurer, Michael Morten, committee member
Andrei, Petru, committee member
Florida State University, degree granting institution
College of Engineering, degree granting college
Department of Electrical and Computer Engineering, degree granting department
Type of Resource: text
Genre: Text
Issuance: monographic
Date Issued: 2015
Publisher: Florida State University
Place of Publication: Tallahassee, Florida
Physical Form: computer
online resource
Extent: 1 online resource (79 pages)
Language(s): English
Abstract/Description: The prospect of SG is green, power efficient, and economical to its customers. Many emerging innovations have reached a consensus that the traditional power grids need to be combined with modern data networks, in order to establish a new platform that supports distributed renewable energy devices, electrical measuring sensors, and intelligent energy management and control systems, etc. For example, an energy management system is proposed to connect data aggregators with renewable energy devices in the network area. A wireless sensor network is used to provide the communications between SG data centers and consumers, and manage residential energy with an optimization-based scheme. In SG, the stability of an energy management scheme becomes heavily dependent on accurate real-time communications among intelligent energy management agencies in residential homes, micro-grids, and main grids. Within a large-scale distributed (or centralized) smart grid (SG), the communication network is designed to connect multiple power management systems and collect data from hundreds or thousands of power sensors over a wide geographical area. One dominant feature of innovated SG communication network is that one power device is coupled with a single Ethernet or non-Ethernet communication agent to exchange control state or management information with others. Generally speaking, an intelligent agent helps its corresponding power device negotiating with other peers to dynamically form an ad-hoc group through the data network infrastructures. Then, many meaningful power management algorithms are operated in the logical group. The grouping topology recognized by a specific agent needs to reform when the participating group members can not satisfy the demand from operating power management algorithms. Upon our ad-hoc ideas, the main problem arises: in a changed group, the networking size, traffic load, queueing effect and security requirement are varied, so an agent experiences different communication cost over group reforming. We define such inevitable difference as communication inconsistency of SG ad-hoc grouping. If the timeout parameters of communication control are set statically in grouping procedures, the inconsistency definitely triggers the timeout, crashes the group and aborts the running cycle of power management algorithms very often. Thus in this work, an adaptive timing solution is developed for connecting distributed intelligent agents in ad-hoc manner to greatly enhance the flexibility and performance of grouping algorithms in SG communication network. A timing adaptive grouping (TAG) protocol is proposed to make every distributed agent capable of adjusting its operational timing configurations (OTCs) in pace with the changing of ad-hoc groups, so that prevents the harmfulness of communication inconsistency to the stability of grouping procedures. More specifically, we first develop a set of queueing model to describe the network traffic of various power management applications among distributed agents in connection with different scale of ad-hoc grouping topologies. Second, the security cost of SG communications is modeled, estimated and validated with various grouping agents' characteristics. Third, based on the network grouping model including both queueing and security cost, we analyze the ad-hoc delay performance and illustrate that the model can be used to predict the average operating delays of networking agents. Fourth, based on the delay parameters derived from previous modeling, the TAG protocol is developed with our Smart Timing Adaptive (STA) algorithm to facilitate each distributed agent dynamically judging variant ad-hoc grouping conditions. Finally, we have implemented a validation testbed with the capabilities of integrated real-time communication and power exchange to demonstrate the ad-hoc grouping operation of SG power management applications. Due to the ripple effect of inconsistent communication delays among the ad-hoc SG groups with dynamic changing topology, the network performance becomes a major concern to support power management applications. To deal with that, in a large NSF project of Future Renewable Electric Energy Delivery and Management (FREEDM), we implement a SG prototype, called the FREEDM Hardware-in-the-loop (HIL) testbed. The so-called Distributed Grid Intelligences (DGIs) act as the distributed intelligent agents in the SG prototype which can group specific peers to exchange power load among power demands and supplies. There are also many other existing works contributing a variety of platforms to integrate power and communication systems. But, in our FREEDM project, we build a SG testbed, which is a new platform that combines an HIL power system and a real-time communication system. The power system devices are managed by the DGIs that are connected into the communication networks. The DGIs act as intelligent energy management agencies for the power system, while information nodes for the communication networks. The DGI instances are coded on embedded computer boards with processing and communication capabilities. A DGI represents its power device to communicate with other DGI instances or DGI nodes. DGIs being connected in LAN and WAN may be grouped together to meet the power demand and supply requirement. A DGI group may cover a LAN, or a LAN and WAN simultaneously, depending on the location of DGI nodes. When electrical faults isolate a section from the power system, in communicational sense, the section is still connected to and can exchange the information of grid states with other sections in the power system. The real-time and HIL features of the testbed are reflected in the design of both power and communication systems. To implement the concept of HIL in the power system, some power devices are implemented by real-world electrical hardware, while other devices are simulated in the Real Time Digital Simulator (RTDS) platform. To implement the concept of HIL in the communication system for the DGIs, the DGI LANs are implemented by Ethernet switches, while the DGI WAN is simulated in real-time by OPNET, a network simulator program. Within OPNET, there is a system-in-the-loop (SITL) interface that interprets DGI traffic between real packet formats and simulated formats.
Identifier: FSU_migr_etd-9562 (IID)
Submitted Note: A Dissertation submitted to the Department of Elctrical and Computer Engineering in partial fulfillment of the requirements for the degree of Doctor of Philosophy.
Degree Awarded: Summer Semester 2015.
Date of Defense: July 8, 2015.
Keywords: ad-hoc communication network, distributed grouping intelligences, network security, queueing network, real-time HIL testbed, smart grid
Bibliography Note: Includes bibliographical references.
Advisory Committee: Ming Yu, Professor Co-Directing Dissertation; Hui Li, Professor Co-Directing Dissertation; Anke Meyer-Baese, University Representative; Michael Steurer, Committee Member; Petru Andrei, Committee Member.
Subject(s): Communication
Computer engineering
Electrical engineering
Persistent Link to This Record:
Owner Institution: FSU

Choose the citation style.
Cai, Z. (2015). TAG: A Timing Adaptive Grouping Protocol for Smart Grid Communications. Retrieved from