PaperlayerView exact revision on arXiv

Source-linked AI summary

For the Grid and Through the Grid: The Role of Power Line Communications in the Smart Grid

Stefano Galli, Anna Scaglione, Zhifang Wang

arXiv:1010.1973v2cs.NIcs.ITeess.SY

TL;DR

The paper asks whether PLC is a good candidate for Smart Grid applications and examines this question through PLC capabilities, Smart Grid communications and control requirements, applications, modeling, and grid topology. It concludes that PLC is mature and useful in Smart Grid settings, while its deployment must account for application fit, coexistence, standardization, and network-structure constraints.

  • Problem

    The paper addresses the open question of PLC’s role in Smart Grid communications amid difficult large-scale control requirements and competing technologies.

  • Method

    The paper surveys PLC history and standards, analyzes Smart Grid sensing, communications, control, and applications, and reviews channel modeling and grid topology.

  • Results

    PLC is mature for deployment, but its suitability depends on matching the PLC technology to the application and managing coexistence, interference, and topology-related vulnerabilities.

  • Takeaways & Limitations

    PLC can support Smart Grid applications using existing grid infrastructure, while hybrid PLC/wireless networking may improve distribution-grid data-network robustness and reliability.

Abstract

from arXiv · show

Is Power Line Communications (PLC) a good candidate for Smart Grid applications? The objective of this paper is to address this important question. To do so we provide an overview of what PLC can deliver today by surveying its history and describing the most recent technological advances in the area. We then address Smart Grid applications as instances of sensor networking and network control problems and discuss the main conclusion one can draw from the literature on these subjects. The application scenario of PLC within the Smart Grid is then analyzed in detail. Since a necessary ingredient of network planning is modeling, we also discuss two aspects of engineering modeling that relate to our question. The first aspect is modeling the PLC channel through fading models. The second aspect we review is the Smart Grid control and traffic modeling problem which allows us to achieve a better understanding of the communications requirements. Finally, this paper reports recent studies on the electrical and topological properties of a sample power distribution network. Power grid topological studies are very important for PLC networking as the power grid is not only the information source \textit{but also} the information delivery system - a unique feature when PLC is used for the Smart Grid.

I. INTRODUCTION

The paper asks what PLC can offer Smart Grid applications and examines its role amid evolving grid-control, communications, and standardization challenges. It presents PLC as a mature, widely deployed technology whose suitability depends on matching technology and application while managing coexistence and interference.

  • PLC background: PLC evolved from early narrowband utility communications to broadband systems and high-data-rate narrowband solutions.Broadband PLC systems in the 2–30 MHz band can reach up to 200 Mbps, while newer high-data-rate narrowband systems operate between 3–500 kHz.
  • Smart Grid motivation: The Smart Grid requires communications, sensing, and control technologies to modernize grid operation and integrate changing generation and demand.The paper identifies wide-area monitoring, two-way communications, enhanced control, and renewable-energy integration as central design concerns.
  • Smart Grid design challenge: Large-scale Smart Grid control remains difficult because simulations and small trials do not capture global stability threats, while existing network-control results rely on restrictive assumptions.The paper also notes that centrally polling and controlling the grid is not scalable and may cause congestion.
  • PLC and the Smart Grid: PLC’s Smart Grid role remains debated, with concerns about standardization, data rates, modem cost, electromagnetic compatibility, and competition from wireless.The paper argues that recent PLC advances address much of this concern, but it does not treat one communications technology as universally sufficient.
  • PLC and the Smart Grid: PLC is mature for deployment and has extensive transmission- and distribution-grid use, including broad AMR and AMI deployment.Its deployment cost is described as comparable to wireless because the power lines already exist.
  • PLC and the Smart Grid: PLC adoption can be hindered by choosing over-designed broadband technologies for unsuitable applications and by multiple non-interoperable standards that create interference risks.The paper emphasizes selecting the right PLC class for each application and using coexistence mechanisms to limit interference among neighboring devices.

C. Organization of Work

The paper surveys PLC history and standards, then analyzes Smart Grid communications, sensing, control, applications, modeling, and grid topology before presenting final recommendations. Its organization follows this progression from background to deployment design and conclusions.

  • Paper organization: The paper begins with PLC history and a review of recent PLC standards.It covers both narrowband and broadband developments before turning to Smart Grid requirements.
  • Paper organization: It next examines Smart Grid communication, sensing, control, and traffic requirements through the evolution of SCADA and sensor networking.The paper treats the Smart Grid as a communications and control problem rather than only a technology survey.
  • Paper organization: The paper analyzes PLC applications in both transmission and distribution networks.The historical context includes utility voice and data communications over high-voltage lines and low-rate telemetering and tele-control.
  • Paper organization: It then addresses channel modeling, grid topology, and the grid’s dual role as Smart Grid data source and PLC delivery infrastructure.The paper concludes with final considerations and recommendations.

B. Ultra Narrowband and Narrowband PLC

The section traces PLC from early utility narrowband deployments to broadband and newer standards, distinguishing low-, high-, and broadband-rate technologies and their application contexts.

  • Early PLC and NB systems: Early utility PLC deployments used ultra-narrowband and narrowband technologies for AMR, AMI, distribution automation, and demand response.Turtle and TWACS systems used voltage and current waveform disturbances for outbound and inbound communication.
  • Broadband PLC: Broadband PLC first targeted Internet access, then shifted toward in-home networking and audio/video applications after disappointing access results.Industry alliances subsequently developed specifications for in-home PLC products.
  • Narrowband standards: LonWorks, FSK, and Spread-FSK illustrate established low-data-rate narrowband standards, while PRIME, G.hnem, and IEEE 1901.2 represent newer high-data-rate efforts.PRIME uses OFDM in the CENELEC-A band and supports PHY data rates up to 125 kbps; G.hnem and IEEE 1901.2 target scalable energy-management communications.

B. The TIA-1113 Standard

The section presents broadband PLC standards, their coexistence mechanisms, and the interference and quality-of-service challenges created by multiple technologies sharing power-line media.

  • TIA-1113 standard: TIA-1113 defines a 14 Mbps OFDM broadband PLC PHY based largely on HomePlug 1.0.It uses BPSK or QPSK according to channel conditions and a CSMA/CA MAC with adaptive contention windows and four priority levels.
  • Broadband standardization: IEEE 1901 targets high-speed communication below 100 MHz for HAN and access applications and includes FFT-OFDM and Wavelet-OFDM PHY/MAC technologies.Its multi-PHY/MAC structure reflects an industry compromise rather than a technical necessity.
  • Broadband performance: IEEE 1901 devices must support at least 100 Mbps, with mandatory features enabling approximately 200 Mbps PHY rates.Higher frequencies can increase rates, but attenuation and short range limit the benefit.
  • Shared medium: PLC channels are shared across homes and premises, so indoor and outdoor interference can reduce data rates through collisions or cause service interruption.This shared-medium behavior makes coexistence central to broadband PLC deployment.
  • Coexistence limitations: The HomePlug hybrid delimiter enables CSMA/CA and virtual carrier sensing across HomePlug generations but cannot eliminate interference or guarantee QoS under growing traffic.It can also defer legacy stations indefinitely, creating efficiency, security, and Smart Grid availability concerns.
  • Coexistence mechanisms: The Inter-PHY Protocol evolved into the mandatory Inter-System Protocol, enabling coexistence among IEEE 1901 PHYs and between IEEE 1901 and G.hn devices.The mechanism was designed for resource sharing and supports distributed features such as time-slot reuse.

IV. THE ROLE OF COMMUNICATIONS IN THE SMART GRID

The paper presents SCADA as the conventional cyber infrastructure for monitoring and controlling the power grid, while highlighting latency, scalability, and data-delivery challenges as sensing becomes faster and more distributed.

  • SCADA commonly combines a human-machine interface, supervisory master server, RTUs or PLCs, IEDs, and communications infrastructure.
  • SCADA networks traditionally use mixed communications technologies and commonly follow a two-level tree topology.
  • Existing links between RTUs and control centers can be inadequate for the increasing data volume generated by faster substation automation.
  • SCADA centrally updates state estimation and system identification, but its architecture has historically underemphasized latency.
  • Asynchronous sensor delivery prevents SCADA from optimally controlling physical responses to contingencies in real time.
  • PMUs generate measurements at typically 30 observations per second, compared with one every 4 seconds for conventional technology, creating new networking requirements.

1) Flexible AC Transmission System (FACTS):

The paper connects Smart Grid control to coupled electrical-network dynamics and communication constraints. It emphasizes synchronous monitoring, distributed control challenges, and the need to model both grid state and network traffic.

  • 1) Flexible AC Transmission System (FACTS):: FACTS uses power electronics and related equipment to control AC transmission parameters, enhancing controllability and power-transfer capability.
  • 1) Flexible AC Transmission System (FACTS):: Distributed FACTS devices are smaller and less expensive than traditional FACTS devices, motivating interest in wide-scale deployment and decentralized control.
  • 2) Smart Grid in the Distribution Network:: The distribution grid is becoming more dynamic through AMI, renewable sources, DER, and distribution automation, increasing uncertainty for state estimation and control.
  • B. Control and Sensing for Cyber-Physical Systems: The grid couples an energy-flow network with a sensor, data-processing, and actuation network, so failures can trigger correlated sensor traffic and congestion.
  • B. Control and Sensing for Cyber-Physical Systems: SCADA leaves substantial control to human operators, while scalable automated control remains difficult under communication constraints.
  • 1) Grid Control Aspects:: At nominal frequency f0, the power network is modeled by YV = I, with admittance determined by topology and electrical parameters.
  • 1) Grid Control Aspects:: Power-flow control acts on injected complex power S or network admittance Y, but centralized polling and control-signal distribution are not scalable and may cause congestion.

2) Traffic Generated by the Physical System:

The paper argues that Smart Grid communications must be designed around the physical grid’s coupled information source and its embedded topology. It therefore emphasizes traffic modeling and joint electrical-topological analysis for PLC planning.

  • Increasing service rates can address congestion by preventing intermediate-node buffers from growing, but overprovisioning does not remove the polling bottleneck.
  • Exploiting correlation and structure in sensor data can reduce information flows and help manage routing and processing complexity.
  • Modeling traffic from phase sensors in the electrical network is identified as an important research direction.
  • For PLC Smart Grid applications, the communication graph is a subgraph of the physical power-delivery infrastructure.
  • Combining topological and electrical characteristics enables analysis of PMU placement for continuous monitoring of channel, network, and distribution-system states.

3) Cooperative Schemes for PLC Networks:

PLC cooperative schemes address scalability and broadcast challenges in Smart Grid networks by coordinating relay transmissions and improving message delivery predictability. The section also situates these approaches within PLC applications across grid voltage levels and reports established and emerging HV capabilities.

  • 3) Cooperative Schemes for PLC Networks:: Relay interference can make naive polling and conventional routing fail to scale as Smart Grid network size increases.Massively deployed relays may interfere with one another, causing throughput and routing problems.
  • 3) Cooperative Schemes for PLC Networks:: Physical-layer cooperation can improve broadcast and multicast forwarding by allowing relay signals to be superimposed in time and frequency.This approach addresses broadcast storms that can arise when control messages traverse many relays.
  • 3) Cooperative Schemes for PLC Networks:: HDR NB-PLC can transform channel contention into cooperation through Single Frequency Network flooding, making delivery more predictable and transmission more power efficient.The reported scheme uses flooding-based routing rather than treating contention solely as a conflict.
  • A. PLC for High Voltage Networks: HV PLC has progressed from early analog telephone links to digital systems supporting 320 kbps over 32 kHz bands across 100 km.The reported 10 bits/s/Hz spectral efficiency exceeds the cited BB-PLC and NB-PLC values.
  • A. PLC for High Voltage Networks: HV PLC supports remote fault detection and real-time sag monitoring, while BB-PLC trials achieved 10 Mbps with about 5 ms latency over an 8 km, 69 kV line without repeaters.The BB-PLC result complied with FCC emission limits, but further testing is needed before commercial deployment over HV lines.

B. PLC for Medium Voltage Networks

PLC is presented as a candidate for medium-voltage Smart Grid monitoring, automation, islanding prevention, and metering because existing power infrastructure can carry communications. Applications and performance depend on equipment age, transformer traversal, network architecture, and emergency traffic conditions.

  • B. PLC for Medium Voltage Networks: Existing MV infrastructure could provide communications for transferring equipment-status and power-flow data between substations without installing new links.This is especially relevant because many MV substations traditionally lack communication capabilities.
  • B. PLC for Medium Voltage Networks: Online MV diagnostics are operationally preferable to expensive truck rolls and temporally disconnected partial-discharge measurements.The cited monitoring need includes extending the lifespan of critical cable connections.
  • B. PLC for Medium Voltage Networks: LDR NB-PLC has been analyzed for detecting unintentional islands and was reported as less expensive than telephone-cable methods.A related photovoltaic study reported superior islanding prevention compared with other existing methods.
  • B. PLC for Medium Voltage Networks: PLC Smart Grid applications on the LV side include AMR/AMI, vehicle-to-grid communications, demand-side management, and in-home energy management.AMI adds two-way exchange with customer devices and real-time electricity-price awareness.
  • B. PLC for Medium Voltage Networks: PLC-based AMR/AMI has extensive deployment experience, with a few hundred million UNB/NB-PLC devices deployed worldwide.UNB-PLC propagates through several MV and LV transformers but offers very low data rates; HDR NB-PLC offers higher rates with reduced range and sometimes requires transformer conditioning.
  • B. PLC for Medium Voltage Networks: Cooperative PLC schemes have a proven track record of avoiding congestion in emergency AMI scenarios involving simultaneous meter access or urgent demand-response signals.The cited comparison contrasts PLC-based AMI with wireless approaches such as ZigBee and WiFi.

2) Vehicle-to-Grid Communications:

The paper examines PLC across vehicle-to-grid, demand-side, in-home, and broader Smart Grid scenarios. It emphasizes application-dependent technology selection, with NB-PLC offering advantages for appliances, meters, and PHEVs while BB-PLC remains relevant to home networking and Internet access.

  • Vehicle-to-Grid Communications: PLC can connect PHEVs, EVSEs, meters, appliances, homes, and utilities for applications such as localized peak-load control.The communication link is described as the key enabler for these interconnected applications.
  • Vehicle-to-Grid Communications: A PLC link physically associates a vehicle with a specific EVSE, supporting security and authentication in vehicle-to-grid communications.The channel is impaired by inverter-generated harmonics, and both BB-PLC and NB-PLC solutions are being tested.
  • Demand Side Management: 2 GW of load can be shed within minutes by Florida Power and Light’s TWACS system, which manages over 800,000 load-control transponders and 1.4 million AMI endpoints.The system has operated for more than twenty years using UNB-PLC technology.
  • In-Home Environment: HEMS can provide utilities with home-state and forecasted-demand information beyond the instantaneous demand reported by smart meters.The paper reports utility benefits including improved grid reliability and reduced peak demand, despite limited compelling financial benefits for residential customers.
  • In-Home Environment: In-home DC lines benefit both NB-PLC and BB-PLC because the channel is time-invariant and appliance cyclostationary noise disappears, except for inverter-generated impulsive noise.
  • Technology Selection: NB-PLC, BB-PLC, and UNB-PLC each have potential application areas, so technology choice depends on technical, regulatory, and business-case factors.NB-PLC is highlighted for appliances, meters, and PHEVs, while separate BB-PLC networks can support home networking and Internet access through gateways or HEMS.
  • Distinctive PLC Advantages: The PLC transceiver can combine sensing and communications, potentially supporting power-quality monitoring and detection of unhealthy grid devices.Existing power lines can also reduce the cost of deploying redundant communication channels, although repeaters and couplers add deployment costs.
  • Technology Selection: NB-PLC offers soft-modem upgrades, worldwide harmonization, and coexistence with BB-PLC, whereas BB-PLC may be over-designed for many Smart Grid applications.HDR NB-PLC standards still require field validation, and the paper notes unresolved questions about the throughput required by future applications.

VI. DEPLOYMENT ASPECTS: CHANNEL MODELING AND NETWORK TOPOLOGY

PLC deployment requires channel models and network-topology models that capture both communication behavior and the power grid’s role as the information source and delivery infrastructure. The channel varies across grid sections and is affected by attenuation, dispersion, and multiple noise processes.

  • Channel Characteristics: The PLC channel is frequency-selective, time-varying, and impaired by colored background noise and impulsive noise.Channel behavior also varies with grid structure and indoor wiring practices across and within countries.
  • Channel Characteristics: Attenuation and dispersion increase toward distribution networks and homes, especially at higher frequencies, while noise remains high across voltage levels.Transmission-side attenuation and dispersion are comparatively small and can be handled more readily.
  • Noise Sources: HV/MV networks experience leakage, discharge, converter, and transformer noise, while LV and HAN environments add appliance-driven time variation and cyclostationary noise.Switching transients, lightning, and other discharge events also produce impulsive noise.
  • Deployment Modeling: Network planning needs accurate, flexible channel models and topological models that characterize both the Smart Grid data source and its physical delivery network.The dual role of the power network is central to PLC deployment planning.

A. Recent Advances in Channel Modeling

Recent PLC channel-modeling work advances from empirical and multipath descriptions toward deterministic transmission-line and multi-conductor models. Key results include statistical attenuation laws, time-varying channel characterizations, and unified treatment of grounded links.

  • Modeling Challenge: A commonly agreed PLC channel model remains absent, which has likely slowed transceiver optimization and the pursuit of general results.
  • Recent Advances: Recent advances include the multipath law, noise classification and modeling, channel isotropy, linear periodically time-varying behavior, and grounded–ungrounded link relations.
  • Statistical Models: PLC channel attenuation and RMS delay spread follow log-normal distributions, providing a statistical basis for channel modeling and throughput calculation.The result is reported for LV/MV PLC channels and is linked to multipath distortion from reflections and impedance discontinuities.
  • Recent Advances: Block models used in wireless and DSL channels have recently been shown applicable to PLC, supporting block transmission and precoding strategies.Most reported results concern broadband PLC, motivated by IEEE 1901 and ITU-T G.hn projects.
  • Multipath Models: The multipath model explains PLC propagation through echoes caused by branches, impedance mismatches, and multiple reflections along power-line cables.
  • Deterministic Models: Transmission-line theory can derive multipath parameters when link topology is known, but computational complexity grows with the number of discontinuities.This limitation has motivated frequency-domain deterministic models based on TL theory.
  • Deterministic Models: Multi-conductor transmission-line models extend two-conductor approaches to additional wires, including grounding, in low-voltage indoor models.The equivalent grounded link can be represented using cascaded two-port networks.

2) Statistical Models:

PLC channel modeling combines deterministic transmission-line calculations with stochastic topology models because real wiring varies widely. The paper also links channel behavior to power-grid topology, whose structure affects PLC network traffic and connectivity.

  • Statistical Models: PLC transfer functions can be calculated deterministically when link topology is known, but variable topologies and wiring practices require stochastic modeling.The paper combines multiconductor transmission-line models with representative field topologies to study coverage and expected transmission rates.
  • Statistical Models: In-home topology generation can use electrical-code constraints on outlets, wire gauges, and inter-outlet spacing to produce realistic channel realizations.Generalizing this approach requires knowledge of the electrical codes used in each country.
  • Statistical Models: LV/MV PLC attenuation is log-normally distributed, supporting channel modeling and achievable-throughput calculations.The model attributes distortion to cable low-pass behavior and multipath echoes from impedance mismatches and discontinuities.
  • Statistical Models: Empirical studies confirm log-normal attenuation in US indoor and outdoor MV PLC channels, and similar behavior has been observed in coaxial and telephone channels.These observations support generalized statistical wireline channel modeling.
  • Power-Grid Topology: Power-grid topology analysis helps explain PLC information traffic and complements fading models with structural effects on network behavior.The paper extends prior transmission-grid analysis to a real-world 396-node rural US MV distribution network.
  • Power-Grid Topology: The sample distribution network is sparse and radially structured, with long communication paths that can grow with network size.Radial topology is common in long rural lines but has the largest maximum communication delay because hop counts increase with network size.

2) Graph Theoretic Analysis of a Sample MV Distribution Network:

The paper analyzes a 396-node rural MV distribution network using graph-theoretic metrics and compares its structure with transmission networks. The sample is sparse, weakly connected, highly branched-limited, and characterized by mostly short branches with an exponential tail.

  • Network Assumptions: The analysis assumes couplers at transformers and switches so PLC connectivity is unaffected by transformer types or switch status.Without couplers, transformers and open switches segment the network into separate sections.
  • Graph Metrics: The study evaluates network size, average degree, average shortest-path length, degree correlation, algebraic connectivity, clustering, and spectral density.These metrics characterize structural size, connectivity, path lengths, and degree relationships.
  • Topological Results: The 396-node MV network has average degree ⟨k⟩= 2.12 and average path length ⟨l⟩= 21.10 hops, with nodes about 16.50 hops from the likely traffic sinks.Its average degree is comparable to the two transmission networks, while its average path length is longer even than the much larger WSCC system.
  • Topological Results: Algebraic connectivity is λ2(L) = 0.00030 versus 0.0094 for IEEE-300 and 0.00076 for WSCC, indicating weak overall connectivity and vulnerability to islanding.The network’s clustering coefficient is zero, compared with 0.0856 for IEEE-300 and 0.0801 for WSCC.
  • Degree Distribution: The maximum node degree is 4: 16% of nodes have one branch, 60% have two, 22% have three, and 2% have four.These values show a strongly degree-limited distribution compared with transmission networks, where maximum degrees of 20 or 30 occur.
  • Branch Lengths: Most branches are shorter than 1,067 m (3,500 ft), while the branch-length distribution has an exponential tail with few extremely long branches.The distribution is shown using probability versus length and log-probability versus length.

C. The LV Distribution Network

The paper presents PLC as a mature, flexible Smart Grid communications option whose coexistence requirements and infrastructure-based benefits shape deployment decisions. It emphasizes PLC’s ability to support diverse applications while leveraging utility-controlled power lines.

  • C. The LV Distribution Network: PLC technologies operating in the same frequency band require coexistence mechanisms to reduce interference and performance degradation.A well-designed analog front end can separate technologies occupying different frequency ranges, but it cannot resolve same-band interference alone.
  • C. The LV Distribution Network: Coexistence provides a transitional path for selecting technologies from field deployment data while allowing Smart Grid and home-networking devices to mature independently.The paper characterizes coexistence as necessary because PLC spectrum usage is not regulated and available bandwidth limits efficient frequency-division separation.
  • C. The LV Distribution Network: PLC can reduce the cost of redundant communications channels by reusing existing wired power infrastructure for utility protection and control.The paper identifies redundant communications as an almost universal utility requirement and treats the existing power network as a deployment asset.
  • C. The LV Distribution Network: PLC transceivers can combine sensing and communications by switching between sensor and modem functions.This blurs the traditionally separate roles of sensing and communicating in utility applications.
  • C. The LV Distribution Network: Power lines can provide direct, low-latency, utility-controlled paths for applications such as tele-protection and communications under deregulated telecom conditions.The paper links PLC’s value to bounded latency, direct utility control, and the availability of varied technologies for different Smart Grid applications.

B. Architecture Must Come First!

The paper argues that Smart Grid architecture and requirements should be clarified before deployment decisions become entrenched. It cautions against assuming that existing technologies or a single common technology will automatically fit Smart Grid applications.

  • B. Architecture Must Come First!: Smart Grid deployments may remain in the field for decades, making early attention to design and architectural frameworks important.The paper calls for frameworks that connect existing standards with the longer-term Smart Grid vision and identify standards for migration.
  • B. Architecture Must Come First!: Migration is sensible, but technologies should be selected judiciously rather than simply carrying old standards into future Smart Grid systems.The paper notes that migration toward new standards may take decades until equipment becomes aligned.
  • B. Architecture Must Come First!: The paper identifies two questionable deployment assumptions: that unrelated off-the-shelf technologies can be used seamlessly and that one technology should serve applications such as DSM or AMI.Both assumptions are criticized as preceding a full understanding of application requirements or normal market alignment.
  • B. Architecture Must Come First!: Statistical-physics analyses and blackout studies treat the power grid as a complex system whose phase transitions and triggering mechanisms matter for distributed-control stability.Published work uses tools such as percolation theory, while analyses of US blackouts provide supporting evidence for complex dynamics.
  • B. Architecture Must Come First!: Blackout-size data show a power-law frequency pattern, but the overall effects of load-side changes such as DSM, DR, and DER remain difficult to determine.The paper states that these interventions could alter blackout-size distributions, while emphasizing that little is currently known about the effect.
Loading 1010.1973v2…