A Road Map

to Integration
Smart grid-related blogs, newsletters,
and conferences have endured numerous debates and
discussions around the issue of whether or not the smart
grid is being integrated correctly. While most debates
focus on approach, methodology, and the sequence of
what needs to be done, there is insufficient discussion
about what is actually meant by smart grid integration. This article attempts to present a holistic view of
smart grid integration and argues for the importance
of developing system integration maps based on a
utilitys strategic smart grid road map.
Faced with diverse technological, organizational, and
business issues that adversely affect the bottom line, utility companies are contemplating immediate changes and/
or upgrades of their technologies, business processes, and
organization. At the same time, however, the realities of
insufficient resources, regulatory impediments, and technological hurdles have prevented the development of concrete plans and concerted actions in this regard.
A closer look at mainstream discussions within the
utility industry reveals that, despite consensus about
the need for change, there is no agreement across the
board in any given utility about a smart grid road map
and integration map. The absence of industrywide
standards and blueprints for smart grid integration has
further compounded the issue. The silo mentality of
the constituent parts of the utility organization drives
the generation folks to push for expanding generation
capacity through the integration of renewables, the
transmission people to urge expansion of transmission capacity through automation, and the distribution community to argue for integration of new assets,
technologies, and intelligence on the downstream side
of the network.

Digital Object Identifier 10.1109/MPE.2014.2301515

Date of publication: 17 April 2014

52

ieee power & energy magazine

Image licensed by Ingram Publishing

By Hassan Farhangi

1540-7977/14/$31.002014IEEE

may/june 2014

Perspectives
on Smart
Grid Development

Furthermoreand given the fact that each group has traditionally been exposed to certain vendors and technology providers for its respective siloeach constituency tends to
regard the technologies and solutions offered by those vendors as the answer to much larger,
systemwide problems. And in the utility environment, these problems by default transcend
the confines of a single silo.
The situation is further complicated by the diversity of views, interests, and approaches
advocated by vendors and technology providers in the field. Influenced (and constrained) by
its core competencies and technologies, each vendor defines the problems, and therefore the
solutions, in the way that best suits its own technologies and products. One should therefore
not be surprised to hear different suppliers put different spins on basic concepts such as distribution automation, demand response, and so on. The irony is that they are mostly sincere
in what they are advocating. The issue is whether any of their prescriptions is the Holy Grail
needed to solve the utilitys smart grid integration puzzle. This seems to be a reenactment
of Rumis story of the blind men and the elephant. Each person has his own understanding
of what the creature is based on what part he has managed to touch. The absence of sight
(or light) has convinced each and every one of the righteousness of his version of the truth,
ignoring the fact that the smart grids systemwide issues require all its constituent parts
to work together and implement a collective strategy for doing what needs to be done. In
Rumis words, If each of us held a candle there, and if we went in together, we could see it.

Smart Grid Development

Utilities in North America have had their fair share of challenges in taking the first step on
their path to full implementation of the smart grid, namely, large rollouts of smart metering
across their distribution circuits. The reaction of the public to the push by utility companies
to implement smart metering took many in the industry by surprise. In addition to open
calls by consumer associations to do away with the idea, many jurisdictions saw the introduction of symbolic resolutions, passed by county and municipal councils, banning smart
meter installation. In response to this backlash from customers, many North American
utilities have had to either slow down smart metering rollouts or devise opt-out programs
while investing in information campaigns to reach and influence their customers.
Despite the specific form that the consumer backlash took (e.g., concern about the health
effects of RF radiation, the privacy and security of customer data, or an imminent rise
in the cost of energy), one could see that such concerns were primarily attributed to an
absence of buy-in for this new technology on the part of utility customers.
What is interesting is that very few, if any, utilities have attempted to answer the more
fundamental question of why their customers should embrace this new technology with
open arms. What will make customers want to be willing participants in this process?
Would smart meters reduce utility bills? Would it provide customers with more reliable
may/june 2014

ieee power & energy magazine 53

This article attempts to present a holistic view of smart grid

integration and argues for the importance of developing system
integration maps.

service? Would smart metering protect customers vital

information and personal data? What would the short- and
long-term impacts of smart metering be on customer engagement? What is next after smart metering? What future functionalities and capabilities would be enabled through smart
meters that would be beneficial to customers?
In fact, not only have very few attempts been made to
answer these questions adequately and convincingly, but
some utilities have added fuel to the fire by suggesting that
smart meters will help with customer behavior change or
load control through time-of-use (ToU) mechanisms and
dynamic pricing. Without a well-formulated plan to prove to
customers that such behavior change will not and should not
happen at the cost of their convenience or at their expense,
such suggestions have only reinforced public perceptions
that smart metering is nothing but a quick money-grabbing
exercise on the part of cash-strapped utilities trying to fill the
holes in their budgets on the backs of rate payers.
The questions that arise here are these: Why wouldnt
utilities confront such misconceptions head on and communicate to their customers the benefits of smart metering? Why
wouldnt they portray smart metering as the first step toward
smart grid integration and all the unprecedented capabilities
that a smart grid will offer their customers? Why wouldnt
they attempt to convince their customers that a smart grid
will effectively empower them to be active stakeholders and
players in energy and service transactions?
Although there could be many reasons for such a disconnect between utilities and their customers, some have speculated that either utilities have not yet managed to develop a
strategic road map for the smart grid or if they have, that there
was very little consensus across their organizations on the
integration plan and on a realistic schedule for implementing it. Regardless of the root cause, pundits have seen this
as a failure on the part of the utilities to formulate the right
communication plans to help their systems, organizations,
staff, infrastructure, assets, and, ultimately, their customers
navigate collectively through this uncertain yet exciting transition to a new set of service transactions, energy paradigms,
and fundamentally different roles and responsibilities.
It goes without saying that no utility has ever discounted
the need for a strategic smart grid road mapand subsequently a smart grid integration mapprior to making such
large investments in their assets and infrastructure. The
question is therefore not the existence of such blueprints
but simply their role in driving (and informing) the major
54

ieee power & energy magazine

technology investment commitments utilities are making

today. The litmus test for this process is to ask a series of
questions so as to ascertain how conducive each investment
is to a seamless transition from a less intelligent grid to an
intelligently integrated smart grid.

Strategic Smart Grid Road Maps

As discussed earlier, the need for the development of strategic road maps for smart grids was recognized early on by
many practitioners and planners in the utility industry. Such
work began by identifying utilities business and corporate
objectives and goals, recognizing the most critical issues and
impediments to reaching those goals, and devising plans for
how to address them. Figure 1 depicts an early attempt by a
group of experts from the British Columbia Institute of Technology (BCIT) and BC Hydro who worked collaboratively
over a period of several months to formulate an R&D as well
as demonstration road map for their joint smart microgrid
initiative at BCITs Burnaby campus. This collaborative
effort took into consideration what each party was hoping to
achieve from the joint project, the modalities of their respective development efforts, the realities on the ground, and the
resources and technologies needed to achieve those goals
over a five-year period.
As Figure 1 demonstrates, the road map highlighted the
need for several constituent streams, each informing as well
as enabling other streams along the way. For example, the
energy management system (EMS) stream included several
layers of sophistication and features, based on the availability of certain assets and capabilities provided by other
parallel streams, such as advanced metering infrastructure
(AMI), communication infrastructure, and load and asset
management. The interplay among these functionalities,
made possible by the integration of their respective technologies, was conceived as enabling stepwise jumps in the range
of capabilities the initiative was expected to provide.
The smart microgrid initiative implemented by BCIT and
BC Hydro reinforced the notion that smart grid integration
consists of several concurrent streams, designed to introduce
intelligence (and thereby command and control) into strategic areas of the system, providing capabilities and functionalities that transcend the legacy silo architecture of the system.
The nature of these capabilities and their intended reach will
determine which assets and/or subsystems must be integrated
in order to realize the target functionality. As Figure 2 demonstrates, smart grid integration does not always have to be
may/june 2014

AMI Database
Architecture

Portal
Technology

EMS Ver 1

ZigBee Network
with Smart Energy
Profile
MODBUS
Integration
Techniques
ANSI-C12.19
Data Aggregation

E2E PLC and RF

SMI Network

MDMS

ANSI C12.22
Smart Metering

Load Control for

HWT and BBH

Mobile EMS
Ver 1
Mobile EMS
Dashboard

End Customer
Experience

Social Science
Factors in Energy
Management

Scientific Paper
On EMS

Load Shedding

Scheduling

EMS Ver 2

Pricing Signal
Broadcast

Time of Use
Tariff
Maximum
Demand Tariff

Asset
Management
Dashboard
Microgrid
Asset
Manager

Integrated
Microgrid Sensor
Network

Microgrid
Substations
Automation

Integration of
Thermal Turbine

Integration of
Storage

Integration of
Wind Turbine

Integration of
Solar Modules

Microgrid
Controller
Integration

SGCC Topology

EMS Ver 3

End-Customer
Experience

Social Science
Factors in Energy
Management

Scientific Paper
On DSM

Net Metering

Protection and
Switching

Synchronization

Energy
Transactions
Revenue Models

EMS Ver 4

DC Distribution
Network

DC Protection
and Switching

Scientific Paper
On DC Microgrid

EMS Ver 5

End-Customer
Experience

V2G and G2V

Load Prediction
and Profiling

EV Charge
Manager

BCIT Campus IPP Pilot

DC Microgrid Pilot

EV Charge Pilot

A6: Microgrid
Islanding

E1: Rural DC
Microgrid

D1: Intelligent
Transportation Network

ieee power & energy magazine 55

may/june 2014

WiMax (1.8 GHz

and 3.6 GHz)

Dedicated Fibre

Scientific Paper
On IEC 61850

2012

ZigBee Sensor,
Thermostat and
IHD
Building
Automation

2011

A5: Expansion of Microgrid

Co-Gen Capacity
A4: Smart Grid
Control Center
Demand-Side
Management

BCIT/BC
Hydro
Smart Microgrid
RD&D
Road Map
C5: Microgrid
Asset Management

A3: Advanced
EMS
B2: Dynamic
Tariffs
C4: Distribution
Automation

A2: Mobile
EMS
Residence Competition
C3: WAN Network

A1: Basic EMS

B1: LAN Network
C1: AMI Infrastructure
C2: HAN Network
2009

Version 0.7
Date: 5 Feb. 2011
Author: Dr. Hassan Farhangi
Status: Proprietary and Confidential

Paper
Tool
IP
Pilot
DC
Stream
EV
Stream
Revenue Automation
Stream
Stream
EMS
Stream
Legend:

2010

2013

figure 1. A typical strategic road map.

Real-Time Simulation and Contingency Analysis

Distributed Generation and Alternate Energy Sources
Self-Healing Wide-Area Protection and Islanding
Asset Management and Online Equipment Monitoring
Demand Response and Dynamic Pricing
Participation in Energy Markets

figure 2. A smart grid functional integration map.

end to end or for all capabilities. Different smart grid functions require the integration of different assets, with different
capabilities and different requirements. As such, a smart grid
integration map must adhere closely to the utilitys strategic
road map and to the intended smart grid functionalities that
need to be enabled in each stage of development.
In practice, smart grid integration has taken many
twists and turns. It is unlikely that North American utilities have followed similar modalities and approaches to
smart grid integration. It is more logical to assume that
each utility has taken a unique path toward implementing
its smart grid plans.

Early Integration Attempts

Given the fact that the electricity distribution network in almost
all the jurisdictions across North America had long been overdue for a major overhaul, the first step many utilities took in
smart grid integration consisted of limited rollouts of one-way
automated meter reading (AMR); this was followed by major
investments in two-way AMI. As Figure 3 suggests, the overwhelming justification for investments in AMI was its enabling
role in facilitating the move toward an eventual realization of
smart grid functions. That understanding convinced many utilities in North America to plan for major smart metering investments. Many projects were announced, and pilots sprang up
across the continent. In the absence of a well-formulated smart
grid integration plan specifying how smart metering would actually lead to a smart grid, pilot project evaluations focused on the
requirements for smart metering rather than its forward compliance with future smart grid functions. As such, most pilots were
perceived to be successful, resulting in substantial follow-on
investments in smart-metering projects.
56

ieee power & energy magazine

As the full cost of ownership of AMI systems became

clear, however, and given regulatory constraintsand in
the absence of clear revenue modelsutilities found it
increasingly difficult to justify AMI capital expenditures.
In addition to the less-than-convincing cost-benefit models
for AMI, the consumer backlash against smart metering
slowed down AMI rollouts in many jurisdictions across
North America. The absence of clear smart grid road maps
and utilities unconvincing arguments in favor of smart
metering prompted many experts in the field to express
doubts about the entire rationale for AMI; many questioned whether or not the smart grid was being integrated
backwards. Some, for instance, suggested that return on
investments (ROIs) would be more palatable if distribution substations were automated first. Others pointed to the
need to start at the top, upgrading the utilitys enterprise
applications in the back office and on the enterprise bus
before attempting to invest at the bottom of the chain.
The fact of the matter is that all those questions were very
valid. And the reason why such doubts were being expressed
had everything to do with the absence of a utility smart grid
integration map that would have demonstrated how AMI
was going to be leveraged to gradually upgrade the functional capabilities of the grid. In The Path of the Smart
Grid [IEEE Power & Energy Magazine, January 2010],
an early attempt to demonstrate linkages between technologies and capabilities along the path from smart metering to the smart grid was presented. In Figure 3, the author
argued that AMI would have to be perceived as an enabling
platform for two-way communication and distributed command and control among all previously unmonitored and
uncontrolled components of the distribution system. The
may/june 2014

Capabilities

Technologies

Intelligent Appliances
Intelligent
Agents

ROI

Intelligent
Applications

Emission Control

Smart
Grid

Load Management
Preventive/Self-Healing
Substation Automation

Network
Management
Distributed
Control
Smart
Sensors

Distribution Automation
Customer Information System
AMI

Asset Management
Outage Detection and Restoration

Two-Way
Communication
One-Way
Com.

Customer Portals
Distributed/Co Gen

Demand Response
AMR

Automated Billing
Investments

figure 3. Smart grid technologies and capabilities.

article further emphasized that the protocols, topologies, and

architecture of AMI systems had to be designed as forwardlooking sources of sensory, status, and alarm information to
allow future (and still undeveloped) smart grid capabilities
to reach and be integrated with the downstream side of the
utility system. In other words, the specification of the AMI
system had to take into account the communication, data
and command exchange, and access requirements of future
smart grid applications.
What the figure attempted to further demonstrate was the
notion that smart grid integration can be broadly divided into
two categories. One operates in local domains using global
system attributes, such as demand response or outage detection, that require access to real-time local data with local
analytics and local decision-making processes. The second
operates over multiple domains, requiring wide-area situational data awareness and an overview of the system constraints as a whole and system operational objectives, such
as management of distributed energy resources, self-healing,
outage prevention, and so on. While the first category is
enabled through a well-designed, forward-looking AMI system with appropriate latency, throughput, availability, and
resilience requirements, the second relies on a well-designed
and optimally integrated network of distributed systems
with suitable security, scalability, and access protocol specification that enables efficient distributed command and control through a multilayer, multitier, and multiagent system.
In other words, the figure emphasized the fact that smart
grid integration should be built on forward-looking information and command and control architectures capable of
may/june 2014

meeting the functional and operational requirements of a

gradually evolving smart grid system, with incremental
needs for higher levels of performance, scalability, and
resilience and without the need for costly departures from
its original design and implementation. It goes without saying that once investments are in place, utilities find it almost
impossible to undo commitments to AMI, substation automation, and so on to upgrade their assets so as to enable
new smart grid functionalities. And that means the cost and
the pain associated with the transition from the legacy grid
to the smart grid will to a large extent depend on the suitability of the utilitys smart grid integration map supporting
that transition.

Building the Smart Grid

The irony is that there is an element of truth in every
approach to building a smarter grid. This diversity of views
can only be attributed to the fact that without a doubt there
is more than one way to integrate a smarter grid. Depending on a variety of potentially conflicting and yet interacting
drivers (priorities, regulations, legacy assets, organizational
and process issues, and so on), different utilities may choose
different points of departure for their long journeys toward
the smart grid. And consequently, the trajectory each utility
takes in integrating its system with different smart grid functionalities, even if similar starting points are adopted, may
prove to be quite unique and dissimilar from others.
Regardless of where that starting point is, however, it is
crucial for utilities to spec out their journeys (as much as
they possibly can, given all the unknowns) in such a way
ieee power & energy magazine 57

Billing and
Accounting

Finance

DOC
Management

Corporate

HR

Biz Ops

Trading

Scheduling

Settlements

Forecasting

System Planning

System
Planning

Capacity
Planning

Network
Planning

Data
Warehousing

Engineering

Maintenance Parts/
Work
Asset
Scheduling Supply Management Management

Sys Ops

EMS

DMS

OMS

Customer Service

CIS

MDM

IVR

Backhaul Coms

Field Coms

Fiber
Network

WiMax
Network

Station Bus
(Revenue Data)
WiMax

Revenue Data

Residential
Metering

IPP

Color Code

Broadband
PLC

Process Bus
(Breakers, Switchgear,
Reclosers, Transformers)

Commercial
Metering

Narrowband/
Broadband
PLC

Industrial
Metering
Gateways

Biofuel

Wind

Data

Data
Warehousing

Microwave
Network

Net Meters

PV

GIS

Public
Proprietary RF
Network

Last Mile Coms

Two-Way Coms

DSM

ERP

COM

HR:

Human Resources Department

ERP:

Enterprise Resource Planning

DOC:

Document Management

EMS:

Energy Management System

DMS:

Distribution Management System

OMS:

Outage Management System

DSM:

Demand-Side Management

GIS:

Geographic Information System

CIS:

Customer Information System

MDM:

Metering Data Management

IVR:

Interactive Voice Response

PLC:

Power Line Communication

WiMax:

IEEE 802.16 Wireless-Networks Standard

RF:

Radio Frequency

PV:

Photovoltaic

Biz Ops: Business Operations

Sys Ops: System Operations
Coms:

Communication Systems

IPP:

Independent Power Producers

VVO:

Volt/Var Optimization

CVR:

Conservation Voltage Reduction

Small
Hydro

Asset

figure 4. A layered smart grid system integration map.

that subsequent moves toward other areas of the system can

be realized seamlessly and without the need to substantially
change or upgrade the assets that have already been rolled
out. As an example, if after AMI rollout the plan is to take
substation automation as the next area for investment, the
utility should ensure that subsequent downstream moves to
implement demand curtailment or load shedding as well as
upstream moves to implement asset management or selfhealing can be realized without having to change or upgrade
the investments made for substation automation and AMI.
Figure 4 shows the extent to which the specifications of
various parts of the system need to be taken into account
to ensure seamless integration of components, assets, and
functions in both the spatial and temporal domains. It
emphasizes the notion that the sphere of influence of smart
grid capabilities varies considerably across different functions. Some span upstream layers of the system, involving
58

ieee power & energy magazine

enterprise functions, utility operations, and revenue management (e.g., contingency management, asset management,
energy market participation, and so on), while others traverse local downstream layers of the system, involving field
and prosumer-facing assets (e.g., demand response, load
management, and so forth). It is the latter group that places
a heavy burden on the AMI system, as it requires close integration and tight operational linkages to AMI system components, protocols, and technologies. A poorly designed and
implemented AMI system would prove to be inhibitive for
the efficient implementation and/or correct operation of such
downstream smart grid functions.

The Smart Grid Integration Map

As discussed above, the point of departure on the path to the
smart grid and the particular set of smart grid capabilities
each utility may want to achieve will not be the same across
may/june 2014

all jurisdictions. Having said that, and regardless of a utilitys current baselines, operational priorities, and organizational abilities to devise smart grid system integration maps,
the ideal way to approach smart grid system integration is
to analyze each smart grid capability in terms of its core
decision-making and data-customer/command-supplier
interface requirements. That analysis will identify to which
domain such functions will belong; to which layer they will
have to reside or be attached; and what their data processing,
command and control, interface protocol, and communication requirements will be.
Figure 4 is an attempt to take the smart grid functions
of Figure 3 and map them across different layers of a fully
integrated smart grid system. In such an approach, smart
grid functions are seen as cutting across multiple layers of
utility structures, including but not limited to corporate,
engineering, field operations, and distribution systems. This
approach turns the utilitys traditional silo structures on its
head, as it traverses organizational boundaries for efficient
and cost-effective realization of target smart grid functions.
What is critical in this approach is not how a particular function needs to be realized, but where it belongs as an entity
providing other entities in its vicinity with the services for
which it is designed. Association with a given layer will then
determine the performance metrics of the assets needed to
support the efficient operation of that capability.
Moreover, the layered approach attempts to identify the
nature of each layer in terms of the dominance of data processing and communication technologies versus the utilitys
traditional assets. This does not necessarily mean that a layer
that is dominated by data processing does not depend on
communication technologies or other existing utility assets.
By its nature, each smart grid capability will have to rely on
all three constituent components of the smart grid: power
systems, telecommunication, and information technology.
Furthermore, the layered approach embeds within it the
notion of the temporal and spatial requirements of each layer.
More stringent requirements for access to real-time data will
place a layer closer to layers that produce such data and vice
versa. In other words, the proximity of layers to each other
is directly proportional to their interface and data and command exchange requirements. As an example, the EMS and
the volt/var optimization (VVO) and conservation voltage
reduction (CVR) layers have to be in close proximity to each
other and to the field assets with which they have a direct,
real-time, and unimpeded data exchange relationship. The
same is not true for the billing layer, which can be placed
further away from and without a need for real-time connections to field assets.
It goes without saying that not all functions within each
layer need to be integrated at the same time. Each utility
could pick and choose one or more functions from each
layer and decide when and how they need to be realized.
Regardless of the integration plan for each function, however, what is critical is to understand which layer it will
may/june 2014

belong toand as such, what its data processing, command and control, interface protocol, and communication
requirements will be. This understanding will ensure that
the architecture of the system, the communication topology,
the adopted technologies, and the associated protocols are
chosen in such a way that they will lend themselves to the
future integration of new functionalities and capabilities.
That is the only way to ensure that the gradual transition to
the smart grid is managed without excessive reengineering
and expensive overhauls.
As discussed, each utilitys enterprise function places a
particular set of requirements on different layers of the system in terms of its vital specifications, such as data structures, protocols, security regime, latency, throughput, and,
last but not least, interactions with the actual assets. In reality, of course, applications can and should reside where their
function is required: some will exist within a substation,
some in the utility back office, and others on the enterprise
bus. Neverthelessand regardless of the environment to
which they are attachedeach application must have the
ability to communicate seamlessly and efficiently with relevant system nodes as and when required. For instance, an
asset management application has to communicate with all
the relevant assets assigned to it from the different domains
of generation, transmission, and distribution.
As an example, a utility that intends to roll out its smart
meters first and subsequently integrate an asset management application over its vital system assets has to ensure
that the AMI system it is integrating will lend itself well
(as a set of distinct assets) to seamless integration with the
asset management application it will be rolling out in the
future. It goes without saying that it would not be acceptable to have patchworks of individual asset management
tools for different categories of assets. In other words, no
utility would be happy using an asset management tool for
its smart meters, another for its relays, switches, reclosers,
and protection components, and yet another for its transmission equipment. One would therefore expect that a major
requirement for the selection of any AMI solution would
be its ability to interface with existing or future smart grid
functionalities, enabling on-demand or event-based reporting of the health, configuration, settings, and maintenance
schedule of all AMI assets, including meters, head ends,
and communication equipment. Similarly, a utility planning
to implement dynamic pricing and ToU tariffs has to ensure
that its AMI system is capable of handling and/or relaying
such real-time information for the systems relevant points
of termination.

Realities on the Ground

The approach advocated in Figure 4 is unfortunately not
the norm. It is probable that most utilities will attempt to
integrate smart grid functions with their operations starting at two extreme ends of the system hierarchy: at the bottom of the chain through rollout of AMI systems and at
ieee power & energy magazine 59

Application Layer

Smart Grid
Smart Grid
Enterprise
Smart Grid
Enterprise
Applications
Enterprise
Applications
Applications

Enterprise Bus

ICT Layer

Information Technology Infrastructure

Security Regime

Power System Layer

Communication
Infrastructure
Utility Assets
Generation

Transmission

Distribution

figure 5. A hierarchical smart grid system integration map.

the top of the chain through adopting and integrating new

enterprise bus functions. That approach is understandable
and very much in line with the current constraints on utility assets and organizational structures. In fact, the early
attempts to modernize the system had to take into account
the realities of a highly compartmentalized system and
operational hierarchy tasked with delivering a critical service to customers while meeting the challenges with which
most utilities grapple.
Figure 5 depicts the approach utilities in general have
taken in their integration of smart grid functions. The point
of entry of new functions into the hierarchical structure of
the utility system has been at the interface with customers
(e.g., smart meters), together with the associated support
functions within the enterprise bus, such as meter data management (MDM) systems. Patches of plumbing to connect
the two ends of the function (e,g., the required communication system to support the capture and exchange of data
for the purpose of billing and revenue management) are thus
inserted within the appropriate information and communication technologies (ICT) layer of the system.
The question utilities have not answered here concerns what
other smart grid capability the chosen AMI technology can
support. The current integration of AMI systems across many

b#3

Public Switch Network

b#

D-Sub#1

T-S
u

D-Sub#2
D-Sub#4

M1

HAN#1

M2

HAN#2

M3

HAN#3

Utility Field Network

T-Sub#2

Power
Plant#1

T-S
u

T-Sub#1

b#

SG-A
pp#5

SG-App#3

SG-A
pp#4

Utility Core Network

Feeder

D-Su

SG-App#1 SG-App#2

Power
Plant#2

Wide Area Network (WAN)

Local Area
Network (LAN)

Home Area
Network (HAN)

figure 6. Disjointed network domains in a legacy grid.

60

ieee power & energy magazine

may/june 2014

jurisdictions in the world assumes a disjointed utility network

(assuming such networks exist at lower layers of the distribution
system, which is often not the case), parts of which can be conveniently bypassed. What typically happens in such installations
is that smart meters are grouped into a local-area network (LAN)
type of association (meshed or otherwise) and exchange their
data and commands through radio frequency (RF) or power line
carrier (PLC) links with a data aggregator unit (DAU) installed
on a pole in their vicinity. The DAU then employs a dedicated
communication link (often a proprietary wireless protocol) to
exchange the aggregated data and commands with an MDM
system within the utilitys enterprise bus.
As shown in Figure 6, the specification of the constituent
components of the system is therefore optimized to ensure
data exchange between smart meters and the associated
MDM system residing within the utilitys core network. The
AMI system as such is not only oblivious to anything that
happens in between, but it also has no provisions for handling or carrying any other information or data, however
critical or important such data may be for other smart grid
functions to be rolled out in the future.
This simply means that critical sensory data and information produced at the downstream side of the network, which
may be critically required by middle layers of the system,
bypass such layers and end up in the upper layers of the
system for a specific function and/or purpose (e.g., billing).
They therefore do not contribute and/or enable future smart
grid functions that may benefit from access to such data. To
elaborate this issue further, the next section of this article
examines typical examples of such applications that are
deprived of access to these critical data as a result of inaccurate smart grid integration maps.

Transformer

WAN
Interface

Field Area
Network
Interface
Feeder A

Recloser
Volt/Var and CVR
Optimization Engine
DMZ

Timing and
Syncronization

LAN
Interface
AMI Headend

DAU

Substation Bus

Substation
Area
Network

One would certainly hope that utilities quest for infrastructure modernization will not come to a screeching halt. Utilities will no doubt attempt to integrate additional smart grid
functionalities based on their particular priorities and road
maps. Two such functionsones many utilities intend to
implement nextare VVO and CVR. The U.S. Department
of Energys recent studies in energy conservation across the
United States indicated that CVR functions, if integrated
across less than half of U.S. feeders, could potentially yield
more than a 2% reduction in demand on the U.S. electrical system. In fact, it is public knowledge that many utilities
regard VVO and CVR as high priorities for implementation
on critically congested feeders. It has been claimed that the
ROI ration for VVO and CVR integration is six to one, with
a payback period of two to three years. That is certainly a
great incentive to regard investment in advanced VVO and
CVR as the next item on the integration map of many utilities after AMI.
Prior to AMI, the VVO and CVR functions in distribution substations (if they existed at all) used a statistically
aggregated profile of feeder load to determine the settings
and configurations of capacitor banks, tab changers, voltage regulators, and other devices used to correct the feeders power factor and to ensure compliance of the voltage
gradient across the entire feeder, from substation to the last
customer, with ANSI and IEEE requirements. Given the
fact that no real-time information about the actual voltage
samples across the feeder was available, the settings and
configurations for such VVO and CVR assets were either
ineffective or inefficient.

VR

Caps

VR

Caps

RF/PLC Interface

Power
SCADA

Data

Typical Issues with Smart

Grid Integration Maps

Substation EMS
Distribution Substation

EMS
HAN

EMS
HAN

EMS
HAN

EMS
HAN

figure 7. Client-server based VVO and CVR.

may/june 2014

ieee power & energy magazine 61

62

ieee power & energy magazine

MCU

Residential Load-H

Smart Meter-H

PLC
Modem

Capacitor Banks
Unit

Sub. Capacitor
Banks Unit

Cap. Bank
Switch

AFE

Line Voltage
Regulator

Cap. Bank Switch

Distribution Transformer
12.5 kV/0.4 kv

..
..
Voltage Regulator
Switch

MCU

Smart Meter-B
MV Breaker

Residential Load-B

PLC
Modem
Substation MV Bus

AFE

PLC
Modem

IA-IED

MCU

Substation Transformer
138-kv/12.5-kV
Transformer On-Load
Tap Changer

MV Breaker

Load Bank

MV
Distributed
Breaker Generation Source
12.5 kV

Capacitor Banks
Unit

MCU

AFE

Smart Meter-A

Residential Load-A

PLC
Modem

AFE
VVO/CVR
Opt. Engine

HV Breaker

Substation HV Bus

To HV Substation

figure 8. Substation-based VVO and CVR.

The advent of AMI changed that

situation. Engineers in charge of planning for new VVO and CVR functions
saw the opportunity to use real-time
voltage, current, and power factor
(V/I/PF) sample values from smart
meters at each customer node to build
a realistic and accurate real-time view
of the load profile across any given
feeder and thus to optimize VVO and
CVR settings based on an accurate
voltage gradient across the feeder, climatic conditions, and ToU. Such an
approach has been named adaptive
real-time VVO/CVR.
To implement adaptive real-time
VVO/CVR, two approaches have
been considered. One relies on capturing smart meters sensory data
through an interface with the MDM
system in the back office, then running VVO and CVR algorithms on
powerful enterprise servers using
the network model of the distribution system, and finally transferring
the new settings to the field VVO
and CVR assets through the SCADA
system. In other words, the VVO and
CVR functions are split into a clientserver configuration, with the server
operating in the back office and relying on MDM system databases to
continuously calculate new settings
for VVO and CVR clients in the field
and transfer the new configurations
to such assets through the SCADA
system. This approach is called centralized VVO/CVR control.
Centralized VVO/CVR control,
depicted in Figure 7, seemed quite
attractive at first. The availability of
accurate network models, combined
with adequate processing power on the
enterprise bus and access to the DMS
system, could indeed result in highly
effective settings for VVO and CVR
assets. But further studies at BCIT (see
Figure 8 and the suggested readings
that follow this article) indicated that
the VVO and CVR functions could be
performed a lot more efficiently (and
at lower cost) if on-demand sample
values of V/I/PF from bellwether smart
meters could be made available much
more frequently than they can be using
may/june 2014

may/june 2014

N5-2 Smart
Meters

D2

T2

Rooftop

B2

Smart Inverters

Net
Meter
N=3

PVs

Var Injection
Point-2

Bi

Var Injection
Point-3
Feeder
Capacitor
Bank

N5-1 Smart
Meters

T3

N=4

N=5

Net
Meter

T4

DK

EV
Net Meter
D1

DG Source

Var Injection
Point-4
Smart Inverter
Cj
C2
C1

N3 Smart Meters

N=1

Var Injection
Point-1

T1

N=2

B1

A8
A2

Communication Flow
Dist. Network Flow

Substation
Capacitor
Bank

Substation Transformer
with Tap-Changer

CVR
Server
IEC 61850

VVO
Engine
EMS
Server

HV/MV Substation

N2 Smart Meters

Feeder
Capacitor
Bank

A1

current AMI systems and their

MDM system interfaces. To achieve
this, VVO and CVR algorithms have
to be processed locally within each
substation, using the local sensory
data associated with each individual feeder. This approach is called
decentralized VVO/CVR control.
Although research in decentralized VVO/CVR control algorithms
is ongoing, early results indicate
that a decentralized approach is
more efficient and cost-effective for
such applications. The difficulty in
realizing a decentralized VVO/CVR
control strategy is that although it
requires the AMI system to supply
it more data more frequently, as a
substation-based function it has no
direct access to the AMI system.
This means that such data must be
extracted from the MDM system on
the enterprise bus and transported
down to the substation through the
SCADA system. The issue there is
that current AMI systems (which
interface directly with an MDM
system in the back office) are not
typically designed to supply such
large quantities of real-time data
from the field to the MDM system
without the risk of network congestion. Second, most SCADA systems
are incapable of transferring such
massive amounts of time-sensitive
information from the back office to
field devices without depriving other
critical functions of access to their
allocated bandwidth. Third, given
the fact that the VVO and CVR
functions are feeder-bound (i.e., the
required inputs and outputs are all
local), there is very little rationale
for involving upper-layer enterprise
functions in their operation.
This example is a clear indication of how critical a smart grid
integration map can be to the
realization of the smart grid. If a
utilitys integration map fails to
accommodate access to time-sensitive data for upper-layer-based
smart grid functions, it will either
have to give up implementing
future smart grid functionalities

figure 9. A substation-based EMS.

ieee power & energy magazine 63

or force its valuable assets (in this case, its communication

systems) to haul time-sensitive data back and forth across
various layers of the ICT hierarchy, thereby risking system
inefficiency and by design failure.
Another example of how critical forward-looking smart
grid integration maps are arises out of the requirement to
achieve tighter and more meaningful interfaces with customer-based assets. It is believed that in the not very distant
future, substation-based VVO and CVR functions will need
to extend their reach beyond smart meters and include some
coordination (or even command and control of) customerside generation assets and loads.
As depicted in Figure 9, such assets will include rooftop
PV modules and electric cars. In fact, recent reports about
the negative impact of uncoordinated rooftop solar cells on
the stability of feeder voltage levels are quite discouraging.
The unpredictable and intermittent behavior of such distributed generation assets cannot be entirely mitigated with utility field assets (e.g., capacitor banks). And even if such costly
assets could be effectively used to help stabilize voltage levels, their useful life spans (and health) could be considerably
compromised by these frequent anomalies (e.g., by voltage
levels pushing outside the ANSI band due to intermittent
generation from customer rooftop PV modules).
The voltage stability issues caused by electric cars used in
their vehicle-to-grid (V2G) mode may be far less severe than
those from rooftop PV modules, but this is still a problem for
which utilities need to make adequate provision. Even though
electric car manufacturers may not enable V2G functionality
for their cars for the foreseeable future due to their concerns
about the cost of battery warranties, utilities need to plan and
be ready for such issues should V2G become a reality.
What is interesting is that both of these threats could be
converted into opportunities for the utility if appropriate
provisions are made in the utilitys smart grid integration
map to take advantage of the availability of such downstream assets and integrate them with future substationbased EMSs. As depicted in Figure 9, such an EMS would
incorporate various command, control, and processing functions, using global system attributes and local feeder data to
configure all of its assets (inside and outside the substation)
so as to achieve its energy management goals.
Obviously, the demands such a level of integration would
place on the AMI system are even heavier than in the previous example. Here, the AMI system would work as the
conduit of communication and coordination between the
substation-based EMS engine and customer-owned cogeneration resources placed behind the meter. As such, it would
be critical to ensure that smart grid integration maps require
AMI systems to support such functionalities without major
engineering and overhaul.

Litmus Test
As discussed earlier, a forward-looking smart grid integration
map is critical for the realization of a smart grid. And given the
64

ieee power & energy magazine

cost involved in integrating new technologies and functionalities into the existing grid, the smart grid integration map could
prove to be either the savior or the Achilles heel of a utilitys
smart grid program. In making that judgment, every utility
has to review the operational requirements of its medium- and
long-term smart grid functions and determine if its smart grid
integration map supports a seamless transition from where it is
now to where it would like to be in the future.
In addition to the examples discussed at length above,
there are several other commonly identified smart grid capabilities that may be considered as a litmus test for ascertaining the suitability of a utilitys smart grid integration
map. These include:
Distributed generation: As discussed earlier, concerns about cogeneration synchronization, var control,
voltage stability, and so on have convinced utilities of
the need to achieve a level of integration (notwithstanding the regulatory impediments that exist in
various jurisdictions across North America) between
feeder assets and behind-the-meter, customer-owned
equipment. Given the fact that the point of common
coupling between the utility and the customer is the
smart meter, such a level of integration must be facilitated by the utilitys smart grid integration map.
Sensor networks on the low-voltage (LV) side of the
distribution system: Although such sensory data on the
LV side (such as those from phasor measurement units)
have not yet been established as a critical requirement,
one should assume that should that become a necessity,
the AMI infrastructure could be the primary means of
supporting such real-time data (through an auxiliary
channel) and transporting them to the substation. The
alternative to using the existing AMI assets for such
data would be to construct a dedicated, low-latency
communication system with a universal communication
protocol and mission-critical availability and resilience,
together with secure and intrusion-resistant multitier
access, as the carrier of such data for the upper layers of
the system. That could be quite costly. Again, no matter
what the chosen architecture for the implementation of
sensor networks, a utilitys smart grid integration map
must include provisions for supporting additional data
networks going forward.
Customer-side EMS: EMSs on the customer side
of distribution systems are often regarded as killer
apps enabling accurate, reliable, real-time, and endto-end energy management functions. Given the
trend toward designing distribution substations as
energy hubs in charge of achieving cost-effective
management of power and services transactions
between producers and consumers (prosumers),
it is paramount to move away from a broadcastbased, global utility pricing and tariff-signaling
system to a real-time, substation-based, local pricing signal. Just as the price of gas is never the same
may/june 2014

Enterprise Bus

Agent
BP20

Agent
BP11

Agent
BP21

Agent
BP00

Agent
BP20

Agent
BP01

Firewall

Agent
BP21

Agent
BP10

Agent
BP20

Agent
BP11

Transport: TCP/IP Over Fiber, Wi-Max, etc.

Agent
BP21

HAN

HAN

HAN

HAN

HAN

RF/PLC

Proprietary
Protocol

ZigBee

SEP 2.0

Operational Intelligent Agents

Geographical Intelligent Agents

Organizational Intelligent Agents

Distribution
Networks

Protocols: IEC 61850/CIM/Web Services for End-to-End Command and Control

Agent
BP21

Agent
BP10

Backhaul
Networks

Transport: TCP/IP Over Fiber, Wi-Max, Microwave, etc.

Agent
BP20

Firewall

figure 10. Integrated network domains in a smart grid.

Utility Data
Warehouse

Data

Enterprise Apps
Servers

DR Server

Outage
Management
Server

Dist Auto Server

Asset
Management
Server

Billing Server

MDMS Server

DAU
DAU
DAU

may/june 2014

ieee power & energy magazine 65

across all the gas stations in a town, so the price of

electricity should vary from substation to substation
in a given jurisdiction. In other words, every substation should be able to price its services based on a
host of local parameters such as load congestion,
the demand profile, and the energy available from
the grid and from prosumers. To achieve this, a utilitys smart grid integration map should facilitate the
required integration between the energy hub (the
substation) and its termination points (prosumers).

Conclusions
The central theme in all of the examples discussed in this
article is the need to have a forward-looking smart grid
integration map that empowers utilities to add incremental
functionalities to their existing grid, if and when required,
without the need to redo any of their previous investments.
Given the examples discussed abovewhich cannot by any
stretch of the imagination be considered comprehensive
the utilities have to be extremely careful about the initial
investments they make in this regard. That does not appear
to be always the case, however, as some of the choices that
have already been made in the early stages of the process
have not been encouraging.
The AMI model implemented in many jurisdictions
across North America, for example, relies on local data
collection units (often referred to as DAUs) as the primary
interface between smart meters and MDM system applications in the back office. In such a model, the local distribution substation will either be totally disconnected from the
AMI system that monitors the customers feeding off its feeders or if there is any communication between smart meters
and substation equipment, the data will have to go through
the round robin of being captured by DAUs locally, passed
on to the appropriate MDM system in the remote back office,
and handed over to the SCADA head end in the back office
before finding its way through the SCADA network from the
back office down into the substation.
It goes without saying that such long delays in data and
command communication would make it nearly impossible
to efficiently run any number of smart grid capabilities
that rely on distributed command and control and as such
require local analytics and decision making. Such applications are by default substation-resident, with a stringent
need for unimpeded access to real-time data from smart
meters, sensors, and other termination points associated
with that substation. In other words, smart meters should
ideally be substations over-the-fence intelligent electronic devices (IEDs), fully engaged in real-time data and
command exchange with substation-resident functions;
failing this, they are nothing more than an interim solution
for automating billing and revenue management.
Finally, a utilitys smart grid integration map must support
the realization of the utilitys integrated network domains,

66

ieee power & energy magazine

as depicted in Figure 10, which emphasizes the need for a

distributed command and control system (using a system of
intelligent agents) running across multiple domains of the
utility network and providing end-to-end communication
and data exchange among all utility assets. In that regard,
no single smart grid asset should be planned as fulfilling an
outlying function divorced from the utilitys existing and
planned operations and capabilities. If it is, one can seriously
doubt the business justification for acquiring such expensive
assets, as well as that utilitys ability to actually implement
cost-effective and efficient smart grid capabilities.

For Further Reading

U.S. Department of Energy. (2012, Dec.). Application of
automated controls for voltage and reactive power management, initial results. [Online]. Available: https://www.
smartgrid.gov/sites/default/files/doc/files/VVO%20Report%20-%20Final.pdf
U.S. Department of Energy. (2012, Mar.). Visioning the
21st century electricity industry: Strategies and outcomes
for America. [Online]. Available: http://energy.gov/sites/
prod/files/Presentation%20to%20the%20EAC%20-%20Visioning%20the%2021st%20Century%20-%20William%20
Parks.pdf
M. Nasri, H. Farhangi, A. Palizban, and M. Moallem,
Application of intelligent agents in smart grids for volt/VAr
optimization and conservation voltage reduction, in Proc.
IEEE Canada Electrical Power and Energy Conf. London,
Ontario, Oct. 2012.
M. Manbachi, H. Farhangi, A. Palizban, and S. Arzanpour, Real-time adaptive optimization engine algorithm for
integrated volt/VAr optimization and conservation voltage
reduction of smart microgrids, in Proc. CIGR Canada
Conf., Montreal, Sept. 2012.
H. Farhangi, Smart grid and ICTs role in its evolution, in
Green Communications: Theoretical Fundamentals, Algorithms
and Applications, J. Wu, S. Rangan, and H. Zhang, Eds. Boca
Raton, FL: CRC Press, 2012.
G. Stanciulescu, H. Farhangi, A. Palizban, and N.
Stanchev, Communication technologies for BCIT smart
microgrid, in Proc. IEEE PES Innovative Smart Grid Technologies Conf., Washington DC, Jan. 2012.
M. Manbachi, M. Nasri, B. Shahabi, H. Farhangi, A. Palizban,
S. Arzanpour, M. Moallem, and D. C. Lee, Real-time adaptive VVO/CVR topology using multi agent system and IEC
61850-based communication protocol, IEEE Trans. Sustainable Energy, vol. PP, no. 99, p. 1, Oct. 2013.

Biography
Hassan Farhangi is with the British Columbia Institute of
Technology, Vancouver, Canada.

p&e

may/june 2014