Smart grid integration of IoT

Andrei Cornel – Cristian, Tudor Gabriel, Mădălina Călin-Arhip, Alexandru Zamfirescu

Department of Power Engineering Systems
Polyethnic University of Bucharest – Faculty of Power Engineering
Bucharest, Romania

Abstract – With the development and introduction of

intelligent electronic devices within the classical
electrical grids and their transformation into smart
grids, emerged the need for extracting and processing the
data from those devices in real time or near real time. In
addition to information extraction, it is necessary for
field devices to be able to communicate with each other
as well as with a central command point. By integrating
internet of thing (IoT) within smart grids communication
between devices that normally were not designed to
exchange information becomes possible, information
that can be accessed from anywhere with Internet access.
This paper will explore the benefits brought by IoT Fig. 1 – Classical electrical grid [2]
application to the smart grid.
The identification and bypass of the equipment or area
Keywords – smart grid, internet of things, intelligent affected by the fault can be achieved through the use
electronic devices, big data, prosumer. of automatic topology reconfiguration system and
electrical fault system detection, which can be
I. INTRODUCTION initialized by a supervisory control and data
acquisition system (SCADA) or it can operate
The first attempts used to shift from the classical
independently to the SCADA system and as result the
electrical grid to the smart grid was the
time of non-supply with power of users will be
implementation of an automated meter reading (AMR)
reduced considerably.
system, which allowed remote reading of power
Unlike the old electrical grid where the flow of
meters installed in the power distribution system.
power from power generation units to consumers was
However, this type of system did not allow the
a direct path in the smart grid vision power generation
application of corrective measures if some restrictions
units are distributed throughout the grid to prevent
or energy quality standards were not respected, and the
overloading of certain areas and deficits in others.
communication in this case was done only between the
equipment and the distribution system operator
(DSO), the end user was not a part of the information
exchange. In order to allow a better review of the
power quality and power flow through the power
distribution system it was necessary to upgrade from
AMR to an advance metering system (AMI) that
allows access to the power meter for both the DSO and
the end user [1].
For a complete transition to smart grids is not
enough only to implement smart meters and their
integration in an advanced metering infrastructure, it
is also necessary to incorporate monitoring and control
systems for every part of the power grid. The Fig. 2 – Smart grid concept [2]
implementation of these systems is required in order
Even more today’s power grids must present a
to identify and bypass fast and punctual any faults that
flexible and dynamic character in order accommodate
occurs at the equipment level in the power grid.
new challenges generated by the smart devices,

978-1-7281-3349-2/19/$31.00 ©2019 IEEE

Authorized licensed use limited to: Universidad Pais Vasco. Downloaded on October 11,2021 at 13:16:14 UTC from IEEE Xplore. Restrictions apply.
ecofriendly products and distributed generation. A few interconnection between a nonIP protocol
of these challenges are represented by integration of and an IP-based protocol.
charging station for electrical cars and grid users who
have installed on their properties power generation With increase in the number of embedded devices
units, mostly using renewable sources, that can supply in smart grids, the amount of data they generate
locally power if their peak load demand is under the exceeds the ability of current software and hardware
generated power. tools to process and analyze. In addition, the data
collected from field is not relevant to all participants
II. IOT AND BIG DATA in the power system in the same way. For example, the
data received from the distribution power system
Internet of things offers. the possibility to connect about the voltage value in the connection point at user
to internet of devices that were not normally designed level is totally irrelevant at the transmission power
with this option. In addition, IoT enables information system level. In the transmission power system is
exchange between different equipment from different necessary to maintain the balance between power
vendors by interconnecting them via the Internet [3]. consumption and generation, balance that dictates
The ability to interconnect equipment from different evolution of power grid frequency.
manufacturers without the need for proprietary If IoT enables physical objects to connect to
solutions provides smart grids a fast and efficient way internet and exchange data between them, the way that
to collect information from the multitude of sensors data sets are structured and analyzed enters the field of
and smart electronics integrated into them. big data. Big data is a concept that can be defined as a
Devices within IoT data can be distinguished in large volume of data, of petabyte order, that requires
three categories depending on the mode of operation processing and organizational abilities beyond the
and the complexity of the generated as follows: capacity of usual technologies [3]. Big data can be
described by five parameters as follow [4,5]:
• Complex devices that are always
connected to the Internet, not limited in • Volume that, as the name suggests,
terms of data traffic or power usage, and refers to very large amounts of data that
generate complex data sets needed to cannot be stored and analyzed by
control from a few hundred to several common computer systems, requiring
thousand parameters; the use of distributed computing systems
• Simple devices mostly represented by connected to each other;
sensors and small equipment both in terms • Velocity that defines how new data is
of physical dimensions and computational generated and how fast it is transmitted
abilities that are used to control a small from one point to another;
number of parameters. Typically, these • Variety refers to the type of data and the
types of devices are powered by batteries way it is classified. Data can be
and are limited in terms of power usage and structured meaning information is
data traffic due to use of a mobile data plan organized according to certain criteria or
for communication, data is transmitted unstructured in the form of text, images
only if the parameters change, for the rest or audio / video files where information
of the time the device is remaining in an is in a raw form;
energy-saving mode; • Veracity refers to the quality and
• Devices that require a gateway to connect accuracy of the information taken; it is
to Internet. This kind of devices use necessary to validate the raw data before
protocols that are part of IoT such as creating a report based on it;
ZigBee or Z-Wave but are not directly • Value refers to the ability to extract
connected to the Internet, these protocols important information from a large
being used only for communication and amount of data and generate reports to
control at the local level. In order to determine what actions are needed to
connect these devices to the Internet, it is take.
necessary to use a transfer medium, a
Big Data can provide the smart grid with the ability to
gateway, which enables the
analyze and generate reports on which to make

Authorized licensed use limited to: Universidad Pais Vasco. Downloaded on October 11,2021 at 13:16:14 UTC from IEEE Xplore. Restrictions apply.
decisions and allow power grid control in real time or Edge computing can improve the efficiency of a
near real-time. smart house or microgrid management system by
analyzing historical data and learning consumer
III. BUILDING A SMARTER GRID WITH preferences through application of artificial
IOT AND BIG DATA intelligence and machine learning techniques. There
are two types of learning techniques that can be used
A. Edge Computing [6]:

The large amount of information generated by • Supervised which uses historical and
embedded IoT devices in smart grids can prove to be behavioral data of the target system to
a real challenge for common cloud computing generate prediction or by modeling
systems. One of the reasons why the cloud computing objects from the target system by using a
servers cannot meet the processing requirements is set of predefined labels in order to
their positioning on different geographic areas, where determine the state of every component
it is necessary to use multiple connection points for from the system;
receiving and transmitting data which introduces a • Unsupervised in which uses models that
considerable delay in communication. The other separate data in subgroups based on
reason is the additional load that raw data transmitted common points of interest, or uses
from IoT devices generates on the communication decision-making models based on a trial
network, introducing an unstable response time. A and error system type where each taken
large amount of the raw date is removed in the decision is evaluated in order to decide
validation process therefore sending data in raw how correct it was, the following
format towards the cloud computing servers takes time decisions will improve based on this
and bandwidth that can be used for other purpose. In evaluation.
the case of a power grid where it is necessary to take
In a real application these techniques can used to
decisions as fast as possible a delay in receiving
predict power usage in smart home based from user’s
critical information can lead to problems in its
behavior or to determine power generation from a
operation and control [6].
photovoltaic plant embedded in microgrid based on
Therefore, edge computing eliminates the
historical data about nebulosity and radiance.
deficiencies presented above by moving the data
processing to the location where they are generated. B. IoT protocols
This away the communication network is no longer
loaded with raw data, the data transmitted to a central The promoters of IoT development are COTS,
point or to a database are fully processed or semi- Commercial Of The Shelf, components. These
processed and the response time becomes acceptable components came into existence due to developments
enough to take decisions. In the edge computing in RFID (Radio Frequency Identification) technology,
environment tasks that were attributed to the cloud which allows devices from the IOT to be tracked and
environment are distributed across layers through the managed by computers using an electronic tag.
IoT infrastructure. So if f one of the lower layers data The rise in FOSH, Free Open Source Hardware,
processing proves to be ineffective, the corresponding created an important market for manufacturers to build
operation is sent to the next layer where more powerful inexpensive components for IoT. This resulted in
computational equipment is present [6]. accessible price ranges for devices, in return, the
Edge computing does not intend to replace the added value is given by the software which is
cloud computing environment but only to eliminate developed for the hardware to enhance its capabilities.
some deficiencies by offloading a part of the tasks The mobile infrastructure has enabled the rise of
performed by cloud computing by processing, a culture that revolves around highly performant
analyzing, and generating reports at a local level. The gadgets which embody accessible but powerful
cloud and cloud computing environment being further processors. These advancements allowed app
used to store and centralize the information received
and generate predictive models based on the received
data.

Authorized licensed use limited to: Universidad Pais Vasco. Downloaded on October 11,2021 at 13:16:14 UTC from IEEE Xplore. Restrictions apply.
 Fig. 4 - ISO Stack [7]
Fig. 3 - CIM Pyramid [7]
limitations from a connectivity perspective of the
development that merges consumer and services
devices and the nodes linking the smart object
throughout the Telecommunications infrastructure.
networks together [8].
Industry 4.0 makes the most from enhancements in the
At the physical link level raw data comes from the
communication technology sector by connecting or
measuring devices. At this level we have limitations in
merging production with information, make use of
the IoT device, such as memory and CPU power. For
customer data by merging it with machine data,
interfacing level one devices, Fieldbus is used at the
efficiently exploit the communication capability
factory level due to the low processing capabilities of
between machines and making the production process
the IoT devices and to reduce the amount of wiring
increasingly autonomous, efficient and resource-
required. In the IoT Scenario we have IEEE 802.15.4
saving.
Zigbee standard in the Physical/Data-Link layers, low-
With the development o high speed computing
power Wi-fi, Bluetooth Low Energy (BLE) and
and intelligent electronic devices supervision and
Protocols in the area of power line communications
control of a smart grid can be realized with little or no
(PLC).
human intervention. This approach is defined by
On the network layer real-time transmission of
computer integrated manufacturing (CIM) concept.
data takes place. This is where the Control Bus is
CIM is a technology that allows systems to function in
located. IP-based IoT devices employ IPv4 and IPv6.
a layered structure. In CIM sensors, transducers and
Due to the depletion of IPv4 addresses we have a
actuators, interfaced by the fieldbus, detect the signals
transition towards IPv6 addressing which provides
which sends data in real time that reaches the
more possibilities in terms of unique device
processing equipment (Remote Terminal Units). The
addressing. At this level Low-power Wireless
processed data is gathered and presented to operators
Personal Area Networks are found (WPANs,
of SCADA systems in order to make informed
6LoWPAN) respectively. At this level Edge
decisions on operating the systems. The available
Computing is defined.
information is sent using the corporate network and is
On the transport layer we have TCP,
organized by MES (Manufacturing Execution
Transmission Control Protocol, and UDP, User
Systems) that bridges the gap between production and
Datagram Protocol. UDP is the protocol of choice for
management and delivers information to the next
IoT because it allows for devices with low power
phase of the process which is ERP (Enterprise
consumption to use it. The downside to this is that
Resource Planning) for daily activities.
retransmission, ordering and congestion control are
Network communication is standardized and
not available and need to be incorporated into other
described by the OSI model (Open System
layers.
Interconnection). The OSI model was developed by
On the Application Layer the connectivity of the
the International Organization for Standardization
low-power and lossy networks (LLNS) is done using
(ISO).
Constrained Application Protocol (CoAP). CoAP is a
Connecting smart objects needs to take into
web-based protocol which mimics HTTP. Request-
consideration the signal propagation from the device
response communication takes place and an example
(range), the frequency bands, power consumption,
of this in action is the MQTT, Message Queuing
topology (mesh, star, peer-to-peer topologies) and
Telemetry Transport, protocol.

Authorized licensed use limited to: Universidad Pais Vasco. Downloaded on October 11,2021 at 13:16:14 UTC from IEEE Xplore. Restrictions apply.
 Fig. 5 - The IoT reference model published by IoT World forum [9]

Most of the protocols used in IoT are striped down With all this information sent in all direction and
version of current in use protocol like TCP/IP or at one click away it is necessary to implement security
derived from it. This way is created new lightweight measures to determine who has access to it and why it
protocols that are suitable for in the new vision of IoT needs clearance to access it, other way anyone could
and smart grid crossing. The need for lightweight tamper critical data. Access control is important
protocols is required mostly because the IoT because it determines the way that users interacts with
embedded devices are constrained by power usage and information based on their clearance level.
bandwidth use. Hence a lightweight protocol assures
that the data is transmitted from one point to another REFERENCES
has little or no impact on both the device and the
communication link. The biggest advantage of these [1] A. B. M. Shawkat Ali, ”Smart Grids Opportunities,
Developments, and Trends”, Springer-Verlag London, 2013
protocols is that they are open source, meaning that [2] Indrajeet Prasad, “Smart Grid Technology: Application and
anyone knowledgeable enough can improve them. Control”, IJAREEIE,Vol. 3, Issue 5, May 2014, in press;
[3] David Hanes, Gonzalo Salgueiro, Rob Barton, “IoT
IV. CONCLUSION Fundamentals: Networking Technologies, Protocols, and Use
Cases for the Internet of Things”, Cisco Press , 2017;
[4] Yang Zhang, Tao Huang, Ettore Francesco Bompard, “Big
Embedding IoT devices and technologies in the data analytics in smart grids: a review, Energy Informatics”,
smart grid assures that data from every point in the grid Springer International Publishing, 2018
is generated and transmitted fast enough to a central [5] Valentina Emilia Balas, Vijender Kumar Solanki, Raghvendra
Kuma, Manju Khari, “Internet of Things and Big Data
command point located on the power distribution
Analytics for Smart Generation”, Springer International
system operator premises. And in combination with Publishing, 2019;
edge computing and the use of artificial intelligence [6] Farzad Samie, Lars Bauer, Jörg Henkel, “Edge Computing for
and machine learning the data received from the Smart Grid: An Overview on Architectures and Solutions:
Design Challenges and Paradigms”, IoT for Smart Grids,
embedded device is used to compile prediction and to
pp.21-42, Springer International Publishing, 2019:
offer the state of the power grid in real time or near [7] Antonio Capasso; Giacomo Veneri, “Hands-On Industrial
real time. Internet of Things”, Packt Publishing, 2018:
The main challenges are the integration of old and [8] Simone Cirani, Marco Picone, Gianluigi Ferrari and Luca
Veltri, “Internet of Things: Architectures, Protocols and
legacy equipment in IoT that is still present in the
Standards”, Wiley & Sons, 2019:
current power grids, that cannot by simply replaced by [9] Gonzalo Salgueiro; David Hanes; Rob Barton, “IoT
newer ones because it involves certain risks for the Fundamentals: Networking Technologies, Protocols, and Use
power grid stability. Cases for the Internet of Things”, Cisco Press, 2017

Authorized licensed use limited to: Universidad Pais Vasco. Downloaded on October 11,2021 at 13:16:14 UTC from IEEE Xplore. Restrictions apply.