Measurement 129 (2018) 479–488

Contents lists available at ScienceDirect

Measurement
journal homepage: www.elsevier.com/locate/measurement

A new PLC-based smart metering architecture for medium/low voltage T

grids: Feasibility and experimental characterization
⁎
Giovanni Artalea, Antonio Cataliottia, Valentina Cosentinoa, , Dario Di Carab, Riccardo Fiorellic,
Salvatore Guaianaa, Nicola Panzavecchiab, Giovanni Tinèb
a
Department of Energy, Information Engineering and Mathematic Models (DEIM), Università di Palermo, Viale delle Scienze, 90128 Palermo, Italy
b
Institute of Marine Engineering, National Research Council of Italy, Via Ugo La Malfa, 153, 90146 Palermo, Italy
c
ST Microelectronics S.r.l., Agrate 20864, Italy

A R T I C LE I N FO A B S T R A C T

Keywords: This paper investigates the feasibility of a smart metering architecture for modern power grids in the smart cities
Smart grids framework. A particular focus is made on Automatic Meter Reading, with the aim of facing some costs and safety
Smart metering drawbacks of current solutions (such as dependence from wireless communication providers, reliability pro-
Automatic Meter Reading blems in bad weather conditions, and exposition to cyber-attacks). The proposed architecture exploits power line
Distributed measurement systems
communications (PLC) at both low and medium voltage level, thanks to new devices and an innovative medium
Power systems communication
voltage PLC coupling system. The new coupling solution and the whole smart metering architecture are ex-
Power line communication
perimentally verified in a wide frequency range, from CENELEC A band up to 200 kHz, by using different
modulation techniques; the on-field measurement campaign has been carried out on the distribution network of
Favignana Island (Mediterranean Sea). Success rate and RTT measured in real environment confirm the feasi-
bility of the proposed solution.

1. Introduction distributed storage systems (DSS), by a proper control of charging and

Smart electric power grids undoubtedly represent a key element in
the smart cities framework, playing a crucial role for reliability, sus-
tainability and resilience in several fields (energy production, dis-
tribution and use, energy efficiency, renewable energies exploitation,
smart buildings and transportation, and so on). The transition towards
smarter cities and communities requires a deep transformation of dis-
tribution electricity networks, at both low voltage (LV) and medium
voltage (MV) level, where a large quantity of energy users, producers
and prosumers are connected. In fact, traditional power grids were
characterized by unidirectional energy flows, from few production sites
to end passive users. On the contrary, in the smart cities framework,
modern distribution grids must be able to cope with new paradigms for
energy production, distribution and use, providing the technical back-
bone for smart energy applications [1,2]. In fact, the widespread dif-
fusion of distributed generation (DG) from renewable energy sources,
lead power grids facing new bidirectional energy flows and random
variations in the electrical quantities, due the intrinsically intermittent
and unpredictable production profiles [3]. For real-time balancing en-
ergy supply and demand, a fundamental role can be played by
discharging periods and power exchange with the grid [4]. Also elec-
trical vehicles (EV) can act as support for electricity storage and load
shaping, by means of their optimal management in grid-to-vehicle
(G2V) and vehicle-to-grid (V2G) paradigm [5,6]. In this framework,
smart electricity users and prosumers are involved in new scenarios for
energy metering and management. Energy demand is expected to be-
come extremely flexible, under demand side management, demand
response and dynamic energy price programs [7–9]. New smart stra-
tegies are expected to be implemented for aggregations of flexible en-
ergy resources and load demands, at different levels (homes, buildings,
virtual power plants, micro-grids) [10–12].
In order to effectively enable the aforementioned scenarios, an in-
creasing amount of measurement data and information must be ac-
quired, transferred, processed, managed and stored. Thus, advanced
information and communication technologies capabilities must be im-
plemented to allow proper monitoring, management, automation and
control features, ensuring safe and effective grid operation [13–16]. In
this framework, advanced smart metering and distributed measurement
infrastructures are key elements for a smarter power grid management,
allowing new modalities of interaction among loads, DG and DSS

⁎
Corresponding author.
E-mail addresses: giovanni.artale@unipa.it (G. Artale), antonio.cataliotti@unipa.it (A. Cataliotti), valentina.cosentino@unipa.it (V. Cosentino),
dario.dicara@cnr.it (D. Di Cara), salvatore.guaiana@unipa.it (S. Guaiana), panzavecchia@pa.issia.cnr.it (N. Panzavecchia), giovanni.tine@cnr.it (G. Tinè).

https://doi.org/10.1016/j.measurement.2018.07.070
Received 1 June 2018; Received in revised form 23 July 2018; Accepted 24 July 2018
Available online 25 July 2018
0263-2241/ © 2018 Elsevier Ltd. All rights reserved.
G. Artale et al.
Fig. 1. Present smart metering architecture based on a wireless and LV PLC hybrid communication solution.
(including those of EV) and the implementation of automatic meter
reading (AMR), demand side management and other smart grid func-
tionalities (real time power flows analysis, islanding detection, fault
detection, power quality assessment and so on) [17–25]. A widespread
diffusion of smart metering technologies is needed to acquire in-
formation on both electrical and non-electrical quantities (powers,
voltages, currents, switches status, distributed generators power pro-
duction, remote commands, security or safety warning signals, and so
on). All these data must be exchanged among all different smart grid
players: distribution system operators (DSOs), users, prosumers, etc.;
thus a pervasive communication network is needed, which must be
reliable, cost-effective and easy to install, in both MV and LV dis-
tribution networks [26–28].
In the aforesaid scenario, AMR systems play a fundamental role for
grids management and the implementation of new interaction strate-
gies between DSOs and users/prosumers. Currently, in many countries
(such as in Italy), AMR systems are already implemented. A scheme of
existing smart metering architecture is shown in Fig. 1. For each LV user
or prosumer, a smart meter provide the information on the energy ex-
changes. In each secondary substation, an AMR concentrator collects
and stores measured data coming from the LV smart meters installed
downstream the substation itself. The AMR server is installed in DSO
control center and it queries the AMR concentrators periodically
(usually once a day), to process and store the collected data (mainly for
billing purposes). Usually, the data connection at LV level, between
AMR concentrators and smart meters, is obtained via power line com-
munications (PLC), while AMR server and concentrators communicate
by means of a wireless modem (typically GPRS) which are widespread
available in urban contexts. Thus DSO copes with economical and in-
frastructural dependence from a wireless communication provider.
Furthermore, such systems can have weak reliability (particularly in
480
Measurement 129 (2018) 479–488
bad weather conditions) and they can also be exposed to cyber-attacks.
To overcome the aforesaid problems, a possible alternative for AMR
communication can be represented by the use of PLC at both MV and LV
level. This solution has several advantages: it has low installation cost,
as power lines are already present; the PLC channel is totally under DSO
control, thus commercial and technical dependence from communica-
tion providers can be avoided; PLC is also more secure from cyber-at-
tacks, as the communication system is not easily accessible from an
intruder. Moreover, PLC can be a suitable solution for last-mile appli-
cations or for reaching smart meters in peripheral urban or rural con-
texts, where other communication links can be not available [28–32].
As mentioned before, PLC is already used in distribution networks at LV
level for AMR and it has been also proposed for DG applications
[13,33]. On the other hand, PLC use at MV level has been investigated
in the last years [34–39]. Some authors have verified that PLC signals
cross power transformer, even if attenuation are expected in depen-
dence of the selected frequency and modulation technique [37,38],
especially in the CENELEC band. Further attenuations due to line length
and derivations can reduce the signal amplitude below noise level, thus
preventing its correct reception. Thus commercial couplers are usually
suggested in substitution or in combination with PLC signal crossing
techniques, to increase signal level in points of MV network where
higher attenuations are expected. On the other hand, the use of MV
commercial couplers generally entails high costs for both equipment
and installation, including those of service interruption. As regards this,
the authors have patented a PLC coupling solution for MV networks,
based on the use of capacitive dividers of voltage detecting systems
(VDS), normally used for revealing the mains voltage presence [40–43].
The patented solution can overcome the limitations of using PLC at MV
level, as it entails low equipment and installation costs, also avoiding
service interruptions.
G. Artale et al.
Fig. 2. Proposed smart metering architecture with low cost and easy to install PLC communication solutions both on LV and MV distribution networks.
Starting from this, a complete smart metering architecture for smart
power grids is proposed in this paper, which is based on the use of PLC
at both MV and LV level. An essential element of the solution herein
proposed is the innovative MV PLC coupler. To verify the feasibility of
the proposed smart metering architecture, an on field experimental
validation is performed in the real case of the distribution network of
Favignana Island (Mediterranean Sea). The paper is organized as fol-
lows. Section 2 presents the general concept of the proposed smart
metering architecture. Section 3 describes the experimental on-field
characterization of the MV-PLC channel. Starting from this, Section 4
describes the experimental on-field set-up of the PLC-based smart me-
tering architecture and the related test results.
2. Proposed smart metering architecture
The proposed smart metering architecture is shown in Fig. 2. Si-
milarly to existing solutions, each LV user or prosumers has a smart
meter. LV AMR concentrators are installed in secondary substations, but
they are connected to MV PLC transceiver boxes (instead of wireless
modems). To this aim, the innovative MV coupling solution, described
in the following, is used to couple the PLC signal to the MV power
network. A MV concentrator (MVC) is installed in the primary substa-
tion; it collects the measurement data of all the related secondary
substations. These data can be used in computation algorithms typical
of new smart city applications. As an example the data can be grouped
to act in the energy market trading as aggregator of the connected users
and prosumers. The MVC are then connected to internet by using pri-
mary substations communication links (LTE wireless or fiber optic).
Finally, DSO control center can acquire measurement data and control
actuators, through the internet connection.
As regards the MV PLC coupler, the solution based on the use of the
VDS capacitive divider is proposed [40]. According to IEC 61243-5
481
Measurement 129 (2018) 479–488
[44], VDSs are worldwide used to reveal the mains voltage presence,
ensuring operators safety. VDS capacitive divider consists in a series of
two capacitances connected between MV bus-bars and the earth. The
intermediate node of capacitance series is made available on a socket in
the MV switchboard panel. When mains voltage is applied at capacitive
divider terminals, a reduced proportional voltage is obtained at socket
terminals. Thus, a voltage indicator, usually a flashing light, can help in
easily detecting the mains voltage presence in the switchboard [44].
The new coupling system uses this socket to inject and receive the PLC
signal. A proper electronic interface card is designed to make the ca-
pacitive divider operate as PLC coupler, having high impedance at low
frequency (thus isolating the transceiver from mains voltage) and low
impedance in the signal frequency band. The interface card bases its
operation on creating a resonant behavior with the VDS capacitances.
To this aim, internal parameters, principally variable inductances are
tuned based on measured VDS capacitances and desired operating fre-
quency range [41]. The proposed interface card is cheaper than com-
mercial MV couplers; moreover, it avoids MV switchboard modifica-
tions, service interruption for its installation and consequent manpower
costs (as the interface card can be safely connected to low voltage VDS
socket).
3. Experimental on-field characterization of the MV PLC channel
To verify the feasibility of the MV PLC solution in a real environ-
ment, an on field experimental characterization was carried on the MV
and LV distribution network of Favignana Island. More in detail,
Favignana electrical network is composed by three MV feeders de-
parting from a diesel power plant, with a total capacity of 12 MW. The
network is isolated from the national grid. Both MV cables and over-
head lines are installed in the three feeders which connect 40 MV/LV
secondary substations. The experimental tests were carried on the MV
G. Artale et al.
Fig. 3. Portion of Favignana distribution network under test: the MV line connecting the two secondary substations “4 Vanelle” and “G. Di Vita” is highlighted in red;
the LV line departing from substation “4 Vanelle” is in green.
Fig. 4. Experimental set-up for the on-field tests of the MV PLC channel between “Gen di Vita” (6) and “4 Vanelle” (9) substations.
feeder which connects 29 secondary substations. The electrical loads
are those typical of touristic and residential districts. Rated data and
some load measurements of the considered network are reported in
[50]. A map of the MV and LV distribution network under test is shown
in Fig. 3. In more detail, the experimental on-field site consists of two
substations, connected by a 1.1 km long MV line, which is made up by
unipolar cables RG7H1R type with 50 mm2 aluminum core cross-sec-
tion and copper shield (in Fig. 3, the MV line route is highlighted in red;
the green route indicates the LV line, L1, used for the tests described in
Section 4). The first substation under test is named “Gen. Di Vita”. It is
in by-pass configuration with three MV switches, connecting the MV
bus-bars to the incoming line, outgoing line and MV/LV power trans-
former (160 kVA), respectively. The second substation, named “4 Va-
nelle”, is a nodal substation with five MV switches for the incoming line,
MV/LV power transformer (250 kVA) and three departing MV lines
supplying further secondary substations. Furthermore, three LV lines
depart from the MV/LV power transformer (one of those is the afore-
mentioned L1, indicated with the green line in Fig. 3, see next section).
A configurable PLC modem with the MV VDS PLC innovative cou-
pler was installed in both substations (“Gen. Di Vita” and “4 Vanelle”), to
inject and receive the PLC signal. The configurable PLC modem and
tunable interface card were assembled into a prototypal box (one for
482
Measurement 129 (2018) 479–488
each substation, see Fig. 4). A modified plug was used to connect the
tunable interface card to the VDS socket at each end of the MV line,
with no power supply interruption and no MV switchgear modification
[43]. Two ST7580 modems were used to transmit nPSK modulated PLC
signals. All on-field tests were performed in the presence of mains
voltage (i.e. 20 kV, 50 Hz).
The MV PLC connection based on the new coupling solution was
tested in three steps:
• firstly, the channel frequency response was used to tune interface
card parameters, i.e. to obtain a resonant but enough flat frequency
behavior around the chosen center frequency;
• secondly, the noise spectrum was measured at the interface card
output, with the PLC modem switched off;
• thirdly, PLC transmission tests were carried out to verify the channel
capacity in term of success rate at different modulation techniques.
The aforementioned steps allow properly setting the interface card
parameters, in order to minimize the impact of stray parameters and
noise on the PLC signal transmission. In fact it should be noted that the
resonance condition between interface card and VDS equivalent capa-
citance is site dependent, being unknown both MV switchboard and
G. Artale et al.
Fig. 5. MV PLC transmission channel frequency response during an on-field sweep test around the center frequency of 78 kHz.
Fig. 6. MV PLC transmission channel frequency response during an on-field sweep test around the center frequency of 150 kHz.
stray capacitances. Thus a tuning procedure is always necessary at first
installation of the interface card in a new site. The channel frequency
response is obtained on-site, by sweeping a sinusoidal signal in the
frequency band under study (50 kHz of SPAN around each considered
center frequency) and analyzing the FFT spectrum of the received
signal. In the experimental tests herein presented, a digital oscilloscope
was used to get the measurements with high input impedance and the
FFT spectrum of the received signal was analyzed with a frequency
resolution of 300 Hz. The frequency response test allows setting the
interface card parameters to have a suitable frequency band, depending
on the chosen modulation. As an example, in the case of nPSK mod-
ulation and 9600 baud/s of symbol rate, a target 6 dB bandwidth of
15 kHz around the center frequency is needed. Thanks to the use of VDS
socket and the high impedance of the interface card at power system
frequencies, the on-field tuning procedure can be carried out in the
presence of mains voltage, thus no service interruptions are needed
during the interface card installation.
In comparison with previous tests reported in [42], the on-field test
483
Measurement 129 (2018) 479–488
campaign herein presented was aimed at demonstrating the effective-
ness of the proposed solution in a wider frequency range, from 78 to
200 kHz. This frequency interval is considered for two reasons: trans-
mission frequencies higher up to 200 kHz are allowed outside Europe
(FCC band in United States, ARIB band in Japan, and so on); nPSK
commercial PLC transceivers normally have a maximum frequency of
200 kHz.
As shown in the next subsections, thanks to the interface card
variable components, a resonant behavior is obtained with a flat fre-
quency response around each different center frequency. Frequency
response and noise analyses show that the increase in frequency leads
to lower noise and larger bandwidth, thus obtaining a better trans-
mission behavior. This is also demonstrated in the transmission tests.
The obtained results confirm the feasibility of the proposed NB-PLC
coupling solution for different center frequencies and modulation
techniques. Thus the proposed solution can be efficiently used to
transmit narrowband modulated signal in a wide frequency range,
reaching a transmission data rate up to 28.8 kbit/s with a high success
G. Artale et al.
Fig. 7. Noise measured in the frequency range 61–111 kHz.
Fig. 8. Noise measured in the frequency range 125–175 kHz.
rate.
3.1. Fequency response
The channel frequency response was obtained by generating a
sweep signal from one substation and measuring the spectrum of re-
ceived signal in the second substation. Thus, the transmission channel
included interface cards, VDS capacitive dividers, MV cable line and all
connected electrical elements, i.e. power transformer, MV switchboards
and departing lines. On-field tests were performed injecting a 1.5 VPP
amplitude signal with frequency span of 50 kHz around selected center
frequencies, assuring that the signal amplitude was less than 134 dBµV,
according to CENELEC [45]. A digital oscilloscope Rohde & Schwarz
RTO 1044 was used to obtain the FFT spectrum of the signal at re-
ceiving interface card output, with a frequency resolution of 300 Hz.
Thanks to the frequency response tests, the interface card parameters
484
Measurement 129 (2018) 479–488
were adjusted in order to find the resonant behavior and to obtain the
desired bandwidth at each center frequency. As an example, the fre-
quency spectra measured at the interface card output for two center
frequencies (78 and 150 kHz, respectively) are shown in Figs. 5 and 6.
The results show a resonant behavior of the PLC channel around the
respective center frequency. Moreover, an increase in center frequency
corresponds to a decrease in received signal amplitude and an increase
in channel frequency band. This is mainly due to the resonant circuit
losses, whose effects are more evident at higher frequencies. A flat
frequency response was more difficult to be achieved on-field than in
laboratory tests [42]. In fact, real network MV lines, power transfor-
mers and metallic MV switchboards introduce additional parasitic ef-
fects. Thus, in comparison with laboratory tests, a tighter frequency
band was obtained, especially at low frequencies. The received signal
amplitude was about 20 dB lower than values measured in laboratory.
This is mainly due to the attenuation introduced by MV line, power
G. Artale et al. Measurement 129 (2018) 479–488

Table 1 understanding which frequency range can guarantee a more reliable

VDS communication tests. transmission.
Frequency [kHz] Modulation Bit rate [bit/s] Success rate [%] In the proposed test site, noise measurements were performed in
both substations, analyzing the acquired signal spectrum at the inter-
78 BPSK coded 4800 100 face card output, with a frequency span of 50 kHz around each analyzed
QPSK coded 9600 97.8
center frequency. During noise measurements, interface card para-
BPSK 9600 95.1
QPSK 19,200 94.1
meters were kept equal to those selected during frequency response
tests. Two measured spectra are shown in Figs. 7 and 8. On the basis of
86 BPSK coded 4800 100
QPSK coded 9600 99.2
PLC noise classification reported in [46,48], the measured noise can be
BPSK 9600 98.2 considered as colored background noise. In fact, the power spectral
QPSK 19,200 97.1 density, decreases with frequency as expected for colored noise,
110 BPSK coded 4800 100 maintaining a selective behavior due to the coupling system. A max-
QPSK coded 9600 99.7 imum level of 94 dBµV was obtained at about 80 kHz; low average noise
BPSK 9600 98.7 was observed at the higher frequency range, instead. For an efficient
QPSK 19,200 98.1
data transmission, signal level should be almost 10 dB above noise
150 BPSK coded 4800 100 level. Thus at higher center frequencies, signal amplitude reduction is
QPSK coded 9600 100 compensated by a lower noise.
BPSK 9600 100
QPSK 19,200 99.8
8PSK 28,800 98.4 3.3. Transmission tests
200 BPSK coded 4800 100
QPSK coded 9600 100 Different transmission tests were performed by varying center fre-
BPSK 9600 100 quency, modulation technique and data rate of nPSK modems. Tests
QPSK 19,200 100 were performed also activating the error correction mode. In this case,
8PSK 28,800 99.0
two coded bits are transmitted for each information bit (code rate ½),
halving the communication bit rate, but increasing the communication
transformers and MV derivations. The effects of these components were robustness through error correction. To overcome the noise, the max-
analyzed in details in previous papers [36,39]. imum signal amplitude was used (compatibly with the interface card
amplifier dynamic), i.e. 6 VPP, assuring that on MV lines the signal
amplitude was less than 134 dBµV. The success rate was measured over
3.2. Noise measurement 1000 transmitted packets. In Table 1 success rate values (higher than
94%) are reported at different center frequencies and for different
Noise measurement in PLC systems is an issue analyzed in different modulations. The worst performance was found around 78 kHz, caused
research studies and standards [46–48]. Based on time or frequency by a higher measured noise. The coupling system performances get
characteristics, the measured noise in the NB-PLC frequency range can better with the center frequency increase, obtaining, for each modula-
be classified as colored background, narrowband, impulsive noise. tion technique, higher success rates. In the case of 8PSK, which corre-
Noise on MV and LV networks is frequency selective and strongly de- sponds to the maximum available bit rate of 28.8 kbit/s, a success rate
pendent from the site. On the other hand, for each site the noise level is higher than 94% was obtained only at center frequencies of 150 and
almost constant, apart from for occasional appearances of narrowband 200 kHz. During transmission tests, the received signal spectrum was
interferences [47]. These observations underline the importance of also measured. Figs. 9 and 10 show the received QPSK modulated signal
noise measurements in each site. In the experimental campaign herein spectra measured at center frequencies of 86 and 150 kHz, respectively.
presented, the noise measurements at MV coupler first installation and Comparing these two spectra it can be concluded that a center fre-
the comparison between noise and signal level are aimed at quency increase corresponds to an amplitude decrease, confirming the

Fig. 9. QPSK transmission test; center frequency of 86 kHz and 19.2 kbit/s.

485
G. Artale et al. Measurement 129 (2018) 479–488

Fig. 10. QPSK transmission test; center frequency of 150 kHz 19.2 kbit/s.

Fig. 11. MV-LV PLC-based smart metering architecture. Experimental set-up scheme. Red and green lines indicate MV and LV PLC channels, respectively

Table 2 Table 3
Data of LV network under test. Smart metering architecture experimental tests.
9 to A A to F F to G Smart meter position

Line length [m] 274 701 947 Substation A F G

Phase cable core section [mm2] 150 95 95 9
Neutral cable core section [mm2] 95 50 50
Correctly received packets 980 972 960 896
Max RTT [ms] 288 350 391 430
Min RTT [ms] 258 258 258 258
results of frequency response tests. On the other hand, a flatter response
Mean RTT [ms] 271 271 270 271
is obtained at 150 kHz, which corresponds to a 99.8% of success rate RTT 95% [ms] 282 284 284 284
with QPSK modulation. Finally, an uneven frequency trend is observed Standard deviation [ms] 7 7 8 11
at 86 kHz, which causes a lower success rate (97.1%).

486
G. Artale et al.
Fig. 12. Frequency histogram of measured RTT from substation 6 to junction
box A.
Fig. 13. Frequency histogram of measured RTT from substation 6 to junction
box G.
4. PLC-based smart metering architecture for medium/low voltage
grid. On field set-up and experimental results
To test the whole MV/LV PLC smart metering architecture, the ex-
perimental on-field MV/LV PLC smart metering set-up (see Fig. 11)
consisted in:
• the configurable PLC modems with a MV VDS PLC innovative cou-
plers installed “Gen. Di Vita” and “4 Vanelle” substations (number 6
and 9, in Fig. 3 – see previous section);
• an AMR concentrator emulator in “4 Vanelle” secondary substation,
communicating via SPI with the MV PLC modem;
• a smart meter, communicating via LV PLC with the AMR con-
centrator emulator.
As already mentioned, the LV line used for the tests was the line L1
(green route in Fig. 3). Different tests were carried out by connecting
the smart meter to various LV junction boxes located at different dis-
tances from “4 Vanelle” substation. For the three tests herein presented,
the LV junction boxes positions are indicated in Fig. 3 with the capital
letters A, F and G, respectively. Distances from “4 Vanelle” substation
and core sections of LV cable lines are reported in Table 2.
The experimental set-up for the MV PLC channel was the same
487
Measurement 129 (2018) 479–488
described in previous section. In 4 Vanelle, the prototypal box with the
PLC modem and the tunable interface card was connected to the AMR
concentrator emulator via SPI. Both the AMR concentrator and the
smart meter used for the tests were based on two demo boards STEVAL-
IPP001V2 (by STMicroelectronics) [49]. They were customized by au-
thors in the function of smart meter and AMR concentrator emulator,
developing both a proper firmware and a dedicated adjunctive hard-
ware. More in detail, the smart meter heart is the STM32 micro-
controller embedded in the STEVAL-IPP001V2. It is connected to an
embedded ST7580 PLC modem and to an external meter board
STEVAL-IPE010V1. This last has a STPMC1 energy calculator and it can
be configured for mono-phase or three phase network applications,
adding STEVAL-IPE014V1 modules, one for each phase. These modules
embed voltage and current transducers for direct LV connection. The
measured data received by the AMR concentrator via LV PLC are then
sent via SPI to the PLC modem board inside the prototypal box. A
proper firmware was developed for each prototype to handle the dif-
ferent measurement, data collection and communication tasks.
The smart metering performances were experimentally verified by
performing ping tests from “Gen. Di Vita” (substation 6) to the smart
meter connected to L1 line. More in detail, the ping request starting
from “Gen. Di Vita” substation were transmitted via MV PLC to “4
Vanelle” substation and then via LV PLC to the smart meter. This last
then sent a ping acknowledgment response, which followed the back-
ward route, i.e. via LV PLC reached the AMR concentrator in “4 Vanelle”
and via MV PLC “Gen. Di Vita” substation. In each ping test, the round
trip time (RTT) was measured as the time from the instant of sending
the message to the instant of receiving its acknowledgment from the
remote smart meter. Different PLC center frequencies were used for MV
and LV transmission in order to avoid interferences between MV and LV
PLC transmissions, due to PLC signals crossing the power transformers
[37]. In detail, the center frequency of 86 kHz was chosen for LV
transmissions with FSK modulation, while 110, 150 and 200 kHz were
chosen for MV transmissions with BPSK modulation. The smart meter
was subsequently connected to the selected junction boxes. A summary
of the obtained results in the case of 86 kHz @LV PLC and 110 kHz @
MV PLC is reported in Table 3. As can be observed, mean and 95 per-
centile values of measured RTTs are not significantly affected by LV line
length. On the other hand, the number of packets correctly received
decreases with the line length. This is due to the signal attenuation
compared to the LV line noise in the PLC frequency band. The fre-
quency histograms of the ping tests performed in locations A and G are
shown in Figs. 12 and 13, respectively. They confirm the low dispersion
of the RTT measurements.
5. Conclusions
In this paper, the feasibility was investigated of an innovative and
low cost smart metering architecture for MV and LV distribution power
grids in the smart cities framework. The proposed solution is low cost
because it is based on the use of PLC both at LV and MV level, thus
significantly reducing installation and service provider costs. In fact, an
innovative solution was proposed to couple the PLC signal to the MV
power line, which is based on the use of VDS capacitive divider as MV
PLC couplers. In this way, the smart meter data, collected by AMR
concentrator at LV level, can be easily sent via PLC to MV con-
centrators, thus avoiding wireless solutions with the intrinsic installa-
tion and service provider costs. The MV concentrator gathered data can
be used to implement different smart grids applications, such as to act
as aggregator of a selected micro-network in future energy market
trading. MV aggregated data are also sent to the DSO control center,
allowing network monitoring and management.
An experimental set-up was developed in the real distribution net-
work of the Island of Favignana, to test the proposed architecture. The
on-field experimental campaign was aimed at verifying the MV PLC
coupling solution performances in a wide frequency range from
G. Artale et al.
CENELEC A band up to 200 kHz. The whole smart metering architecture
was also verified; the measured success rate and RTT obtained, in real
environment, confirm the feasibility of the proposed solution.
Moreover, the low cost of the proposed solutions guarantee easily
scalability favoring its vast employment in modern smart cities.
Acknowledgment
The authors would like to thank the local electrical energy dis-
tribution operator SEA S.p.a. (Società Elettrica di Favignana) for the
assistance during the on-field tests.
This research was supported by STMicroelectronics Srl (STM),
Agrate Brianza, MB, Italy (research contract “smart grid”).
References
[1] R. Morello, S.C. Mukhopadhyay, Z. Liu, D. Slomovitz, S.R. Samantaray, Advances
on sensing technologies for smart cities and power grids: a review, IEEE Sens. J. 17
(23) (2017) 7596–7610.
[2] M.L. Tuballa, M.L. Abundo, A review of the development of Smart Grid technolo-
gies, Renew. Sustain. Energy Rev. 59 (2016) 710–725.
[3] P.T. Manditereza, R. Bansal, Renewable distributed generation: the hidden chal-
lenges – A review from the protection perspective, Renew. Sustain. Energy Rev. 58
(2016) 1457–1465.
[4] P.-Y. Kong, A distributed management scheme for energy storage in a smart grid
with communication impairments, IEEE Trans. Ind. Inf. 14 (4) (2018) 1392–1402.
[5] N. Saxena, I. Hussain, B. Singh, A.L. Vyas, Implementation of a grid-integrated PV-
battery system for residential and electrical vehicle applications, IEEE Trans. Ind.
Electron. 65 (8) (2018) 6592–6601.
[6] M. Farhoodnea, A. Mohamed, H. Shareef, H. Zayandehroodi, Power quality impacts
of high-penetration electric vehicle stations and renewable energy-based generators
on power distribution systems, Measurement 46 (8) (2013) 2423–2434.
[7] K.G. Di Santo, S.G. Di Santo, R.M. Monaro, M.A. Saidel, Active demand side man-
agement for households in smart grids using optimization and artificial intelligence,
Measurement 115 (2018) 152–161.
[8] D. Croce, F. Giuliano, I. Tinnirello, A. Galatioto, M. Bonomolo, M. Beccali, G. Zizzo,
Overgrid: a fully distributed demand response architecture based on overlay net-
works, IEEE Trans. Autom. Sci. Eng. 14 (2) (2017) 471–481.
[9] G. Zizzo, E. Riva Sanseverino, M.G. Ippolito, M.L. Di Silvestre, P. Gallo, A technical
approach to P2P energy transactions in microgrids, IEEE Trans. Indus. Informatics
(2018), https://doi.org/10.1109/TII.2018.2806357.
[10] R. Morello, C. De Capua, G. Fulco, S.C. Mukhopadhyay, A smart power meter to
monitor energy flow in smart grids: the role of advanced sensing and IoT in the
electric grid of the future, IEEE Sens. J. 17 (23) (2017) 7828–7837.
[11] M. Pasetti, S. Rinaldi, D. Manerba, A virtual power plant architecture for the de-
mand-side management of smart prosumers, Appl. Sci. (Switzerland) 8 (2018) (3).
[12] N. Shaukat, et al., A survey on consumers empowerment, communication tech-
nologies, and renewable generation penetration within Smart Grid, Renew. Sustain.
Energy Rev. 81 (2018) 1453–1475.
[13] E. Kabalci, Y. Kabalci, A measurement and power line communication system de-
sign for renewable smart grids, Meas. Sci. Rev. 13 (5) (2013) 248–252.
[14] A. Cataliotti, V. Cosentino, D. Di Cara, P. Russotto, E. Telaretti, G. Tinè, An in-
novative measurement approach for load flow analysis in MV Smart Grids, IEEE
Trans. Smart Grid 7 (2) (2016) 889–896.
[15] S. Rinaldi, D. Della Giustina, P. Ferrari, A. Flammini, E. Sisinni, Time synchroni-
zation over heterogeneous network for smart grid application: design and char-
acterization of a real case, Ad Hoc Netw. 50 (2016) 41–57.
[16] Q. Sun, H. Li, Z. Ma, C. Wang, J. Campillo, Q. Zhang, F. Wallin, J. Guo, A com-
prehensive review of smart energy meters in intelligent energy networks, IEEE
Internet Things J. 3 (4) (2016) 464–479.
[17] S. Jamali, A. Bahmanyar, E. Bompard, Fault location method for distribution net-
works using smart meters, Measurement 102 (2017) 150–157.
[18] G. Del Prete, D. Gallo, C. Landi, M. Luiso, Real-time smart meters network for en-
ergy management, ACTA IMEKO 2 (1) (2013) 40–48 ISSN 2221-870X.
[19] A. Cataliotti, V. Cosentino, D. Di Cara, G. Tinè, LV measurement device placement
for load flow analysis in MV Smart Grids, IEEE Trans. Instrum. Meas. 65 (5) (2016)
999–1006.
[20] A. Cataliotti, V. Cosentino, N. Nguyen, P. Russotto, D. Di Cara, G. Tinè, Hybrid
passive and communications-based methods for islanding detection in medium and
low voltage smart grids, Int. Conf. Power Eng., Energy Electric. Drives (2013)
1563–1567.
[21] G. Artale, A. Cataliotti, V. Cosentino, D. Di Cara, S. Nuccio, G. Tine, Arc fault de-
tection method based on CZT low-frequency harmonic current analysis, IEEE Trans.
Instrum. Meas. 66 (5) (2017) 888–896.
[22] G. Crotti, D. Gallo, D. Giordano, C. Landi, M. Luiso, M. Modarres, M. Zucca,
Frequency compliance of MV voltage sensors for smart grid application, IEEE
Sensors J. 17 (23) (2017) 7621–7629 7976280.
488
Measurement 129 (2018) 479–488
[23] L. Morales-Velazquez, R.D.J. Romero-Troncoso, G. Herrera-Ruiz, D. Morinigo-
Sotelo, R.A. Osornio-Rios, Smart sensor network for power quality monitoring in
electrical installations, Measurement 103 (2017) 133–142.
[24] V. Pop, H.J. Bergveld, P.H.L. Notten, J.H.G. Op het Veld, P.P.L. Regtien, Accuracy
analysis of the state-of-charge and remaining run-time determination for lithium-
ion batteries, Measurement 42 (8) (2009) 1131–1138.
[25] G. Artale, A. Cataliotti, V. Cosentino, D. Di Cara, S. Guaiana, S. Nuccio,
N. Panzavecchia, G. Tinè, Smart interface devices for distributed generation in
smart grids: the case of islanding, IEEE Sens. J. 17 (23) (2017) 7803–7811.
[26] Y. Kabalci, A survey on smart metering and smart grid communication, Renew.
Sustain. Energy Rev. 57 (2016) 302–318.
[27] D.D. Giustina, S. Rinaldi, Hybrid communication network for the smart grid: vali-
dation of a field test experience, IEEE Trans. Power Del. 30 (6) (2015) 2492–2500.
[28] E. Kabalci, Y. Kabalci, I. Develi, Modelling and analysis of a power line commu-
nication system with QPSK modem for renewable smart grids, Int. J. Electr. Power
Energy Syst. 34 (1) (2012) 19–28.
[29] P. Kiedrowski, Toward more efficient and more secure last mile smart metering and
smart lighting communication systems with the use of PLC/RF hybrid technology,
Int. J. Distrib. Sens. Netw. (2015).
[30] S. Souissi, O. Ben Rhouma, C. Rebai, Statistical modeling of mains zero crossing
variation in powerline communication, Measurement: J. Int. Measurement
Confederation 90 (2016) 158–167.
[31] M.C. Bali, C. Rebai, Improved maximum likelihood S-FSK receiver for PLC modem
in AMR, J. Electric. Comput. Eng. (2012) 452402.
[32] B. Masood, S. Baig, Standardization and deployment scenario of next generation
NB-PLC technologies, Renew. Sustain. Energy Rev. 65 (2016) 1033–1047.
[33] A. Cataliotti, V. Cosentino, D. Di Cara, S. Guaiana, N. Panzavecchia, G. Tine, A new
solution for low-voltage distributed generation interface protection system, IEEE
Trans. Instrum. Meas. 64 (8) (2015) 2086–2095.
[34] T.A. Papadopoulos, Ch.G. Kaloudas, A.I. Chrysochos, G.K. Papagiannis, Application
of narrowband power-line communication in medium-voltage smart distribution
grids, IEEE Trans. Power Del. 28 (2) (2013) 981–988.
[35] A.I. Chrysochos, T.A. Papadopoulos, A. ElSamadouny, G.K. Papagiannis, N. Al-
Dhahir, Optimized MIMO-OFDM design for narrowband-PLC applications in
medium-voltage smart distribution grids, Electr. Power Syst. Res. 140 (2016)
253–262.
[36] A. Cataliotti, D. Di Cara, G. Tinè, Model of line to shield power line communication
system on a medium voltage network, in: Proc. IEEE I2MTC, Austin, TX , 3–6 May,
2010, pp. 1459–1462.
[37] A. Cataliotti, V. Cosentino, D. Di Cara, G. Tine, Oil-filled MV/LV power-transformer
behavior in narrow-band power-line communication systems, IEEE Trans. Instrum.
Meas. 61 (10) (2012) 2642–2652.
[38] T.A. Papadopoulos, A.I. Chrysochos, A. ElSamadouny, N. Al-Dhahir,
G.K. Papagiannis, MIMO-OFDM narrowband-PLC in distribution systems: impact of
power transformers on achievable data rates, Electr. Power Syst. Res. 151 (2017)
251–265.
[39] A. Cataliotti, V. Cosentino, D. Di Cara, G. Tinè, Measurement issues for the char-
acterization of medium voltage grids communications, IEEE Trans. Instrum. Meas.
62 (8) (2013) 2185–2196.
[40] R. Fiorelli, A. Cataliotti, D. Di Cara, G. Tinè, Coupling circuit for power line com-
munications, US Patent 8896393 B2, Nov. 25, 2014.
[41] G. Artale, A. Cataliotti, V. Cosentino, D. Di Cara, S. Guaiana, R. Fiorelli,
N. Panzavecchia, G. Tine, A new low cost power line communication solution for
smart grid monitoring and management, IEEE Instrum. Meas. Mag. 21 (2) (2018)
29–33.
[42] G. Artale, A. Cataliotti, V. Cosentino, D. Di Cara, G. Tinè, Development of a coupling
system for medium voltage power line communication in the CENELEC a frequency
band, in: Proc. IEEE AMPS 2016, Aachen, Germany, Sept. 2016, pp.103–107.
[43] G. Artale, A. Cataliotti, V. Cosentino, D. Di Cara, R. Fiorelli, S. Guaiana, G. Tinè, A
new low cost coupling system for power line communication on medium voltage
smart grids, IEEE Trans. Smart Grid 9 (4) (July 2018) 3321–3329.
[44] Live working – Voltage detectors – Part 1: capacitive type to be used for voltages
exceeding 1 kV a.c., IEC 61243-1, 2009.
[45] Signalling on low-voltage electrical installations in the frequency range 3 to 148.
5 kHz—Part 1: general requirements, frequency bands and electromagnetic dis-
turbances, CENELEC EN 50065-1, 1995.
[46] IEEE 1901.2-2015, IEEE Standard for Low-frequency (Less Than 500 kHz)
Narrowband Power Line Communications for Smart Grid
Applications—Amendment 1, IEEE Std., 2015, (2017).
[47] I.H. Kim, A. Dabak, D. Rieken, G. Gregg, Evaluating the Low-frequency Power-line
Communications Channel in Rural North America, Texas Instruments, Texas, USA,
2012.
[48] M. Zimmermann, K. Dostert, Analysis and modeling of impulsive noise in broad-
band powerline communications, IEEE Trans. Electromagn. Compat. 44 (1) (Feb
2002) 249–258.
[49] A. Cataliotti, et al., A prototypal architecture of a IEEE 21451 network for smart
grid applications based on power line communications, IEEE Sens. J. 15 (5) (May
2015) 2460–2467.
[50] A. Cataliotti et al., A simplified approach for load flow analysis in MV smart grids
based on LV power measurements, in: 2017 IEEE International Instrumentation and
Measurement Technology Conference (I2MTC), Turin, 2017, pp. 1–6.