Vous êtes sur la page 1sur 6

Advanced Metering Infraestructure Case Study in

Federal University of Santa Maria


William Dotto Vizzotto
Filipe Gabriel Carloto
Lucas Maziero
Flávio Garlet Reck
Fernando Garcia Vedoin
Carlos Henrique Barriquello
Daniel Pinheiro Bernardon
Tiago Bandeira Marchesan
Federal University of Santa Maria
Santa Maria, Brazil
vizzotto@foxiot.com.br

Abstract—The increase in electrical energy consumption is energy flows and significantly increase the complexity of the
causing the grid to be more complex in the last years. Utilities electrical systems, as shown in Figure 1 [1].
need to improve the quality of the service, as the efficiency of
delivering energy, while electric vehicles and distributed energy
resources are becoming a reality in the market. The advance of
electronic communication technology, the processing capacity of
computers, cloud computing, among others, enables many fea-
tures for Smart Grids. The Advanced Metering Infraestructure is
one of the tools enabled by Smart Grids, which is an information
system for collecting, storing and analyzing power data. The
structure, communication technologies, and benefits of this tool
are discussed in the paper. Then, the case study of the Federal
University of Santa Maria Advanced Metering Infrastructure is
shown, followed by its results.
Index Terms—AMI, Distribution system, LPWAN, LoRaWAN,
RF, Smart Meter.

Fig. 1. Present and future of the electrical grid [1].


I. I NTRODUCTION
The Advanced Metering Infraestructure (AMI) is an infor-
The global climate change is causing many countries pro- mation system for collecting, storing and analyzing power
mote various policies for the Smart Grids. The recent growth, data and consists of three parts, namely, smart meter, com-
together with the characteristic of distributed energy resources munication system and MDMS (Meter Database Management
(DER), indicates that the diffusion of these technologies System). The Smart Meter is the device that senses electrical
present a high disruptive potential, capable of profoundly parameters and energy, and is able to send this data through
transforming the electric systems that today are predominantly the communication network to the MDMS, which receives,
operated with larger generators and centrally managed. In analyzes and manages data, and further integrates with other
fact, the World Energy Council (2017) points out that in application systems [2].
the period between 2017 and 2025 several countries will go AMI can also be applied to home energy management, help
through a transition from predominantly centralized power reducing the cost and manpower for manual meter reading,
systems to hybrid systems. In this context, the transition from and providing power information for users or utilities to plan
a centralized to a more distributed model should change the energy management, achieving the purpose of energy saving,
and carbon reduction [3].
The authors would like to thank the technical and financial support of The energy management feature is very important when
CPFL Energia Power Utility to the project Soluções Inovadoras de Eficiência considering public universities from Brazil. UFSM has around
Energética e Minigeração em Instituição Pública Federal de Ensino Superior: 30,000 students and 5,000 workers, that demand a huge
Uma abordagem na UFSM (PD/ANEEL). This study was also financed in
part by the Coordenação de Aperfeiçoamento de Pessoal de Nivel Superior amount of electricity over the day to lighting, air-conditioning,
Brazil (CAPES) Finance Code 001.. electronic devices, and laboratories. Besides that, The Federal

978-1-5386-8218-0/19/$31.00 ©2019 IEEE


University of Santa Maria (UFSM) supports many projects outages and power quality, enabling the service provider to
related to sustainability and energy efficiency. So in order to rapidly address grid deficiencies. The AMIs bidirectional com-
evaluate the results of its actions, an energy monitoring system munications infrastructure also supports grid automation at the
is needed. station and circuit level. The vast amount of new data flowing
In order to monitor energy consumption over the buildings, from AMI allows improved management of utility assets as
an AMI was implemented. The AMI consists of 60 smart well as better planning of asset maintenance, additions, and
meters installed in the main buildings, a LoRaWAN network replacements. The resulting more efficient and reliable grid is
to cover the university area, and a local MDMS (time-series one of AMIs many benefits [4].
database, broker and dashboard).
The paper is divided as follows: Section II shows the main B. Communication Technologies
concepts of AMI (structure, communication technologies and An important feature of the Smart Grid is the interaction
applications), Section III explains the case study of UFSM and among its entities using bidirectional communication. In the
in Section IV the results are presented. Smart Grid, AMI is the distribution level building block
comprising a network of smart meters. Smart meters collect
II. A DVANCED M ETERING I NFRASTRUCTURE
and report the consumption data to a Service Provider (SP)
AMI is not a single technology implementation, but rather periodically (less than 1 hour) rather than monthly total con-
a fully configured infrastructure that must be integrated into sumption recorded in conventional meter, as well as allows the
existing and new utility processes and applications. consumers to get involved in electricity trade by selling home-
A. AMI Structure generated, unused electricity to the grid [5]. AMI also caters to
the SP, the control and monitoring with outage management,
The AMI includes home network systems, smart meters, demand response, disaster prevention, and disaster recovery.
communication networks from the meters to local data concen- Consequently, the communication in AMI is necessarily bidi-
trators, back-haul communications networks to corporate data rectional. The domain of communication of a smart meter is
centers, meter data management systems (MDMS) and, finally, segregated into two groups - Home Area Network (HAN) and
data integration into existing and new software application Neighborhood Area Network (NAN). Smart home appliances
platforms. Additionally, AMI provides a very intelligent step communicate with the smart meters in HAN whereas meter-
toward modernizing the entire power system. Figure 2 graph- to-meter communication is termed as NAN. The networking
ically describes the AMI technologies and how they interface: of the smart meters is proposed by a wireless network [6],
[7], [8]. Although, interference is a major issue in wireless
communication [9].
Power Line Communication (PLC) or Radio Frequency
(RF) are the most common techniques in AMI field. PLC
is a communication method that uses power line to carry
data, but the communication quality may be affected by
power line noise or the impedance which does not match at
load change, and the long transmission distance may cause
signal attenuation. PLC viability also depends on the quantity
of consumers by distribution transformer. RF technology is
another LAN communication option, each country has its
own RF channel to choose RF specification. The general RF
Fig. 2. AMI structure overview [4]. technology is used in open areas and need meter concentration,
and the data transmission have a high success rate. The big
At the consumer level, smart meters communicate consump- challenge of LAN communication technology is complexity
tion data to both the user and the service provider. Smart me- environment and communication interference at present, it
ters communicate with in-home displays to make consumers may be the reason that communication technology cant be
more aware of their energy usage. Going further, electric applied in all environments. The gateway to MDMS or meter
pricing information supplied by the service provider enables to AMI MDMS WAN communication usually is wired or wire-
load control devices like smart thermostats to modulate electric less. A wireless solution often apply Cellular communication
demand, based on pre-established consumer price preferences. technology , and a wired solution is the optical fiber. It can be
The consumer portals process the AMI data in ways that seen that the AMI can be composed of different heterogeneous
enable more intelligent energy consumption decisions, even transmission technologies, to use in different communication
providing interactive services like prepayment. environments such as cities, suburbs, mountains, islands and
The service provider (utility) employs existing, enhanced other environments [2]. Figure 3 shows a brief comparison of
or new back-office systems that collect and analyze AMI data bandwidth and range of most common wireless technologies.
to help optimize operations, economics and consumer service. It is important to address the communication requirements
For example, AMI provides immediate feedback on consumer in order to choose the one that best fits. [11] and [12] present
participation, the utility companies will more likely provide
electricity at a lower rate for their customers [13].
2) Distributed Network Management: The distributed net-
work helps the utilities in providing useful information to more
accurately size new transformers and circuits to match the
peak load conditions. Additionally, it also provides information
about factors stimulating peak consumption [13].
3) Variable Pricing Structure: AMI enables the price struc-
ture to better align as per customer usage, so those consuming
energies at peak time will be charged more. Also, monitoring
the energy usage on a daily or hourly basics creates an
opportunity for services like variable pricing to encourage off-
peak usage without the reduced rate [14].
Fig. 3. Comparison of bandwith and range of most common wireless 4) Improved Data Quality: Automatic remote data col-
technologies [10].
lection sleeks the processing for billing, asset management,
and outage management. Machine to machine data transfers
the relation of communication requirements and different reduces the number of misreads. Overall this meter reading
smart grid applications, that are summarized in Table I. reduces billing error issues and customer disputes [15].
5) Shorter Billing Cycle: Changing the traditional meter
TABLE I reading with AMI cuts down the billing process by reducing
C OMMUNICATION REQUIREMENTS FOR S MART G RID APPLICATIONS . the time and the number of steps between consumer usage
bills and bill distributions [16], [17].
Data Max. Traffic
Application QoS
rate latency type
AMI
III. C ASE STUDY OF UFSM’ S AMI
500kbps 50ms-1min periodic moderate
Metering This section presents the AMI implemented in UFSM. First,
Demand 14- periodic/
Response 100kbps
2s-5min
random
strict the network infrastructure is described, as well the reasons for
Customer choosing the specific communication technology. The Smart
13 kbps 2s periodic moderate
feedback Meters functionality are presented, and finally, the aimed
Customer benefits of the AMI for the university.
32 kbps N/A random moderate
bill savings
Data 32 kbps N/A random moderate
A. Network Infraestructure
Voice 8 kbps 200ms random moderate
File transfer 8 kbps 200ms random moderate The choice of the network infrastructure is based in the
Substation communication requirements. Latency, data rate and QoS were
800 kbps 40ms periodic strict
automation
considered in order to attend as many as possible applications
Interference due to home appliances (like microwave oven, of Smart Grids, and not just AMI, as the network is also going
cordless phone, and garage door), proprietary use, etc., along to be used for other researches. Furthermore, the network
with Rayleigh fading, co-channel interference, and attenuation should be able to cover a large area and points that are distant
due to obstacles (like buildings, trees, etc.) are the common from each other, operates on public unlicensed frequency.
issues on public unlicensed frequency band. However, reliable The Low Power Wide Area Network (LPWAN) and Cellular
wireless communication is obligatory from a commercial point technologies fit the needs, however, cellular needs recurrent
of view as well as for robustness of Smart Grid. To nullify payment (generally, monthly), as well as SigFox, for example,
the effect of attenuation and some interference (like co- while LoRaWAN has just the initial cost of infrastructure.
channel interference, harmonics from home appliances, etc.), So, the university engineering team decided by LoRaWAN
a frequency band with more channels and channel spacing is technology. Another advantage of LoRaWAN is that it is
advantageous. If a specific channel is attenuated or interfered an open protocol. Table II summarizes the characteristics of
by the same frequency, other channels can carry the data. LoRaWAN [18].

C. Applications TABLE II
L O R AWAN LIMITATIONS .
The main benefits are explained briefly. Some of them have
already been implemented in many countries, and others still Technology
Data Max. Traffic
QoS Distance
need development or economic viability. rate latency type
40ms to
1) Managing Supply and Demand: The demand manage- 0.3- periodic/ 2.5km urban
LoRaWAN several moderate
27kbps random 15km suburban
ment saves result from proper load selection and optimized seconds
use of electricity. It also allows people the opportunity to shift
their own demand for electricity during peak hours and reduce The installed gateways are RHF2S008, from Rising HF, it
their energy consumption overall. Hence through customer is an IOT gateway based on LoRaWAN and target to LPWAN
network. It is an IEEE 802.3 af/at compatibility PD, which • Meter: Three-phase - 4 wires;
could be powered by PoE. Both Ethernet and LTE-4G are • Wireless communication: LoRaWAN stack v1.0.2;
supported to connect to the cloud server. With an integrated • Wireless frequency: 915MHz - Australian Standard
GPS module, the GW could support LoRaWAN Class B • Model: SME;
protocol with the synchronous clock from GPS PPS signal. • Nominal Voltage (Vn): 90V to 240V;
This device integrate an high performance CPU ARM • Nominal Current (In): 50 to 1500A (depends on CT);
Cortex-A53 core, one baseband processor SX1301, that it • Frequency: 50 or 60Hz;
could support 8 multi-SF channel (SF12 to SF7), 1 single-SF • Active Energy: Ke = Kh = 1,00 Wh/pulse;
channel and 1 GFSK channel. Output power could achieve to Furthermore, the SME has a digital output, that can be
27dBm max. Sensitivity is as low as -141dBm@300bps. With configured in the MDMS software. It allows the user to control
specified payload length and transmit period, one gateway loads and trigger alarms using this digital output. There have
could support 10k nodes. It is a smart device but with been installed 60 smart meters in the main buildings of UFSM.
high reliability, that it could work outdoor or in a complex Figure 6 shows where the smart meters have been installed.
environment [19].
The gateways were installed in order to cover all the
LoRaWAN smart meters, as can be seen in Figure 4. They
are positioned in the top of two of the highest buildings in the
university. They are supplied by no-breaks and connected to
the institutional internet network through Ethernet cables.

Fig. 6. Buildings where there are smart meters installed.

They are configured to send the data periodically, every 15


seconds. The data packet is composed of individual phases and
Fig. 4. Gateways installed in UFSM. total voltages, currents, active, reactive and apparent powers,
energy and power factors.
B. Smart Meters C. Meter Data Management System
The smart meters are model SME, from Fox IoT company, In order to receive, manage, store, analyze, and provide
similar to the Figure 5. access to the data sent by the smart meters, it is necessary
to have the MDMS. In this case, the MDMS is composed
by the LoRaWAN network server, MQQT broker, database,
application server, and user interface. The LoRaWAN network
server is responsible for management of the connections and
gateways, and package receiving. The MQQT broker does
the bridge between the data received by the network server
and application server. The application server is the core of
MDMS, it is responsible by getting the data and storing in the
database, as well as executing the commands of the user. The
user interface is the dashboard of the system, composed by
graphs, gauges, buttons, and others.

D. Benefits
Fig. 5. Fox IoT SME Smart Meter [20]. There are many benefits that could be achieved by AMI
implementation. The UFSM’s engineering team designed the
Their main characteristics are [20]: system in order to have better knowledge about the energy
consumption, power quality, and enough information in order TABLE III
to guide energy efficiency activities in the university. D ELIVERY RATE OF THE 30 ANALYZED SMART METERS .
The energy consumption is about how much each building Delivery
Meter id Building
spend in electricity. Power quality considers information about rate (%)
voltage levels, power factor, load balancing and the load sm01 Dança 85,1
sm02 Casa Estudante Indı́gena 98,4
level of each distribution transformer. The information for sm06 Sistema Irriga 92,9
energy efficiency includes the data from electricity usage sm07 Prédio 74B 72,7
and consumer habits, relation between energy consumed and sm11 Centro de Educação 80,5
sm17 CTISM I 97,9
how many students have classes in each building. The data sm18 Música 100
extracted from the smart meters allows the possibility of a sm19 Geração Distribuı́da 98,8
wide range of new researches, like demand response in public sm20 Destilaria 95,3
buildings, real-time distribution loss evaluation, automatic load sm23 Matemática 96,3
sm24 Prédio 40 78,9
control, among others. sm26 Terapia Ocupacional 96,1
sm29 Casa do Estudante - 34 95,7
IV. R ESULTS sm31 Agittec 97,8
sm34 Restaurante Universitário 98,4
The UFSM’s AMI consists of 2 LoRaWAN gateways, 60
sm37 Prédio 74C 81,4
smart meters, and the MDMS. However, for this case study, sm36 Almoxarifado 54,4
30 smart meters will be evaluated and tested, because not all sm38 Casa do Estudante - 35 56,5
of them are installed and completely working already. The sm40 Casa do Estudante - 33 97,7
sm43 CFED Piscinas 98,8
results will show the functionality of the network (Availability sm45 Centro de Tecnologia 80,3
and latency) and some benefits that already are possible in sm46 CFED 100
concerning with electrical parameters (Energy quality, load sm47 Laboratório Virologia 97,9
balancing, and energy efficiency). The analysis was performed sm48 Prédio 26 86,9
sm50 Anfiteatro I 100
between March 15, 2019, and March 23, 2019. sm53 Centro de Educação 2 92,8
sm54 Anfiteatro II 72
A. Availability sm55 Reitoria II 89,4
sm57 Biblioteca Central 77,1
The availability of the system is measured considering
sm58 Odontologia 88,1
the relation between the number of packets received by the
network server and number of packets that should be sent by
the smart meter. The results are presented in Table III.
Through the 30 smart meter analyzed, the average delivery
rate is 88.6%. 3 smart meters had 100% delivery rate, 36.67%
between 95% and 99.99%. On the other hand, 13.33% deliv-
ered less than 75% of the packets.

B. Latency
The latency is achieved by the network server, that manages
the communication network. It considers the air time of the
package, which depends on LoRaWAN network parameters,
like spreading factor, bandwidth, packet size, and others. The
currently set of parameters results in 369ms. This results shows
Fig. 7. SM46 CFED voltages.
that LoRa AMI allows many Smart Grid features, but is not
suitable for distribution automation, for example.
It is verified that the deviation between the phases and the
C. Energy Quality
average can be as high as 60.14%. This implies in the need of
In order to evaluate the energy quality supplied to the balancing the load over the phases, in order to have a better
buildings, it was verified the voltage levels. Figure 7 shows energy quality and less losses.
an example of CFED building voltages.
It is possible to see that maximum voltages measured are D. Energy efficiency
231.61V, 234.01V, and 232.64V of phases A, B, and C, The energy effiency is a tool used to do the same activities
respectively. The nominal voltage is 220V, while the limits using less energy, as well as paying less. The AMI of UFSM
imposed by the Brazilian Regulatory Agency are 231V and allows to rank the top 10 energy consumption buildings and
202V, for the maximum and the mininum, respectively. In power. Figures 8 and 9 shows these two ranks.
order to evaluate the load balance, the maximum difference These 10 buildings are the ones where energy efficiency
between the average powers. Table IV presents the top 5 activities should be implemented first, as they have the greatest
unbalanced loads. consumers among the analyzed smart meters.
TABLE IV R EFERENCES
T OP 5 UNBALANCED LOADS .
[1] EPE, “Distribution energy resources: Impacts on energy planning stud-
Meter Avg A Avg B Avg C Avg total ies,” vol. 1, no. 1, july 2018.
Building dev(%) [2] T. Sangsuwan, S. Thepphaeng, and C. Pirak, “Experimental performance
id (kW) (kW) (kW) (kW)
Geração analysis of powerline communication technologies in ami systems,” in
sm19 1.48 3 1.14 1.87 60.14% The 20th Asia-Pacific Conference on Communication (APCC2014), Oct
Distribuı́da
sm31 Agittec 1.98 2.51 4.32 2.94 47.11% 2014, pp. 382–386.
Casa [3] S. Chren, B. Rossi, and T. Pitner, “Smart grids deployments within eu
sm02 Estudante 1.81 1 2.65 1.82 45.60% projects: The role of smart meters,” in 2016 Smart Cities Symposium
Indı́gena Prague (SCSP), May 2016, pp. 1–5.
[4] N. E. T. Laboratory, “Advanced metering infraestructure,” NETL Modern
sm55 Reitoria II 15.23 7.42 15.3 12.65 41.34%
Grid Strategy, vol. 1, no. 1, february 2018.
sm20 Destilaria 6.97 3.91 7.05 5.98 34.58%
[5] I. Parvez, A. Sundararajan, and A. I. Sarwat, “Frequency band for han
and nan communication in smart grid,” in 2014 IEEE Symposium on
Computational Intelligence Applications in Smart Grid (CIASG), Dec
2014, pp. 1–5.
[6] C. Bennett and D. Highfill, “Networking ami smart meters,” in 2008
IEEE Energy 2030 Conference, Nov 2008, pp. 1–8.
[7] H. Shabani, M. M. Ahmed, S. Khan, S. A. Hameed, and M. Hadi
Habaebi, “Smart zigbee/ieee 802.15.4 mac for wireless sensor multi-hop
mesh networks,” in 2013 IEEE 7th International Power Engineering and
Optimization Conference (PEOCO), June 2013, pp. 282–287.
[8] C. Prapasawad, K. Pornprasitpol, and W. Pora, “Development of an
automatic meter reading system based on zigbee pro smart energy profile
ieee 802.15.4 standard,” in 2012 IEEE International Conference on
Electron Devices and Solid State Circuit (EDSSC), Dec 2012, pp. 1–
3.
Fig. 8. Top 10 highest energy consumption. [9] W. L. Firestone, “A review of communication interference problems,”
IEEE Transactions on Vehicular Communications, vol. 14, no. 1, pp.
57–66, March 1965.
[10] Addvalue. Addvalue internet of things (iot) , extending global connectiv-
ity for iot. [Online]. Available: https://www.addvaluetech.com/satcom-
iot-via-lora-wan/
[11] and and, “The design of communication infrastructures for smart das
and ami,” in 2010 International Conference on Information and Com-
munication Technology Convergence (ICTC), Nov 2010, pp. 461–462.
[12] M. Islam, M. M. Uddin, M. A. A. Mamun, and M. A. Kader,
“Performance analysis of ami distributed area network using wimax
technology,” in 2014 9th International Forum on Strategic Technology
(IFOST), Oct 2014, pp. 152–155.
[13] S. Nimbargi, S. Mhaisne, S. Nangare, and M. Sinha, “Review on ami
technology for smart meter,” in 2016 IEEE International Conference
on Advances in Electronics, Communication and Computer Technology
(ICAECCT), Dec 2016, pp. 21–27.
Fig. 9. Top 10 highest power.
[14] A. R. C., A. E., R. S. B., and S. K. Vasudevan, “Smart meter
based on real time pricing,” Procedia Technology, vol. 21, pp. 120
The top 10 power demands are important because it is – 124, 2015, sMART GRID TECHNOLOGIES. [Online]. Available:
possible to reduce the peak demand, thus reducing the total [15] http://www.sciencedirect.com/science/article/pii/S2212017315002467
V. S. Munisami Sarma, C. Selvam, A. , and K. Srinivas, “Advanced
power demand of the bill. This is possible by changing devices metering infrastructure for smart grid applications,” 06 2012.
by newest and efficiently ones, and changing consumption [16] T. N. Le, W.-L. Chin, D. K. Truong, and T. H. Nguyen, “Advanced
metering infrastructure based on smart meters in smart grid,” in Smart
habits. Metering Technology and Services-Inspirations for Energy Utilities.
IntechOpen, 2016.
V. F INAL C ONSIDERATIONS [17] R. R. Mohassel, A. Fung, F. Mohammadi, and K. Raa-
This paper presented a case study of AMI implemented in hemifar, “A survey on advanced metering infrastructure,”
International Journal of Electrical Power Energy Systems,
UFSM. The requirements of some smart grid applications were vol. 63, pp. 473 – 484, 2014. [Online]. Available:
presented in order to compare with the results obtained. The http://www.sciencedirect.com/science/article/pii/S0142061514003743
structure of UFSM’s AMI consists of a LoRaWAN network, [18] F. Adelantado, X. Vilajosana, P. Tuset-Peiro, B. Martinez, J. Melia-
Segui, and T. Watteyne, “Understanding the limits of lorawan,” IEEE
network server, SME Smart Meters, and an application server Communications Magazine, vol. 55, no. 9, pp. 34–40, Sep. 2017.
customized to the needs of the university. The results prove [19] R. HF. Lorawan iot industrial gw rhfs008p4g specifications.
that LoRaWAN network fits many smart grids applications, in- [Online]. Available: https://wiki.ai-thinker.com/ media/lora/docs/ai-
thinker lorawan industrial gw rhf2s008p4g specifications v0.3.pdf
cluding AMI. The latency achivied was 369ms and the average [20] F. IOT. Sme specifications. [Online]. Available:
packet delivery rate was 88.6%. Electric measurements allows http://foxiot.com.br/energia.html
different analysis to improve the energy efficiency and quality.
The top 5 building with unbalanced loads, top 10 with highest
consumption and power demand were ranked, allowing the
engineering team to take actions. Besides that, the buildings
with voltages above or below the limits were detected too.

Vous aimerez peut-être aussi