Smart Grid: A Demanding Use Case For 5G Technologies
Smart Grid: A Demanding Use Case For 5G Technologies
technologies
Helen C. Leligou, Theodore Zahariadis, Lambros
Sarakis, Eleftherios Tsampasis Artemis Voulkidis, Terpsichori E. Velivassaki
Electrical Engineering Dept. Power Operations Ltd.
Univ. of Applied Sciences of Central Greece Swindon, UK,
Psahna, Evias, Greece artemis@power-ops.com
{leligou, zahariad, sarakis, etsampasis}@ teiste.gr
Abstract— The energy sector represents undoubtedly one of patterns. Accordingly, the next generation network should be
the most significant “test cases” for 5G enabling technologies, due capable of handling the complex context of operations and
to the need of addressing a huge range of very diverse support an increasingly diverse set of new and emerging
requirements to deal with across a variety of applications services, all of them with extremely diverging requirements,
(stringent capacity for smart metering/AMI versus latency for which will push mobile network performance and capabilities
supervisory control and fault localization). However, to to their limits. Furthermore, it should provide flexible, smart
effectively support energy utilities along their transition towards and scalable adaptation and/or association of the available
more decentralized renewable-oriented systems, several open network resources to the requirements of the supported
issues still remain as to 5G networks management automation,
services, enabling a dramatic paradigm shift from legacy
security, resilience, scalability and portability. To face these
issues, we outline a novel 5G PPP-compliant software framework
CAPEX to the OPEX “Everything-as-a-Service” driven
specifically tailored to the energy domain, which combines i) business models.
trusted, scalable and lock-in free plug ‘n’ play support for a Although a variety of software frameworks and reference
variety of constrained devices; ii) 5G devices’ abstractions to architectures have already made available for 5G enabling
demonstrate mMTC (massive Machine Type Communications), technologies, there is still a clear gap to bridge for 5G seamless
uMTC (critical MTC) and xMBB (Extended Massive deployment within a number of “vertical” sectors such as smart
BroadBand) communications coupled with partially distributed,
grid and smart city, which pose significant new requirements.
trusted, end-to-end security and MCM to enable secure, scalable
Among others, the energy “vertical” represents one of the most
and energy efficient communications; iii) extended Mobile Edge
Computing (xMEC) micro-clouds to reduce backhaul load, demanding “use/test case” for 5G enabling technologies,
increase the overall network capacity and reduce delays, while mainly due to the need of addressing a huge range of very
facilitating the deployment of generic MTC related NFVs diverse requirements to deal with across a variety of
(Network Function Virtualisation) and utility-centric VNFs applications (stringent capacity for massive smart
(Virtual Network Functions). metering/AMI (Advanced Manufacturing Infrastructure)
services versus stringent latency for supervisory control and
Keywords—Smart energy grids; 5G; preventive maintenance, fault localization) [1]. As a matter of fact, the combined effect
resilience; demand response; Machine Type Communications; of growing penetration of distributed Renewable Energy
Virtual Network Functions. Sources (RES) in the generation portfolio together with the
European Union’s “Customer Centric Energy Systems” vision,
I. INTRODUCTION1 which aims at turning energy consumers into active
“prosumers”, are dramatically changing the way in which
The wide deployment of IoT devices, broadband and energy distribution grids operate. To underline the penetration
mission critical services along with a large variety of scenarios, of RES with unpredictable energy generation patterns
ranging from smart city to factory automation, are paving the integration, please consider that on 15 May 2016, RES
way for a novel and disruptive 5G communication network, supplied nearly all of German domestic electricity demand[2]
which will enable huge capacity, zero delay, faster service while the political importance is evident by the agreement that
development, elasticity and optimal deployment, less energy took place on 6 June between the North Sea region countries
consumption, enhanced security, privacy by design and (Belgium, Denmark, France, Germany, Ireland, Luxembourg,
connectivity to billions of devices with less predictable traffic the Netherlands, Norway and Sweden) to create good
1
conditions for the development of offshore wind energy [3].
The work presented in this document was funded through
H2020- NRG-5 project. This project has received funding from Technological advances, political visions and market
the European Union’s Horizon 2020 Programme (H2020- 5G- liberation are transforming the energy network from a closed,
PPP) under Grant Agreement No. H2020- ICT-762013. monolithic and highly predictable infrastructure to an open,
multi-owned, decentralized ecosystem and pose huge
challenges, both in functional (i.e. stability, resiliency and Energy as a Service at large Scale, placing emphasis on
highly availability) and in non-functional (i.e. sustainability, security, privacy, trust and high availability. This work was
security, privacy and CAPEX/OPEX) directions. In this new developed in the framework of H2020 -NRG-5 project which
and time varying energy landscape, 5G initiative is challenged aims to additionally deliver innovative open-source prototypes,
to guarantee optimal communications of the energy grid, which state of the art laboratory experiments and heterogeneous real-
is believed to be the most complex, heterogeneous and gigantic life trials to draw valuable.
machine ever made in human history.
The rest of the paper is organized as follows: in section II
In particular “last mile” of the smart energy network has the we explore the main challenges that smart grid operations
highest potential for demonstrating the added value of the 5G impose on 5G technologies. These challenges have been
unified approach. While smart energy grids observability (in validated by smart grid operators from the NRG-5 consortium.
particular in the case of smart electricity grid) is already in Then, in section III, we define novel concepts that can
place in the High and mostly in the Medium Voltage branches efficiently address these challenges and are in line with the 5G
of the energy networks, situational awareness of Low technology principle and vision. In section IV we propose a
Voltage/Low Pressure branches is lagging behind. The state of novel architecture that proposes concrete developments that are
the art is actually for substation-level/pumps monitoring via required to make 5G technology a perfect enabler for next
SCADA, without considering real time energy consumption or generation smart grid operations. Finally, section V concludes
energy production feedback from prosumer, which would the article.
allow a finer-grained prediction of the demand and an
improved load balancing of the energy networks. Hence smart II. UTILITIES’ COMMUNICATION CHALLENGES
energy “last mile” network represents an ideal vertical for
extensive 5G deployment, where different applications with Starting from the communication challenges that smart
different requirements have to be managed: energy faces in view of the Smart Grids establishment, we
investigate the stringent Machine to Machine (aka M2M) and
• Smart Grid applications, such as supervisory Machine Cloud Machine (MCM) Machine Type
monitoring (cyber monitoring and physical/aerial Communications (MTC) networking requirements of a range
surveillance), fault localization, isolation/self-healing of today’s smart (electricity and gas) meters, intelligent energy
and energy re-routing, requiring more stringent devices sensor/actuator devices and drones, identifying and
latency, highest availability and security (Mission handling current and anticipated limitations of communication
Critical) networks. These challenges and associated requirements will
• Advanced metering applications enabling the massive drive the definition of research concepts that relate to achieving
and lock-in free integration of end-users’ Smart Energy-as-a-Service by applying generic Network
infrastructure requesting more stringent capacity and Function Virtualisation (NFVs) and domain-specific Virtual
privacy (Massive IoT application) Network Functions (VNFs), leveraging on extended Mobile
Edge Computing (xMEC) and distributed energy flexibility.
• A combination of the above such as smart Electric
Vehicle charging, where 5G technology should be According to the 5G PPP Architecture Working Group, 5G
able to incorporate and address both latency and networks aim to natively meet the requirements of three groups
capacity more stringent requirements. of use cases[5] :
Beyond applications requirements, significant challenges a) Massive broadband (xMBB) that delivers gigabytes
are posed from the energy infrastructure complexity and of bandwidth on demand
heterogeneity. The huge diversity in variables such as b) Massive machine-type communication (mMTC) that
population density, service territory size, control and connects billions of terminals.
monitoring technology, terrain and topology, power plants
location and fuel, RES capabilities and budget of utilities for c) Critical machine-type communication (uMTC) that
new deployments, as well as the different bandwidth and allows immediate feedback with high reliability and enables
latency requirements of applications within each utility, has for example remote control over robots and autonomous
resulted in the deployment and management of several, legacy driving.
communication solutions. Only 24% of utilities manage just Targeting the Smart Energy vertical domain, we define use
one communications network, with 58% of utilities have cases covering all three groups of MTC communications and
between 2 to 6 operating networks, 14% between 7 to 10, and aim to examine how already defined 5G radio features could be
4% have 10 or more networks[4], with a significant complexity used in Energy use cases and define NFV concepts as
and financial burden to manage. contributions to 5G studies which would optimise the
In this paper, we focus on the exploration of requirements performance for 5G network in the upper layers, to support the
and identification of innovative concepts to be contributed to use case requirements.
the 5G PPP/5G Initiative research and development activities
towards the realization of a Smart Energy as a Service use case A. Device identification and network autoconfiguration
that will stress 5G current results. We aim to advance beyond According to the Smart Grid paradigm, customers will play
state-of-the-art in virtualization-based communication an active role in energy flexibility, being equipped with
networks technologies, making them suitable to support Smart components for decentralized energy generation and local
(possibly moving) energy storage. In such a framework, smart the load will peak in the evenings at about 38GW, introducing
metering devices are quickly proliferating in number and need very severe stability risks to the Utilities due to the mobility
to exhibit a far more complex profile than today, offering patterns of the EVs that introduce uncertainty in the grid
services beyond traditional 15 minutes reporting, including management. As the penetration of EV rises above 10% the
support for real-time measurements, service discovery, risk will further increase, demanding that solutions be
infrastructure automation and AAA. However, for security, researched and tested and to be ready for deployment before
cost and operational rationality reasons, implementing the the volume of EVs becomes too large, and the cost of retro-
above features at full scale at smart meters level is, usually, fitting solutions become prohibitive.
impractical, also due to the high numbers of devices that need
to access the network; smart metering should be kept at low In the same context, low cost, Phasor Measurement Unit
cost, hence resource constrained. In turn, the currently (PMU), enabling precise state measurements to be made across
employed smart meters’ authentication solutions and an entire gird, support fast monitoring of distribution feeders,
centralized key management repositories create an informal with data refresh of 10 to 50 times per second. However, such
customers’ lock-in. update frequencies open up vast new possibility for the fine-
grained network control needed to manage the complex future
The challenge that 5G technologies has to face is to provide grid, but also permitting potentially devastating attacks though
a framework that will allow for easy, real-time, automated manipulation and data mirages.
devices identification so that network auto-configuration can be
achieved automatically. Unified AAA should be achieved in a Last but not least, the increasing proliferation of EVs has
homogeneous manner, to reduce the chances of AAA been paving the way for the massive deployment of smart EV
misconfigurations among different services of the same or chargers, which will offer near real time communicating and
different tenant, to address multi-tenancy under geographically interfacing with EV driver, with EV battery management
unbound mobility scenarios. Last, secure communications system and with DSO, with a view to determine what
should be achieved irrespectively of the network service appropriate scheduling of EVs may provide flexibility benefits
provided and the physical entity initiating the connection. to DSO (Vehicle to Grid service).
The challenge that 5G technologies has to face is to support
B. Predictive Maintenance for utility infrastructures fast response. Given that most of the times, storage is not
Predictive Maintenance of distributed generation plants, available on-site, ultra-low (below 5ms) response from the
energy transmission and distribution networks, like electricity energy operation centre is of vital importance. The enablement
cables and isolators, and gas/LNG tanks, pumps and pipelines, of large scale DDR requires extreme (for today’s standards)
is an activity of utmost importance in achieving highest power communication requirements as metering and associated
network reliability. Especially in case of electrical isolators and computational processes should be performed at very high
the transmission lines along the electricity grid or the natural frequencies.
gas storage tanks, pumps or pipelines is critical with high
accompanying cost. III. APPROACHES TO MEET THE CHALLENGES
Recently, the Energy industry started to adopt manually To meet the described challenges, we have outlined novel
driven UAVs/Drones to perform visual inspections, but, again, approaches as follows. To meet the challenge for device
special flight control certification has been necessary and the identification and network autoconfiguration, our approach is
time required for such operations is, still, hindering wide grounded on a basic analysis of the considered use case which
adoption; using manually driven drones, an electricity shows the need for efficient mMTC communications. We
transmission network may need to be traversed three or four propose the realization of the “plug & play vision” in metering
times as each electricity transmission network has at least one resources by implementing a novel and scalable xMEC
power line for each electrical phase and a safety lightning paradigm. Beyond edge computing acceleration, we propose
protection cable, often with fibre optics inside, for the design of a group of VNFs to facilitate distributed, scalable
communication purposes. and trusted plug ‘n’ play functionality of hardware constrained
devices. In more details, we focus on:
The challenge that 5G initiative has to face is to offer low-
delay, 5G-enabled predictive maintenance for more efficient a) A virtual Terminals Self-Discovery (vTSD) VNF
operation, accidents avoidance and fast restoration of energy which, whenever needed, will initiate and broadcast discovery
networks, leading to reduced maintenance costs and increasing beacons and messages to identify nearby nodes and MEC
the QoE offered by the Utilities to the citizens. nodes.
b) A virtual Self-Organizing Networks (vSON) VNF,
C. Resilience and high availability via Dispatchable Demand
which will realize a lightweight path selection paradigm in a
Response (DDR)
multi-RAT landscape by selecting and establishing
Energy cannot be easily stored at large scale, so utilities networking paths between the nodes and the edge-routers
have traditionally matched demand and supply by shaping based on functional and non-functional requirements, e.g. trust
demand through Demand Response campaigns, inserting (very metrics, expected transmission count (ETX), Expected
expensive) peaking plants to cover peak demand, or importing
Transmission Time (ETT) and explicit policy rules.
power from other utilities. The FP7 FINESCE Project has
estimated that when EVs take-up reaches even 10% in the EU,
c) A virtual Machine-Cloud-Machine (vMCM) VNF, processor to a dedicated edge cloud accelerator, sense the
which will provide caching services for the smart meter values traffic demand and the mobility/distribution of the drone
on the xMEC, enabling scalable remote access and significant swarms. Additional vMPA and vDFC instances may be
reduction of the device energy consumption, as the devices deployed and/or their resources may be dynamically scaled-up
may remain more at sleep mode and consuming less energy. in locations where there is an increased traffic processing
demand. In this way, the access and especially the backhaul
d) A virtual Mobility Management Entity (vMME) network capacity is preserved, thus offering the capability to
VNF, which will provide for idle mobile devices paging and the network infrastructure to allow remote control and conserve
tagging (e.g. safeguarding the location of EVs or mobile more users/terminals, while mobile operators monetize on their
terminals in proximity). access/edge resources. Here we addresse: a) xMBB
e) A trust-service provisioning VNF based on communications via the vMPA VNF for video streaming from
Blockchains Processing (vBCP), which will provide a truly the drones and analysis to the xMEC and the utilities control
distributed key management mechanism based on the centre, and b) uMTC communications via the vDFC VNF for
blockchains technology. controlling the flight of drones.
f) A virtual Authentication, Authorization, and To meet the challenge for resilience and high availability,
Accounting (vAAA) VNF, which will combine novel, low we propose the realization of decentralized, communication
complexity cryptographic approaches implemented on the demanding services such as AMI as a Service (AMIaaS),
smart meter with blockchains implemented by the vBCP. As Dispatchable Demand Response as a Service (DDRaaS),
Energy Infrastructure Fault Localization as a Service (FLaaS).
blockchains offer a truly distributed keys repository,
All these services may be realized as closed loop applications,
combination with inherited physical functions may create an
integrating application specific logic with a complex
AAA that overcomes utilities lock-in. forwarding graph of collaborating utility-specific VNFs, such
It is important to note that all proposed VNFs may be as:
applied well beyond utility networks to any type of mobile
hardware constrained terminal. a) A virtual Phasor Measurement Unit (vPMU) VNF,
enabling precise state measurements to be made across an
To meet the challenge for predictive maintenance of the
energy infrastructures, we propose the use of semi-autonomous entire gird. In principle, fast sampling will be enriched with
swarms of drones and parallel survey from different software based synchronization algorithms. To ensure
views/cameras. The UAVs/Drones swarms need to run reusability, this VNF will be a result of combined SFC and
complex, bandwidth demanding, computationally heavy and VNFFG techniques. VNF
time critical applications, meeting a) operational requirements, b) A virtual Electricity Substation & Rerouting (vESR)
such as to define the flight plan for each drone in a swarm, so VNF, enabling control of the local substation and electricity
that they have optimal coverage with minimal resources, taking rerouting activities.
into account the flight capability of each UAV/drone and the
remaining energy, b) communication requirements, either by c) A virtual DER (vDER) VNF, enabling the optimized
cellular or satellite links controlling the drones flight and management and local control of distributed energy storage
uploading captured video and c) mission requirements, such as resources (like second life batteries, EVs on the move and EVs
object (i.e. lines, pipes, tanks, blades, towers) video analysis chargers), which will be in charge for providing low-latency
and inspection. flexibility provisioning services to the grid operator.
Pursuing this approach, we address: a) mMTC
Towards an energy infrastructure Predictive Maintenance
as a Service (PMaaS) application, integration of application communications via the huge number of RES, DES and
specific logic with a complex forwarding graph of VNFs is controlling units and b) uMTC communications as most VNFs
needed, such as: require real-time control of the smart energy services.
Mini-cell
Mini-cell
Network
Network
Open APIs
Pluggable
Power Plant EV
(Extra) High High Voltage Distribution Smart
Voltage (110kV) Grid Meters
(265 to 275kV)
Hydro-Electric Power Wind
Plant Farms
Plant PV Parks
Fig. 1. The proposed framework which exploits 5G concepts to build an enhanced smart energy grid
we consider many novel utility functions offered following With respect to traceable Critical Infrastructures SLAs (CI-
the “X as a Service” approach, separated by an Open SLAs) and Extended MEC, trust and security are considered to
Northbound API (NBI). play a major role in 5G systems. Especially critical
infrastructures such as electricity and gas storage and
With respect to decentralized trust & identity management, distribution infrastructures request for high availability and
we proposed the development of a smart meter device and resilience. We introduce the Critical Infrastructures SLAs (CI-
software protocol stack as a generic 5G terminal to validate SLAs) as the ability of the network, from a security and
decentralised trust and identity management. The idea of using resilience perspective, to provide explicit, measurable and
the Smart Meters for delivering value-added services, beyond enforceable capabilities enabling the network to behave
just billing (e.g. Demand Side Management [6], power quality consistently in a prescribed way. Such CI-SLAs will specify
monitoring [7] and fault detection [8]) is not new. Currently, the relevant security constraints for traffic, such as traffic
efforts on developing the NORM (Next Generation Open Real isolation and network slicing requirements, desired levels of
Time Smart Meter) are in place. Yet, little effort has been availability or the level of security assurance of network
placed so far in services requiring high accuracy and precise elements, while taking into account the energy asset severity,
time-tagging. So, we propose to leverage the H2020 SUCCESS the cascading effect versus the time to recover from a
unbundled smart meter architecture and introduce a scalable malfunction (incident, accident or attack) and the relevant
cloud-based stack, optional multi-RAT access interface and CAPEX and OPEX.
enriched security and trust features, to provide a next
generation smart meter as 5G device prototype. PUF To enable traceability and enforcement of the CI-SLA, we
encryption will be combined with Blockchains to offer a introduce the concept of Blockchains technology based VNF
decentralized trust & identity management mechanism, Descriptors (VNFDs) and expand the ETSI MEC concept to
supporting end-users’ privacy by design. In parallel, the self-* become a trusted Multi-RAT MEC (xMEC) enabling process
features will accelerate deployment. An abstraction layer will offloading from hardware constrained devices to the edge-
hide metering specific functionality (e.g. legacy meter for cloud infrastructure. By utilizing blockchains VNFDs as
electricity or gas, hard real-time features such as PMU processes tags, we introduce a traceable mechanism to monitor
functionality) and physical network interfaces (offering event and enforce CI-SLAS. At execution time, an offloading
satellite access in case cellular coverage is not available). In decision will be taken, based on the remote node and xMEC
this way, we will offer a novel, open solution, applicable not status (e.g. CPU load, memory, network delay, etc), policies
only for smart meters but any hardware constrained device, and license agreements. xMEC will employ fault-tolerant
with built in trusted and vendor/ utilities lock-in free Plug ’n’ replication techniques, to re-execute any failing task execution,
Play functionality. if this is feasible due to time constraints. Likewise, NRG-5 will
be able to recall any incident, and modify the permissions place real and diverse challenges on 5G technologies. We
granted to a migrated service or even revoke access to proposed novel concepts and approaches as well as a concrete
previously stored data. The combination of self- architectural framework that constitutes 5G an enabling
identifiable/traceable VNFDs and CI-SLAs monitoring play a technology for the next generation of smart grids. The
crucial role in the considered use cases and exploitation proposed architecture will be implemented under the 5G
potential, as they guarantee a realistic business model. We Public Private Partnership (5G-PPP) framework and the EC
propose the support of novel business models via migration funded project H2020 ICT-762013 NRG-5
micro-contracts, which could be negotiate and unexceptionably (http://www.nrg5.eu). The complete framework will be
monitored, while protecting the hosting node from malware, as
validated in laboratory environment and in real life trials
it could directly identify and quarantine malicious nodes and
offered by ASM Terni (electricity network) and ENGIE
applications at any time.
(Gas/LNG network).
Turning our attention to elastic VNF optimal sizing,
placing and chaining which is crucial to achieve smart energy
requirements, beyond development of generic and utility REFERENCES
specific VNFs, we propose the development of an automated, [1] 5G-PPP whitepaper: 5G Empowering Verticals (February 2016)
machine-learning based VNF handling mechanism for either https://5g-ppp.eu/wpcontent/uploads/2016/02/BROCHURE_5PPP_BAT
automatically sizing the deployed VNFs to much operational 2_PL.pdf
needs, or automatically deciding the correct VNF architecture [2] http://www.renewableenergyworld.com/articles/2016/05/germany-
and size to operate under the CI-SLA required. In a multi- achieves-milestone-renewables-supply-nearly-100-percent-energy-for-a-
tenant environment, the machine learning VNF handling day.html
mechanism will interface the northbound interface NBI of both [3] European Commission, “North Seas Countries agree on closer energy
cooperation,” Press release, Luxembourg, 6 June 2016,
the single- and the multi-tenant MANOs so that hierarchical http://europa.eu/rapid/press-release_IP-16-2029_en.htm
state-awareness is achieved. In this framework, elasticity of the [4] GTM Research, “Trends in Utility Smart Grid Communications
network services will be able, even when collaboration of Management,” sponsored by ERICSSON, 2013,
multiple communication service providers is mandatory. To https://www.ericsson.com/res/docs/2013/trends-in-utility-smart-grid-
support high flexibility, instead of relying on static definition communications-management.pdf
of VNFDs, a mix of machine learning (e.g. K-Means [5] https://www.ericsson.com/res/docs/whitepapers/wp-5g-security.pdf
Clustering combined with singular value decomposition and [6] Di Leo, G.; Liguori, C.; Paciello, V.; Pietrosanto, A.; Sommella, P.,
principal component analysis) and regression techniques (e.g. "Smart meters and demand response: A double challenge,"
logistic regression and support vector machines for regression) Computational Intelligence and Virtual Environments for Measurement
Systems and Applications (CIVEMSA), 2015 IEEE International
will be employed to reduce data dimensionality, cluster the Conference on , vol., no., pp.1,6, 12-14 June 2015
terminals based on their characteristics and forecast their [7] Wambaugh, J.O., "Smart metering for the distribution control room,"
behaviour in terms of communication requirements. Electricity Distribution (CIRED 2013), 22nd International Conference
and Exhibition on , vol., no., pp.1,4, 10-13 June 2013
V. CONCLUSIONS [8] Kuroda, K.; Yokoyama, R.; Kobayashi, D.; Ichimura, T., "An Approach
to Outage Location Prediction Utilizing Smart Metering Data,"
This paper explores the requirements imposed by smart grid Modelling Symposium (AMS), 2014 8th Asia , vol., no., pp.61,66, 23-
operations on 5G designs and it was found that smart grids 25 Sept. 2014.