UNIT V HIGH PERFORMANCE COMPUTING FOR SMART GRID

APPLICATIONS
HAN:
In-House Communication Technologies
By establishing an ICT infrastructure for metering gateways, enhanced energy management
for decentralized power generation (e.g., solar panels, wind power, power-heat coupling) and
Demand Side Management (DSM) (e.g., household appliances, electric vehicles [EVs]) are
feasible for private households. The integration of these components combined with a
decoupling of ICT and energy components comes along with the necessity for reliable in-
house communication networks. Several solutions are discussed for these use cases, like
Wireless MBus [10], ZigBee [11], Z-Wave, [12] and KNX-RF [13] for transmission of
metering data. For in-house energy management and home automation ZigBee [11], Wireless
LAN [14] and narrowband RF systems are used. As wired technologies and bus systems,
Broadband Powerline Communications (BPLC) [15], Ethernet [16], KNX [13], Konnex RF,
[13] and M-Bus [17] are widely deployed.
In order to ensure transparent connectivity to all in-house components, the comprehensive
introduction of HAN gateways is one of the central elements for combining the high demands
on security and providing an extensive connectivity to the prosumer‘s household. On the one
hand, the gateways provide firewall functionality; on the other hand, the gateways provide
connectivity to all HAN entities. Thereby functionalities like smart metering for multiple
metering devices, DSM for loads and decentralized power generation, as well as user
interaction are provided by the gateway.
The HAN gateway collects and stores metering data from several metering devices, such as
electricity, gas, water, and heating meters using dedicated wireless technologies. The
collected data are bundled and securely transmitted to the meter reading operator typically
using wireless wide area point-to-multipoint technologies.
The energy management can be done either through the prosumer itself, motivated by tariff,
or through the distribution network operator for controlled or emergency load reductions.
Therefore, interfaces to the prosumer‘s appliances, loads, and local power generation
components are provided by integrating these components into the prosumer‘s in-house
network. Depending on external pricing information, several loads can be controlled, e.g., the
charging process of EVs as well as controlling home heating systems. An intelligent washing
machine makes use of the dynamic tariff information by starting the washing procedure in
low-tariff periods and avoids starting it in high-level tariff periods. In addition to this, the
connection interface can be used for maintenance, remote configuration issues, and firmware
updates.
One of the key capabilities is the integration of decentralized power generation systems,
which will make up a large part of the power generation in future systems. Nowadays an
increase in local power generation installation can be observed, e.g., solar panels, wind power
plants, and combined heat and power generation. Referring to the communication market, a
dedicated infrastructure has to be provided by an operator that is also providing management
and installation services to the prosumers. In order to maintain a
reliable ICT infrastructure between the HAN gateways and the marketplace, a HAN gateway
operator provides the reliable ICT components for software updates, administration, and
configuration issues.
Local HAN ICT Gateways
In order to ensure transparent connectivity to all in-house components, the comprehensive
introduction of HAN gateways is one of the central elements for combining the high demands
on security and providing an extensive connectivity to the prosumer‘s household. On the one
hand, the gateways act as firewalls, and on the other hand, the gateways provide connectivity
to the HAN entities. Thereby two different forms of HAN gateway can be considered:
1) Metering HAN gateways
2) Management HAN gateways
Metering HAN gateways:
The metering HAN gateways collect and store metering data from several metering devices,
like electricity, gas, water, and heating meters. The collected data are securely transmitted
bundled to the meter reading operator. Security in this context comprises two services:
integrity protection and nonrepudiation for metering data, which is often done by using
digital signatures allowing the meter reading operator to validate that the metering data have
not been altered by an intermediate component. Furthermore, the bundled transfer may be
encrypted to ensure the privacy of user related data in environments, where physical access to
the transmission path cannot be secured. Moreover, a local feedback system may give the
prosumer transparent insight into his current energy consumption. In conjunction with
available tariff information, motivation for reducing overall power consumption can be
achieved.
Management HAN gateways
The management HAN gateway, which can be realized as an integrated HAN gateway or as a
separate hardware component, represents an enhancement of the metering HAN gateway and
offers management functionality to the prosumer‘s HAN entities. The energy management
can be done either through the prosumer itself, motivated by tariff or sustainability, or
through the Distribution System Operator (DSO) for controlled or emergency load
reductions. Therefore, interfaces to the prosumers‘ appliances, loads, and local power
generation components are provided. Depending on an external pricing information, several
loads can be controlled, e.g., charging of EVs or controlling a home heating system. For
example, an intelligent washing machine makes use of the dynamic tariff information by
starting the washing procedure in low-tariff periods and avoids starting it in high-level tariff
periods. In addition to this, the connection interface can be used for maintenance, remote
configuration issues, and firmware updates.
The reference architecture presented in the previous section introduced the integration of both
HAN gateways at the prosumer‘s premises. In general, three modes of operation are
presented in this section that combines various components of the reference architecture
differently:
_ Mode 1: Joint HAN gateway operator
_ Mode 2: Separated HAN gateway operator
_ Mode 3: Split HAN gateway
In modes 1 and 2 only one HAN gateway exists. Mode 1 provides the HAN gateway operator
and the meter reading operator at the same instance, which ensures the operational
infrastructure between the HAN gateway and the remaining service infrastructure. Compared
to this, mode 2 explicitly separates the HAN gateway operator and meter reading operator.
Hereby an independent HAN gateway operator is responsible for the operational
communication infrastructure between the consumers‘ gateways and the remaining service
infrastructure. The gateway gets installed, configured, and administrated by the HAN gateway
operator. The acquisition of measurement data is done through a dedicated infrastructure,
which is provided by the HAN gateway operator.
The HAN gateway in mode 3 is split into two physical entities, the metering HAN gateway
maintained by the HAN gateway operator and the management HAN gateway maintained by
the customer himself. The metering gateway is installed, configured, and administrated by the
HAN gateway operator. By using this operational scenario, both gateways are separated and
only linked by interface IF 5, which is used for administrative purposes in the management
HAN gateway and for transmission of power consumption data to the user interaction system.
The gateways may be virtually separated or physically separated, depending on the
underlying deployment scenario.
HOME AREA NETWORKS (HANs):
One potentially beneficial function of the Smart Grid is management of residential energy
consumption. Home energy management includes a wide spectrum of devices, including
lights, appliances, heaters, air conditioners, local generation facilities (such as solar panels),
and electric vehicles, if present. Home energy management systems (HEMS) are being
developed and deployed to manage these energy consumption, storage, and generation
devices. Communications among these devices are supported over a home area network. (See
Fig. 5.5.)
HANs may include wireless networking technologies such as the ones defined by ZigBee
standards. ZigBee standards are developed and specified by the ZigBee Alliance. ZigBee
2007 is a suite of high-level communication protocols supported over the PHY and MAC
layers of IEEE 802.15 wireless networks [802-15-11]. Zig-Bee enhances the IEEE 802.15.4
standard by adding mesh networking and security functions required in HAN environments
[ZIGBEE12]. ZigBee is predominantly used in the 2.4 GHz unlicensed band but may also be
regionally used over the 915 and 865 MHz unlicensed bands in the United States and in
Europe, respectively. PLC is another networking technology used in HANs. The HomePlug
networking specifications developed by the HomePlug Alliance and leverage existing home
wiring are examples of a PLC technology that can be used in HANs [HOMEPG12].
Existing appliances and thermostats may be monitored and controlled using external devices
specifically developed for the appliance. In addition to on/off control, such devices may also
provide electric measurements such as consumption as well as the status of the appliances,
thermostats, and lights. These attached devices may be connected to the HAN for the purpose
of remote monitoring and control. Newer appliances, thermostats, and light switches may
already be equipped with such control and communication functions. Homes with local DG
(such as solar panels) and/or electric vehicles may also be connected to the HAN.

Depending on the functions supported, the HEMS remotely monitors energy consumption,
storage, and generation units connected into the HAN. The HEMS monitors energy
consumption/supply into and out of these systems. Further, HEMS may control their
operation, turning them on and off as needed and managing their energy efficiency. If
allowed by the utility, the (smart) meter may be connected into the HAN, allowing the HEMS
to get access to meter measurements. HEMS may also provide access to the devices attached
to the HAN from smart phones and PDAs from outside of the home.
HEMS can connect to the home Wi-Fi network to connect to the Internet. In this case, the
HEMS may receive energy management services from a third-party provider of these
services. To support such functions as automated demand response (to be described later in
this chapter), a utility may allow the HEMS to communicate with the utility EMS (UEMS).
In addition, the UEMS may report to the HEMS information including past consumption,
future consumption predictions, and consumption anomalies. The HEMS–UEMS
communication may be carried through the meter connected into the HAN and the utility
network. However, if the utility does not allow for such connection through the meter, the
UEMS–HEMS communication will be carried over the Internet or other IP network through
the utility Web portal.

WIDE AREA NETWORK:

We formalize the concept of WAN as an interconnection of WAN routers (WRs) as
shown in Fig. 6.2.
In this figure, no specific interconnection between the WRs within the WAN is shown or
implied. In particular, the oval in the figure is only an iconic representation of the WAN and
should not be misconstrued as a ring interconnection of the WRs. All network endpoints
(including endpoints of utility applications – both traditional as well as the new Smart Grid
applications) connect to WRs. For most part, no direct physical (L1) or link layer (L2)
connections between endpoints of an application are allowed in the architecture definition.
(For example, no direct connections between an IED and SCADA master control or between
a meter and theMDMS are allowed.) These endpoints connect to WRs. Traffic between the
pairs is routed by the respective WRs in the WAN. The end-to-end connection between
endpoints of an application will be an IP connection. Direct connections between application
endpoints may be allowed in cases where the network implementation cannot support
application requirements such as delay (e.g., teleprotection).
WAN Architecture
The network design will dictate the placement of WRs and the physical interconnections in
theWAN. For network reliability, it is important that there be at least two
disjoint physical paths between every pair ofWRs. Additional routers called Interior Routers
(IRs) may be deployed in the WAN (based on the design) for shorter paths between pairs of
WRs (see Fig. 6.3). Many utilities already have extensive fiber plants in the form of OPGW
and/or leased optical fiber. Utilities may also have deployed microwave infrastructure to
support point-to-point links between endpoints in its service area. These existing fiber (and
additionally the microwave) assets can be used as WAN links between WRs and IRs. A few
typical examples of the L1/L2 protocols over the physical links between WRs and IRs are
shown in Fig. 6.4.
Legacy network connections may need to be upgraded. Deployments of Packet over
SONET/SDH (POS) or Ethernet over (dark) fiber or overWDM protocols over the fiber links
are typical upgrade paths. POS (Fig. 6.4a) can be implemented on a fiber strand between the
two routers or over a single DWDMor CWDMwavelength. Similarly, Ethernet connections
can be used to connectsWRs/IRs (Fig. 6.4b). There are many advantages (including cost
advantages) to the use of Ethernet compared to SONET/SDH. These advantages make
Ethernet the protocol of choice for new and existing WAN deployments.
Packet microwave products provide point-to-point Ethernet connections over microwave
links (Fig. 6.4c). In order to support legacy TDM connections, the Ethernet MAC layer needs
to be deployed over a physical layer capable of transmitting and receiving bits to support time
synchronization. One way of supporting this functionality is through the use of standardized
network-based timing protocols such as IEEE 1588v2. An alternative approach for the
introduction of point-to-point Ethernet connections is to reuse legacy TDM microwave links
and deploy PPP at the link layers.
Packet microwave products support Ethernet ports, which can be used to connect to the router
or to the Ethernet LAN at that router location. There are other options possible for L1/L2
connections in addition to the three shown here. For example, point-to-point Ethernet can be
implemented over SONET/SDH. T1/T3 (or E1/E3) leased lines and digital microwave
(T3/E3) are other possibilities for L1 connection with appropriate L2 protocol such as PPP.
Frame Relay connections over NSP Frame Relay services is another option. WRs not only
serve the purpose of providing connectivity to the endpoints but also aggregating the traffic
to/from these endpoints. Thus, traffic from one endpoint of an application connected to a
WAN router WR1 may need to be routed to another router WR2 to which the other endpoint
of the application is connected. The routed traffic may traverse the WAN from router WR1 to
WR2 over zero or more intermediate WRs. It is also possible to consider the WRs and IRs as
LERs and LSRs so that non-IP traffic is supported over the LSPs created in the MPLS
infrastructure (see Fig. 6.4d).
Generally, the WRs (and IRs) may be conveniently located at existing utility facilities such as
DCCs and some substations. It may be necessary to deploy WRs and IRs at additional
locations based on the overall network design of the WANs and the FANs for connecting the
network endpoints.
WAN over Network Service Provider Networking Service
The WAN interconnection described so far requires that the utility own and/or lease the
individual L1 and L2 links between the routers, so that the required IP connections can be
configured. The utility may also create a WAN by connecting the WRs to one or more NSP
services. Five example NSP services are shown in Fig. 6.5.
1. Leased TDM Lines (Fig. 6.5a):
The required WAN connections are realized by connecting the WRs over leased TDM lines.
If IRs are required, the WR–IR connections are also supported over the leased lines (not
shown in Fig. 6.5a). L2 and IP configurations as well as the network operations are the
responsibilities of the utility. PPP (or POS with SONET/SDH connections) can be used at the
L2 level. Ethernet over SONET/SDH is another viable option. If additional TDM line
redundancy is required, the WRs/IRs may be connected with redundant leased lines or by
using ring protection if SONET/SDH connections are used.
2. Frame Relay Service (Fig. 6.5b):

The FR option is included here only for the completeness of presentation. NSPs are phasing
out (―sunsetting‖) FR service, giving way to other network services such as MPLS services.
Links between the WRs (and between WRs and IRs if IRs are deployed) are FR PVCs with
appropriate CIRs. All WRs (and IRs if deployed) connect to FR switches over TDM
connections. A utility can then configure IP over FR in the WAN. There are other Frame
Relay alternatives offered by the NSPs, for example, AT&T‘s M-VPN service.
3. Metro Ethernet Service (Fig. 6.5c):
Metro Ethernet service can provide pointto- point Ethernet connections between the WRs (or
between WRs and IRs, if deployed). The available data rates are 1 and 10 Gbps. NSPs may
provide other data rates by rate limiting these connections. All WRs (and IRs, if deployed)
connect to Metro Ethernet service switches over fiber connections. If available, Metro
Ethernet service can provides direct Ethernet connectivity at very high data rates. A utility
can then configure IP over Ethernet in the WAN.
4. VPLS Service (Fig. 6.5d):
If a Virtual Private LAN Service is available (which is an L2 MPLS service), it can be used
to provide Ethernet broadcast domain for the WRs (and IRs, if deployed). All WRs (and the
IRs) must connect to the VPLS service PEs over Ethernet connection over fiber or over TDM
connections. However, the L2 protocol between the WRs (and between the WRs and IRs) is
Ethernet. A utility can then configure IP over Ethernet for the WAN.
5. VPRN Service (Fig. 6.5e):
Virtual Private Routed Network service (which is an L3 MPLS service) can be used to create
a private IP network among the WRs (and IRs, if deployed) over the NSP service. All WRs
connect to the VPRNservice‘s PEs over TDM connections. PPP (or POS) is used as the L2
protocol between aWR acting as the CE router of the VPRN service and the PE. Ethernet
connection between the WR (acting as the CE) and the PE is another option. VPRN service
can be used to create independent IP connections between the WRs, analogous to four NSP
services defined above, with each IP connection being a single VPRN. Creating a private IP
network of WRs has many advantages over the other four options of Fig. 6.5. The utility does
not need to manage the internals of the WAN. There is no need to deploy IRs. (We are
assuming that the VPRN service provides for the required reliability of connections between
the WRs.) Further, from the utility perspective, each WR is only one IP hop away from every
other WR in the private IP network (though, in reality, a packet from one WR to the other
goes through at least one PE router and possibly over multiple PE and P routers in the NSP’s
MPLS network). The utility must make sure that the NSP supports the requirements for
network reliability, security, and performance. Depending on the service availability, the
WAN may be created using one of the NSP services shown in Fig. 6.5 or a combination of
more than one type of these services. Even when the utility owns fiber (and possibly
microwave) infrastructure, the WAN may be extended to additions of WRs by using one or
more of these NSP services.

BPL(BROAD BAND POWER LINES):

Distribution networks at Medium Voltage (MV) levels (i.e. 1 - 36 kV) comprise mainly
underground cables, overhead lines and related infrastructure, including secondary
substations (SS). SSs are the natural locations in which Smart Grid services are concentrated.
One of the main challenges for the Smart Grid is to connect all relevant SSs with central
utility systems. Applications can include the transmission of bi-directional aggregated
Advanced Metering Infrastructure (AMI) or sensor data between the Low Voltage (LV)
network and the utility as well as Distribution Automation (DA) such as the extension of IEC
61850 applications between the SSs [1]. The use of Broadband Power Line Communications
(BPL) technology within the MV distribution network can provide a number of advantages
compared to competing communication technologies [2]. Advantages provided by BPL
include the fact that the communication network remains largely in control of the utility, the
number of necessary backbone connections to the utility through the Wide Area Network
(WAN) may be significantly reduced as well as supporting Ethernet/IP-based real-time and
multicast applications. The current requirement for existing Smart Grid applications in the
MV network is the ability to transmit a minimum of 100 kbps of real-time simultaneous
bidirectional application throughput for all deployed SSs. The design of the
telecommunications network must cope with this performance, and must be guaranteed ex-
ante at the design and planning phase with little or no previous on field assessment. Next to
the performance requirements, there is also a need to provide a solution based on an
international standardized BPL technology (e.g. IEEE P1901 or ITU G.hn) in order to ensure
equipment availability and reduce the investment risk for the utility. Furthermore, especially
considering DA applications, the BPL network must provide high availability and quick
recovery even in the presence of network disruptions. While the use of BPL in the MV
network has been investigated since the days when last-mile Internet distribution was driving
the market [3], it has received relatively little attention compared to LV-indoor and LV-
access applications. Much of the existing academic work has focused on the modeling of the
MV BPL communications channel or providing optimizations at the link level. However, in
order to provide a large-scale deployment of BPL in a MV network, there are several system
level and practical aspects which must be considered. Therefore, this paper will focus on
describing some of the system level challenges in deploying BPL for MV smart grid
applications.

MV NETWORK CHARACTERISTICS
The MV network mainly differs from the LV access and indoor network in terms of the
physical topology, cable/wire types and link distances. A broad range of SSs can be found
depending upon the area (urban, suburban or rural) and consumption levels. Big utilities may
operate hundreds of thousands of SSs. Such a heterogeneous set of SSs, however, presents a
fundamental common infrastructure: MV lines which interconnect SSs among themselves
and to Primary Substations (PS). The underlying topology of this interconnection can be
considered as a meshed ring topology in which a certain amount of redundancy is provided
between SSs and PSs (see Fig. 1 for an example). Links between SSs are usually very
heterogeneous with several different cable types as well as a combination of overhead lines
and underground cables being found in a single geographical area. While underground cables
represent point-to-point links with relatively stable loads and impedances, overhead MV
lines, on the contrary, may present taps in a tree-like topology [4]. As links are repaired or the
network is modernized, multiple cable segments spliced together may be found across a
single link. Splices are a rather common occurrence and, depending upon the types of the
spliced segments, may result in almost no transmission loss or a rather high loss (>10dB).
Another rather common characteristic of underground MV networks is that multiple cables
may be buried together in a common duct for a certain distance. Depending upon the length
of the parallel runs as well as the cable types, the crosstalk attenuation between parallel
cables may be rather low which will have an impact on the channel between SSs with cables
in a common duct. This can lead to the introduction of connections or a change in the quality
of the communications channels between SSs which are located multiple physical ―hops‖
apart from each other in the electrical topology.
One typical misconception regarding the use of BPL in MV networks is that link distances
are too long to support any reliable communications. Attenuation is a very important factor
and will increase with distance and frequency, and as a consequence, longer MV links have
to use frequencies in lower bands to guarantee a minimum performance. While it is true that
BPL cannot achieve 100% coverage of all links in a typical MV network (especially in rural
areas), measurements have shown that raw data rates of several 10‘s of Mbps are possible for
links up to 500 m [5]. A general rule-of-thumb is that older Paper Insulated Lead Covered
(PILC) and newer polyethylene (PE) insulated cables will support sufficient throughput on
cables lengths up to 450 and 900 meters, respectively. Measurements on overhead lines are
currently being performed with initial results showing that longer link distances compared to
underground cables can be supported. An analysis of urban MV networks in Spain as
reported in [6] show that roughly 90-95% of the links will generally support BPL, i.e. provide
sufficient throughput to meet application requirements. It should be noted that the inability to
provide 100% coverage will be no different than any other telecommunications technology.
BROADCAST DOMAIN AND REDUNDANCY
As previously mentioned, the topology of a MV network can be described as a ringed mesh
topology in which SSs may have redundant paths to a single or multiple PSs. This means that
SSs will have anywhere from one MV feeder (endpoint) to several feeders per station. For the
case of multiple feeders, individual phases (3-phase system) are connected across a common
bus bar. Feeder lines are switched within the electrical network such that a connected tree
structure without loops is achieved. Load management and fault isolation can lead to manual
or automated switching being performed in the network. In order to achieve independence
from the underlying electrical network topology, but also to provide increased reliability
through redundancy (redundant paths) coupling is generally performed on the feeder side of
the switch (opposite the bus bar). This ensures that the logical PLC network topology remains
independent of the MV network‘s current switched state. Complete independence of the
switch state, however, cannot be achieved as the coupler loss will be dependent upon the state
of the switch due to fact the impedance seen by the coupler will change during switching.
Inductive coupling will generally function better in a closed switch state (low-impedance)
and capacitive coupling in an open switch state (high-impedance). Therefore, changes in the
switch state can instantaneously introduce a change in signal attenuation of ±20dB or higher.
Unfortunately, for fault isolation applications, it is exactly in this time that reliable real-time
communications between SSs are necessary. Meaning that a quick recovery time at the
network level is required, which can be very challenging.
The selection of a coupling method is influenced by the cable/wire type as well as the
characteristics of the transformer and switch. However, coupler selection may also be
influenced by practical limitations such as confined space, safety distance requirements or
other installation constraints. Safety regulations will generally require that couplers be
installed on non-energized lines. In order to ensure that no power outages occur (as they
could affect a very large area); this requires the preparation of a detailed network switching
plan for the temporary re-routing of the MV distribution. Any such switching, of course,
brings with it an inherent risk and must be carefully planned. In some cases, the installation
of couplers may require the approval of the local or national regulations body. This means
that close support from the utility is required not only for the planning, but also the
installation phase. This additional risk and planning effort is one of the reasons that channel
measurements in MV-networks have not been made on the scale as is found in similar LV
networks as the access is limited and the costs of coupler planning and installation can be
very high.
The potential presence of multiple feeders connected across a common bus bar means that a
PLC signal injected at one SS may propagate in multiple directions (across each feeder), but
also over multiple hops (across the bus bar of neighboring SSs), thereby creating a broadcast
channel. On the one hand, the broadcast nature of the PLC signal in a MV network provides
the potential for increased redundancy both through multiple paths (feeders) as well as across
multiple hops along a single feeder. If a cable fault occurs, data may be routed across an
alternative path or if a node failure occurs an intermediate node may be used as a relay. This
is a very important aspect for Smart Grid communications in order to increase the network
reliability against node and link failures. On the other hand, the interference domain of a SS
may be rather large which can reduce the potential for resource reuse in the network. This
problem is further complicated by the previously mentioned parallel cables in a common duct
which leads to the fact that SSs located a large number of ―hops‖ from each other may still
exist in each other‘s interference domain. The size of the broadcast domain for each node
may change dynamically over time as the channel conditions and network switch state
changes. With the standardized BPL technology there is currently no practical method
available for limiting the interference domain size other than reducing the transmission power
spectral density. Coupling to alternating phases on neighboring links has been investigated,
however has been found to only provide minimal signal attenuation.