IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 22, NO.

2, APRIL 2007 1021

Application of Sensor Network for

Secure Electric Energy Infrastructure
Ramón A. León, Student Member, IEEE, Vijay Vittal, Fellow, IEEE, and G. Manimaran, Member, IEEE

Abstract—Wireless sensor networks are becoming the tech- center only receives indication that an electrical fault occurred,
nology of choice for sensing applications mostly due to their ease but not if it is temporary or permanent. Therefore, the operating
of installation and associated lower costs. This paper proposes standards state that he/she has to try to reinsert the faulted line
a novel conceptual design for an application of wireless sensor in order to check the temporary/permanent condition of the
technology for assessing the structural health of transmission
lines and their implementation to improve the observability and
event. Once all of the attempts fail, then the line is marked
reliability of power systems. A two-layer sensor network model is as permanently out of service. The recent blackout events in
presented for overcoming the communication range limitations of the U.S. [3] and Italy [4] have shown that failure to assess and
smart sensors, and two operational modes for enhanced energy understand the condition of the power system and delay in
efficiency are introduced. Simulations integrating the output of taking appropriate corrective actions after just a single outage
the sensor network with an energy-management system were can lead to widespread blackouts of large areas of the power
conducted, obtaining improvement in the security of the power system.
system. We propose the utilization of wireless sensor network tech-
Index Terms—Energy-management system (EMS), network sen- nology for detection of mechanical failures in transmission lines
sitivity factors, power system security, power transmission lines, such as: conductor failure, tower collapses, hot spots, extreme
wireless sensor networks. mechanical conditions, etc. The proposed application involves
the installation of sensors for mechanical monitoring in prede-
termined towers of a transmission line and communicating via
I. INTRODUCTION
a wireless network. The main goal is to obtain a complete phys-
ITH the increasing threat of terrorism around the world, ical and electrical picture of the power system in real time, di-
W more attention is being paid to the security of the elec-
tric transmission infrastructure. Experiences in countries like
agnose imminent as well as permanent faults, and determine
appropriate control measures that could be automatically taken
Colombia, which has faced as much as 200 terrorist attacks on and/or suggested to the system operators once an extreme me-
its transmission infrastructure per year [1], demonstrate the vul- chanical condition appears in a transmission line.
nerability of the power system to these kinds of events. Although For evaluating the feasibility of the concept, a dispatcher
it is very difficult to avoid or predict when and where these ter- training simulator (DTS) based on the energy-management
rorist acts can occur, quick assessment of the situation can help system (EMS) platform from AREVA T&D was used for
operators to take optimal actions to avoid cascading events and simulating the operation of the electric power system in real
the consequent partial or total blackouts. time as it is monitored at an actual energy control center.
The mechanical failures resulting from malicious attacks on
a transmission line are basically the same as those that would II. MECHANICAL CHARACTERISTICS OF OVERHEAD
result when extreme natural events affect a portion of the trans- TRANSMISSION LINES
mission line. Thus, any analysis conducted in this regard can
also help in taking preventive and corrective action when acts A. Transmission Lines Components
of sabotage are directed on the transmission infrastructure. The design of overhead power lines conceptually considers
The current method to assess the damage caused by any them as composed of four individual components: foundations,
unexpected physical event on the transmission grid is the visual supports, interfaces and conductors, all of which have limited
inspection of the transmission infrastructure [2]. With problems mechanical strength [5]. The design process takes into account
that occur in concentrated environments, like substations or the coordination of the mechanical strengths since the failure of
generating plants, it is not difficult to find and assess the damage one of the four components may lead to the collapse of the entire
with a fairly small crew or with adequately localized video transmission facility.
surveillance. But in transmission lines which are geographically When the different components of a transmission line are sub-
dispersed over hundreds of miles, this task is more difficult. jected to their limits of strength, their failure could be sudden,
Nevertheless, once an event occurs, the operator in the control occurring in fraction of seconds and can result in instability, rup-
ture or complete separation; or it could be progressive, resulting
in loss of strength that eventually leads to damage after long pe-
Manuscript received April 15, 2005; revised April 4, 2006. This work was
supported by the National Science Foundation under Grants ECS-0428040 and riods of time.
COLCIENCIAS. Paper no. TPWRD-00226-2005. There are two main types of support structures, strain and
R. A. Leon and G. Manimaran are with the Department of Electrical and suspension supports.
Computer Engineering, Iowa State University, Ames, IA 50014 USA (e-mail: Strain and angle-strain supports carry the conductor tensile
[email protected]; [email protected]).
V. Vittal is with the Department of Electrical Engineering, Arizona State Uni- forces in the direction of the conductor and serve as rigid points
versity, Tempe, AZ 85287-5706 USA (e-mail: [email protected]). in the line. They are designed for conductor tensile forces dif-
Digital Object Identifier 10.1109/TPWRD.2006.886797 fering in both line directions and to provide protection against
0885-8977/$25.00 © 2007 IEEE
Authorized licensed use limited to: Muranga University College. Downloaded on February 15,2025 at 14:45:17 UTC from IEEE Xplore. Restrictions apply.
1022 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 22, NO. 2, APRIL 2007

cascading structural failures. For long, straight line sections, Bombing produces a more direct impact over transmission
strain supports are placed every 3 to 6 miles [5]. structures and is widely used as the main form of attack over
At suspension supports, conductors are fixed to suspension the electric infrastructure [1]. The impact of bombing a trans-
insulator sets, carrying the conductors in a straight vertical posi- mission line support goes from collapse of the entire structure
tion, swinging in the prevalent direction of the conductors. They to limited damage at the point of placement.
are not designed to transfer conductor tensile forces to the sup- The explosive blast produced by a bomb placed at the foot of a
ports other than for abnormal conditions, given that the design transmission structure produces vibration by means of the direct
provides for the longitudinal forces to cancel in both directions. impact of the expansive wave and indirectly due to ground in-
Any medium to large imbalance can potentially produce the col- duced vibration [10]. The duration of the vibration phenomenon
lapse of the entire structure. depends on a series of factors, from the size of the explosion
Given its relatively lower cost, large sections of a transmis- to the constructive characteristics of the support. If the blast is
sion line are compounded by suspension supports. Thus, be- not powerful enough to produce a complete collapse of the sup-
cause of its construction characteristics, when a failure occurs port, it would probably produce a tilt of the tower. In this case,
on any of the suspension members of a determined section, there differential longitudinal stresses can appear in suspension sup-
is a very high probability of collapse of the remaining suspen- ports and increased tension can be detected in strain/dead-end
sion supports in the affected section, producing a structural cas- supports.
cading failure.
C. Temperature Concerns
B. Mechanical Loads on Structures
The effect of the current flow on the rise of temperature on the
Overhead transmission lines, by the nature of their exposed conductors is well known. The loading calculations presented in
constructional characteristics are subjected to loads imposed [11] establish thermal limits for determined weather conditions
by the environment such as wind, ice, snow, earthquakes and and conductor characteristics.
flooding; and also to human related hazards such as accidents Hot Spots present another concern of temperature rise in
and terrorism. The effects of these loads produce detectable transmission lines, appearing in the coupling between ener-
forces in the components of the line that can affect their ability gized conductors and the interfaces. The appearance of hot
to withstand the operating conditions. spots may degrade the mechanical reliability of conductors and
Three different categories of wind induced conductor motion connectors producing a thermal runaway situation that could
are recognized, being differentiated by their frequency, ampli- lead to catastrophic failure of the point of attachment. Once a
tudes and effects on conductors, interfaces and supports [6]. hot spot appears, a reduction of the current flowing through the
Aeolian vibrations are the most common. Their main character- line is in order until it can be repaired.
istics are small amplitude and relatively high frequency. They
increase the tension stress on lines, produce conductor “turn”
and create vibration on the structures [5]. Conductor gallop III. PROPOSED WIRELESS MECHANICAL SENSOR NETWORK
is another form of wind effect characterized by vertical low
frequency and high amplitude conductor motion. It is usually A. Proposed Sensor Selection and Placement
caused by relatively strong and steady winds on asymmetrically There are a number of sensors capable of monitoring mechan-
iced conductors [6]. Galloping magnifies loads in a conductor ical variables on the line that could be used to detect abnormal
and especially the vertical end forces on the supports [7]. The conditions when extreme environmental events or human re-
third form of effect is the wake induced oscillation. It is peculiar lated accidents or sabotages appear. Since transmission line sup-
to bundled conductors, and occurs when relatively moderate port structures have constructive characteristics closely related
or strong winds act upon the line. Experience has shown that to building structures, it is proposed to use a set of sensors based
damage is largely localized in a few places on the line. in that application, where researchers have found that the utiliza-
Accumulation of snow and ice on conductors affect them in a tion of acceleration, strain and displacement sensors can provide
two fold way. It increases the tensile forces on the wires due to an appropriate level of observability for earthquakes and wind
the added weight, and additionally changes their aerodynamic [13], [14].
characteristics by changing the shape of the surface exposed to Given that temperature is also a concern in electric energy
the wind, with the effects related to those described previously. transmission; the application can take advantage of the sensing
The effects of human related events over transmission lines infrastructure to place temperature sensors at the attachment
can be detected by the same set of sensors intended to monitor point of conductors to detect possible hot spots and overheating
the effects of winds and ice because any accidental or malicious problems related to overloads.
event involves some kind of disturbance over the normal me- Table I presents an application matrix for selected sensors and
chanical operating characteristics of the line. The extent of the their response to any mechanical event appearing on the line.
damage caused by any accidental or malicious event cannot be This is based on observed characteristics following the event.
predicted since it depends on many variables. It is proposed that tension or strain sensors will be mounted
In October of 2003, in Oregon and California, there were at the interfaces of the strain supports. It is recommended to
some reported cases of bolt removal from the legs of 500 kV install them at all the conductor attachments of all strain struc-
transmission lines [8]. The act was apparently aimed at weak- tures because of their important role in maintaining the physical
ening the mechanical strength of the structures; however, no me- integrity of the transmission line. This results in a high level
chanical collapses occurred. There have been reports in the U.S. of observability on any transmission line for mechanical events
of cutting guy wires of strain/dead-end structures [9], also, re- that involve change over normal tensile conditions, such as high
ports from U.S. and Colombia show that cutting the legs of lat- winds, ice accretion or compromise of the structural integrity of
tice towers is a method used by saboteurs [1]. surrounding structures.
Authorized licensed use limited to: Muranga University College. Downloaded on February 15,2025 at 14:45:17 UTC from IEEE Xplore. Restrictions apply.
LEÓN et al.: APPLICATION OF SENSOR NETWORK FOR SECURE ELECTRIC ENERGY INFRASTRUCTURE 1023

Fig. 1. Overview of the proposed sensor placement.

TABLE I
SENSOR APPLICATION MATRIX
We propose the installation of accelerometers in the support
body for vibration and tilt monitoring, and in the conductor at-
tachment points for detecting wind induced vibration. Installa-
tion in conductors is recommended for maintenance optimiza-
tion, but if cost is a constraint, their application can be omitted.
Installation of temperature sensors for the detection of over-
heating can be optimized given that heating conditions due to
overloads are uniform in relative long portions of the line. How-
ever, since hot spots are highly localized phenomena, they can
only be detected by placing temperature transducers close to all
the points of attachment of conductors. Again, cost is the de-
termining factor for selecting the complete application for hot
spots.
It should be noted that all the measurements from the pro-
posed sensors are related in various ways. The relationship can
be between different sensors applied to the same structure, as in
the case of tension and strain, or between vibration and tilt, or it
can be between the same kinds of sensors at different locations,
as in the case of strain sensors applied to the conductors at both
ends of a suspension interface.
Additional event classification can be obtained by taking into
For a more complete assessment of the mechanical condi- account the difference on the dynamics of the failures. Wind
tions of the line, as in detecting the unlikely event of an isolated and ice accretion do not appear suddenly; in the case of winds,
failure, it is recommended to measure tensile forces on conduc- they increase in time with deteriorating climatic conditions. Ice
tors attached to a number of suspension supports, placing them buildup is a progressive phenomenon that gradually adds weight
on both sides of the point of attachment of the conductors at to the conductors, consequently, increasing tension and strain in
every third tower. In this manner, each support monitors only them. Accidents and sabotage are sudden; their effects may ap-
one phase conductor, but the system does not loose observability pear as a sharp increase of the sensed variables. An act of sabo-
because each attachment point not being monitored directly is tage involving weakening of the structure can be distinguished
monitored by an adjacent support. The overview of the recom- if there is a collapsed structure in which no anomaly in the vari-
mendation is shown in Fig. 1. ables was previously detected.
Authorized licensed use limited to: Muranga University College. Downloaded on February 15,2025 at 14:45:17 UTC from IEEE Xplore. Restrictions apply.
1024 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 22, NO. 2, APRIL 2007

greater than 100 feet given the dimensions of typical transmis-

sion line supports. A local data and communications processor
(LDCP) installed at each support is used to aggregate the
information from the LSG. Its radio can achieve a larger range
and comes with an increased communication bandwidth due to
the fact that it does not have size and power constraints. For that
matter, it can harvest power from an inductive source placed
near the closest phase conductor and can also come with a
bigger rechargeable battery. The normal range expected for the
application varies from 300 to 1500 feet, using more powerful
radios in particular structures where longer spans exist.
Sensor data on every LSG is aggregated and analyzed for ver-
ification purposes on the LDCP. Data verification is possible
Fig. 2. Two layers model. thanks to the inherent relation between the sensed variables, as
was discussed previously. Sensors on the LSG and their corre-
sponding LDCP will form the Layer 1 of the WMSN.
B. Proposed Architecture of a Wireless Sensor Network for The interaction between the LDCPs on each support is the
Transmission Lines basis for the Layer 2 of the WMSN and forms the inter-sup-
For the recommended implementation of a wireless mechan- port communications and collaboration (ISCC) Layer. This
ical sensor network (WMSN) for transmission lines, the local- layer handles all the message processing and transmission
ization of each node is determined in the pre-deployment phase, required for delivering the mechanical status information to the
with sensors of particular characteristics placed in predefined substation.
locations along the support structures and conductors. The total In this paper, a method is proposed to enable collaboration
cost of deploying the WSMN would depend on the cost of the between LDCPs in adjacent supports as well as data collection
sensor package, installation cost, and maintenance cost. The from all the LDCPs in a transmission line that involves sequen-
cost of a typical sensor package depends on the various phys- tial message broadcasting across the line. For executing both
ical quantities that it measures (variable cost) and the cost of functions, two modes of operation (partial mode and full mode)
wireless transceiver (fixed cost), and it is typically in the range are proposed for the WMSN.
of few hundreds to one thousand US dollars. The installation In the proposed operating principle, a local substation pro-
and maintenance costs depend on the geographical spread of cessor (LSP) at one end of the transmission line commands the
the transmission line (flat surface, terrain, forest, etc.) and other start of a data collection sweep, interrogating the first LDCP.
factors which is beyond the scope of this paper. The sensors can The message control header will contain the parameters indi-
be programmed to implement one or more communication pro- cating if the sweep corresponds to partial or full operation mode,
tocols, which can be activated on need basis. thus, instructing the LDCP what kind of processing to apply to
It is proposed to rely on wireless communication between its contents and its own data from the LSG. Directionality is
towers, since they would offer a reliable transmission path in achieved by including the sender and receiver address in the
the event of a failure of a support structure, provided that the message header. In this manner, when a substation initiates a
causal event does not damage the transmitter (as in the case of sweep, it places a zero value on the sending node address and
an explosion caused by an act of sabotage). 1 as the receiving node. This will instruct to process
The inherent linear characteristic of a transmission line drives the message and broadcast it to the next LDCP. This process
the overall topology of the sensor network. Communications will continue until the sweep reaches the substation’s LSP at
between nodes in such a topology are reduced to their adja- the other end of the line. Then, that LSP will trigger a data col-
cent node and at most two hops ahead (communications range lection sweep in the opposite direction by placing its address
permitting). Thus, for messages originating from a node in the ( ) as the sending node and as the receiving node.
middle of the line to reach the substation, they should be relayed The message body contains a vector of data values and the as-
through all the intermediate nodes. sociated addresses of the sensor nodes from which those values
The construction characteristics of transmission lines, with come. The size of the vector depends on the operating mode,
supports separated hundreds or even thousands of feet between being larger for full mode and reduced for partial mode.
each other pose a hard constraint for the range requirement for In the partial mode, messages exchanged by the LDCPs con-
wireless communications between sensor nodes localized on tain only a reduced set of the total data present on every LSG.
different structures. By design, the communications range of It contains the maximum value of each variable group (vibra-
smart sensors is not very long and extending it is not efficient tion, strain and temperature) and its associated sensor localiza-
due to power supply limitations [15]–[17]. tion. As a consequence, the size of the message is small enough
A two layers model is proposed to overcome the restrictions to provide high communication speeds. However, the time re-
imposed by the range/energy management issue on the sensor sponse characteristics that can be obtained from the WMSN
nodes as shown in Fig. 2. do not make it viable for fast acting applications like primary
The sensor nodes installed in each structure form a local or backup protection. On the other hand, it is suitable for in-
sensor group (LSG) with required communication ranges not tegration with the data acquisition functionality provided by
Authorized licensed use limited to: Muranga University College. Downloaded on February 15,2025 at 14:45:17 UTC from IEEE Xplore. Restrictions apply.
LEÓN et al.: APPLICATION OF SENSOR NETWORK FOR SECURE ELECTRIC ENERGY INFRASTRUCTURE 1025

the SCADA system. In the partial mode, various complete line • Normal (N): There are no indications of variables being
sweeps could be executed within one SCADA cycle. outside the normal operating limits. Tensions are within de-
The full mode collects the status information from all the sen- sign parameters and balanced at suspension supports. Vi-
sors in the transmission line, enabling the LSP to obtain a com- brations are minimal, and structures are in the vertical po-
plete picture of the mechanical status of the line. With the infor- sition.
mation obtained, the LSP can execute verification algorithms • Suspicious (S): Some of the variables are outside the
oriented to detect any inconsistencies within the collected data. normal range. Ice accretion, strong winds or maximum
Since the message size in this mode is larger than in the par- thermal loading can be events associated with this status.
tial mode, bandwidth utilization is also larger. For that reason, There could be measurable vibration and the inclinometers
it should be expected that end to end transmission would be sig- detect small tilt angles.
nificantly higher than that of the partial mode. • Imminent (I): Several variables are outside the thresholds
Failure detection of any LDCP is provided by means of for detecting mechanical problems, but there is no clear
implementing a timeout function in the preceding (sender) indication of a mechanical failure. High vibration levels
LDCP. Thus, when there is no receiving confirmation from the are detected and structure’s tilt angles are measurably dif-
next LDCP, it is marked as a broken link and the sweep reverses ferent from the vertical. Although an indication of failure
direction, relaying the failure information to the initiating LSP. is not present, this condition deserves attention from the
This could lead to further analyses for establishing if there power system operator since it could be an indication that
was in fact a mechanical event (from measurements around the there can be an electrical failure and/or a network topology
broken link) or just a LDCP failure. change some time soon.
1) Time Response Characteristics: Failures in the trans- • Fault (F): Mechanical failures can be detected by the
mission system that involve grounding of phase conductors, WMSN in different ways, by the excursion off limits of
or contact between them, induce short circuits that require to some variables, by values near to zero in others, or by the
be cleared in the shortest possible time. Electrical protection complete lack of signal from the sensors. Tension sensors
equipment has been used since the inception of the electric grid indicate a broken conductor by detecting abnormally high
for taking the appropriate corrective actions in a fast, selective or low values. Collapsed structures would be detected by
and reliable manner. Common response times for clearing tilt angles close to the horizontal.
faults are in the order of 50 to 100 ms. Therefore, for the The transitions between the proposed 4 states establish the
WMSN to be a plausible tool to carry out protective functions permanent or temporary nature of a failure and their dynamics
and provide fault signaling with adequate timing, messages can establish the natural or human nature of a mechanical event.
from the faulted element should reach the substation in about In normal operation and even under high wind/ice conditions
50 to 100 ms through the WMSN. This is an almost impossible in which there is no compromise of the mechanical integrity
requirement for today’s technology, because of limitations on of the line, it is unlikely that electrical faults will occur unless
bandwidth and processing speed at LDCPs. Hence, it is con- vegetation plays a role. Therefore, a direct relation between the
cluded that the WMSN cannot be used to provide principal or WMSN and the electrical protective elements of the line is not
backup protective functionality. On the other hand, the WSMN expected. When events develop leading to a compromised or a
technology can be used to satisfy a slightly coarser timing re- completely failed structure, it is expected that the conductors
quirements, such as the SCADA system. The SCADA collects will have a contact or grounding, resulting in phase to phase,
information from the substations typically every 4 s. Therefore, phase to ground or even three phase short circuits. In this case,
it would be desirable that the LSP at the substation achieve at high correlation can be obtained between the output of the elec-
most one overall diagnostic of its supervised lines for every trical protective equipment and the output of the WMSN.
SCADA cycle. Given the latter requirement, it is expected that
the transmission for delivering a complete information package B. IPSS Software
from one end of the line to the other, plus the processing time For each of the different failure modes, different actions need
at the substation, should not exceed the SCADA cycle time. to be taken in the power system in order to avoid cascading
For small size messages, this communication requirement can events that could lead to a collapse of the interconnected power
be satisfied with current technology. system. For testing the feasibility of using the WMSN, we fo-
cused on transmission overload management driven by the fault
IV. PROPOSED TOOL FOR WMSN/EMS INTEGRATION classification. That is, if a line trips and that produces an over-
load in another transmission line, the fault classification pro-
vides two ways to handle the event. For a temporary fault, the
A. Mechanical/Electrical System Failure Modes
recommended actions are to wait and try to reclose the faulted
As discussed in Section II, the mechanical health status of a line in a few minutes due to the purely electrical nature of the
transmission line can be influenced by the environmental con- fault, but for a permanent fault, the recommendation begins
ditions and by human actions. The severity of the mechanical with blocking any further reclosure and immediate initiation of
status can be established by comparing the values of the vari- the recovery process that can include generation rescheduling
ables against the different limits defined by the design of the and/or load shedding.
transmission line. A classification for the mechanical status is By using the network sensitivity factors proposed by Wood
introduced here as follows: and Wollenberg in [19], the power injections that contribute
Authorized licensed use limited to: Muranga University College. Downloaded on February 15,2025 at 14:45:17 UTC from IEEE Xplore. Restrictions apply.
1026 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 22, NO. 2, APRIL 2007

most to a line flow can be identified by examining the resultant

generation shift distribution factor (GSDF) vector for a partic-
ular line, taking into account that any change in the power in-
jection is compensated by a change in the opposite direction
by the slack bus. It is to be noted that the GSDF factors de-
pend on the network configuration and also generation and load
pattern. Under different contingencies these factors change. In
this paper, without loss of generality, the GSDF factors have
been applied to illustrate the potential benefit of using mechan-
ical sensors data in operations setting to aid operator decision
making.
For determining the combination of injection increase and de-
crease that contribute most to line flow changes, it is necessary
to first identify the node with the maximum value of GSDF in
the original GSDF vector as Fig. 3. IPSS output.

Then, the GSDF of each node referenced to the pivot is ob-

tained by

Once this is established, the combination of injections that

contribute the most to line flows can be selected by taking the
previously obtained pivot node and the maximum value of the
new vector (opposite node 1). In the same manner,
the second most contributing combination is identified by se-
lecting the second largest (opposite 2) and the second smallest
(pivot 2) values of the vector. We use the previ-
ously identified set of four nodes for determining the appropriate
changes in their injections for managing the overload in a par-
ticular line. If the generation at the nodes obtained is involved
in the AGC function, the next largest or smallest node will be Fig. 4. EMP60 test system.
selected.
The direction of generation change for each node depends A Visual C++ program was developed for implementing the
on the signs of both the line power flow and the GSDF vector. generation shift strategy previously described. The integrated
When the opposite 1 and opposite 2 nodes have the same sign power system security (IPSS) program performs the real-time
as the line flow, they have to decrease generation in order to al- assessment of the mechanical/electrical situation collecting the
leviate the overload. When they have different signs, they have power system’s electrical status data from the EMS/DTS and
to increase generation. The pivot and pivot 2 nodes change gen- the mechanical status information from the WMSN every 8 s.
eration in the opposite direction of the opposite nodes. An xml data file produced by the central processor supplies the
Once the nodes to reschedule generation are identified, the mechanical status information to the IPSS, while the interface
generation allocation for each direction of change (up or down) with the AREVA e-terra platform is built on the DDE service
that would provide the fastest overload relief in the line can be provided by Microsoft Windows, using AREVA’s HABDDE
established by server.
The mechanical/electrical status processing flowchart is pre-
sented in the Appendix wherein option exists to classify faults
into imminent, suspicious, or permanent using sensor data.
The IPSS user interface is shown in Fig. 3.
where
V. SIMULATION RESULTS
change of generation required in unit ;
unit generation rate of change; The integrated operation of the IPSS software, AREVA’s DTS
GSDF value for the unit referenced to and the WMSN concept was tested on the EMP60 power system
line ; model (Fig. 4), assuming that a wireless mechanical sensor net-
work is installed in the line Martdale–Ceylon 345 kV for mon-
overload in the line . itoring its mechanical health.
Authorized licensed use limited to: Muranga University College. Downloaded on February 15,2025 at 14:45:17 UTC from IEEE Xplore. Restrictions apply.
LEÓN et al.: APPLICATION OF SENSOR NETWORK FOR SECURE ELECTRIC ENERGY INFRASTRUCTURE 1027

Fig. 7. Line flows showing relief actions–imminent fault case.

Fig. 5. Line flows–collapse case.

Fig. 8. Line flows showing relief actions–sudden fault case.

Fig. 6. Voltages–collapse case.

TABLE II
GENERATION SHIFT RECOMMENDATIONS With the reduction of the line flow as a result of the rec-
ommended actions by the IPSS, the line Chenaux–Picton will
achieve 100% loading (700 MW) after the outage of Mart-
dale–Ceylon (Fig. 7).
If the outage of Martdale–Ceylon is sudden (i.e., without a
progressively deteriorating mechanical condition), the combi-
nation of WMSN and IPSS can help the operator to take fast and
appropriate actions to reduce the time interval during which the
The following simulations model different mechanical failure line Chenaux–Picton is overloaded as shown in Fig. 8.
modes in the monitored line and their associated dynamics. The
objective of the test is to verify the appropriate response and
VI. CONCLUSION
recommendations provided by the IPSS software as the power
system is simulated in the dispatcher training simulator. This paper proposed a novel approach for using wireless
Figs. 5 and 6 show the evolution of the system without sensor technology to assess the mechanical health of transmis-
WMSN after the outage of the line Martdale–Ceylon and the sion lines. The proposed two layers architecture provides an
consequent overload of Chenaux–Picton. It can be seen that if approach to overcome the range limitation of the smart sensors
during the time spent for reclosing the line, an additional outage installed in the supports, while offering a complete monitoring
occurs in Chenaux–Picton due to its overloaded condition, the environment for a transmission line.
system experiences a voltage collapse as shown in Fig. 6. The simulation studies show that the WMSN can help opera-
Assuming the presence of the WMSN monitoring the line tors take fast and appropriate decisions based on the mechanical
Martdale–Ceylon, a simulated imminent failure mode will failure modes detected by the WMSN.
trigger the IPSS to recommend generation shift actions before Optimized maintenance practices can also be achieved by the
the actual outage of the monitored line as shown in Table II. analysis of the measurements provided by the proposed sensor
Authorized licensed use limited to: Muranga University College. Downloaded on February 15,2025 at 14:45:17 UTC from IEEE Xplore. Restrictions apply.
1028 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 22, NO. 2, APRIL 2007

system. Collected statistical information about stress and vibra- [7] T. Wipf, F. Fanous, M. Baezinger, S. Gupta, and R. Anjam, in Ice Storm
tion in conductors can help maintenance engineers to optimally Damage Assessment of the Leigh-Sycamore 345 kV Transmission Line,
Ames, IA, Jul. 1991, Iowa State Univ. rep.
schedule preventive maintenance in lines with excess mechan- [8] “Acts of sabotage involving high voltage transmission towers,” in
ical stress. The ability to detect hot-spots can provide surveyors Department of Homeland Security Information Bulletin, Washington,
D.C., Oct. 24, 2003.
with select locations to perform more thorough infrared analysis [9] U. S. Congress Office of Technology Assessment Physical Vulnera-
in structures. Our current work focuses on developing models bility of Electric System to Natural Disasters and Sabotage U.S. Gov.
that establish the relationship between various physical quanti- Printing, Washington, D.C., 1990, Rep. OTA-E-453.
[10] F. J. Lucca, Tight Construction Blasting: Ground Vibration Basics,
ties and the eventual fault scenarios, and integrating the model Monitoring and Prediction. Granby, CT: Terra Dinamica LLC, 2003.
into state estimation and other applications. [11] IEEE Standard for Calculating the Current-Temperature Relationship
of Bare Overhead Conductors, IEEE Std. 738, 1993.
[12] J. Fraden, AIP Handbook of Modern Sensors. New York: American
APPENDIX Inst. Phys., 1993.
STATUS PROCESSING FLOWCHART [13] N. Kurata, B. F. Spencer, Jr., and M. Ruiz-Sandoval, “Risk monitoring
of buildings using wireless sensor network,” presented at the Int. Work-
shop on Advanced Sensors, Structural Health Monitoring, Smart Struc-
tures, Mita, Japan, Nov. 10–11, 2003.
[14] J. P. Lynch, A. Sundararajan, K. H. Law, A. S. Kiremidjian, T. Kenny,
and E. Carryer, “Embedment of structural monitoring algorithms in
a wireless sensing unit,” Struct. Eng. Mechan., vol. 15, no. 3, pp.
285–297, 2003.
[15] I. F. Akyildiz, W. Su, Y. Sankarasubramaniam, and E. Cayirci, “A
survey on sensor networks,” IEEE Commun. Mag., vol. 40, no. 8, pp.
102–114, Aug. 2002.
[16] D. Culler, D. Estrin, and M. Srivastava, “Overview of sensor networks,”
Comput., vol. 37, no. 8, pp. 41–49, Aug. 2004.
[17] F. L. Lewis, “Wireless sensor networks,” in Smart Environments: Tech-
nologies, Protocols, and Applications, D. J. Cook and S. K. Das, Eds.
New York: Wiley, 2004.
[18] P. Jalote, Fault Tolerance in Distributed Systems. Upper Saddle
River, NJ: Prentice-Hall, 1998.
[19] A. J. Wood and B. F. Wollenberg, Power Generation, Operation and
Control. New York: Wiley, 1996.
[20] J. R. Berryhill, C++ Scientific Programming. New York: Wiley,
2001.
[21] B. Stroustrup, The C++ Programming Language. Reading, MA: Ad-
dison-Wesley, 1997.

Ramón A. León (S’04) received the B.E. degree in electrical engineering from
Universidad Tecnológica de Bolívar, Cartagena, Colombia, in 1997, and the
M.S. degree from Iowa State University, Ames, in 2005.
Currently, he is a Planning Analyst with the Interconexion Electrica S.A. ISA,
Medellin, Colombia, where he has been since 1997.

Vijay Vittal (M’82–SM’87–F’97) received the B.E. degree in electrical engi-

neering from B.M.S. College of Engineering, Bangalore, India, in 1977, the
M.Tech. degree from the Indian Institute of Technology, Kanpur, India, in 1979,
and the Ph.D. degree from Iowa State University, Ames, in 1982.
He served on the faculty of the Department of Electrical and Computer En-
gineering at Iowa State University from 1982 to 2004. He is currently the Ira A.
ACKNOWLEDGMENT Fulton Chair of the Electrical Engineering Department, Arizona State Univer-
sity, Tempe.
The authors would like to thank Areva T&D for providing Dr. Vittal received the 1985 Presidential Young Investigator Award and the
the Dispatcher Training Simulator and the help required for its 2000 IEEE Power Engineering Society Outstanding Power Engineering Edu-
successful implementation in our power systems laboratory. cator Award. He is a member of the U.S. National Academy of Engineering.

REFERENCES
[1] Internal Rep. Guerrilla Attacks on the National Interconnected System G. Manimaran (M’99) received the Ph.D. degree in computer science and en-
Interconexión Eléctrica S.A. Medellín, Colombia, 1998–2004.
gineering from the Indian Institute of Technology (IIT), Madras, India, in 1998.
[2] A. J. Pansini, Transmission Line Reliability and Security. Liburn,
GA: Fairmont, 2004. Currently, he is an Associate Professor in the Department of Electrical and
[3] Final Report on the August 14, 2003 the Blackout in the United States Computer Engineering at Iowa State University, Ames. Prior to this, he was
and Canada: Causes and Recommendations U.S.–Canada Power an Assistant Professor in the same department. His research expertise is in the
System Outage Task Force, 2004. areas of Internet QoS and security and real-time systems. He is coauthor of
[4] Interim Rep. Investigation Committee on the 28 Sep. 2003 blackout in many research publications, of which two conference/workshop papers received
Italy UCTE, 2003. the best paper awards. He is a coauthor of Resource Management in Real-Time
[5] F. Kiessling, P. Nefzger, J. F. Nolasco, and U. Kaintzyk, Overhead Systems and Networks (MIT Press, 2001). He was Guest Co-Editor for a special
Power Lines – Planning, Design, Construction. Berlin, Germany: issue of IEEE Network, the Journal of High Speed Networks, and the Journal of
Springer-Verlag, 2003. Systems and Software, and served as Chair for several reputed workshops.
[6] Transmission Line Reference Book (Wind-Induced Conductor Mo- Dr. Manimaran has given tutorials and was Session Chair and member of the
tion). Palo Alto, CA, Electric Power Res. Inst., 1979. program committee in several IEEE conferences. He is a member of the ACM.

Authorized licensed use limited to: Muranga University College. Downloaded on February 15,2025 at 14:45:17 UTC from IEEE Xplore. Restrictions apply.