1962 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 25, NO.

3, JULY 2010

Reliability Evaluation of an HVDC Transmission

System Tapped by a VSC Station
Sajjad Zadkhast, Mahmud Fotuhi-Firuzabad, Senior Member, IEEE, Farrokh Aminifar, Student Member, IEEE,
Roy Billinton, Life Fellow, IEEE, Sherif Omar Faried, Senior Member, IEEE, and
Abdel-Aty (Aty) Edris, Senior Member, IEEE

Abstract—Reliability evaluation of multiterminal HVDC sys- I. INTRODUCTION

tems has received very little attention in the past. This paper
presents reliability modeling and analysis of an HVDC trans-
mission system incorporating a voltage-sourced converter (VSC)
tapping station. The use of VSC tapping stations enables the
D URING the last three decades, reliability analysis of
power systems considering transformers, breakers, trans-
mission lines, and generators failure has been widely applied.
supply of power along the route to areas with comparatively little
consumption. In this paper, a comprehensive detailed reliability This analysis has been done especially in the generation and
model is developed and then converted to a manageable and transmission levels and is mainly restricted to high voltage
computationally efficient model. Using this equivalent reliability alternating current (HVAC) systems. HVDC transmission tech-
model, various reliability indices are calculated at the load point nology has proved to be well suited for specific applications,
of the system and the impacts of the VSC tapping station on including long-distance bulk power transmission, particularly
these indices are illustrated. Sensitivity analyses are conducted
to investigate the impacts of the load level and the location of the between remote generation points and load centers, relatively
tapping station on the reliability indices. The conducted studies long cable interconnections, interconnections between large
are numerically applied on a typical HVDC system and thorough isolated HVAC systems, and asynchronous tie-lines between
discussions are presented. or internal to HVAC systems. Investigations have shown that
Index Terms—Hybrid HVDC system, reliability modeling, tap- HVDC systems are more economical than HVAC systems
ping station, voltage-source converter (VSC). particularly for long transmission distances, where equal power
can be transferred using two wires as compared to three wires
using an HVAC system.
NOMENCLATURE There have been studies conducted on the reliability analysis
of HVDC transmission systems over many years [1]. The relia-
bility of an HVDC system is dependent on its design and on the
investment made in the system. Cost-benefit aspects have been
ACF AC filter.
included in the reliability analysis of HVDC systems in [2]. A
DCF DC filter. technique has been published in [3] based on failure modes and
Brk Breaker. effects analysis (FMEA). In this method, the event tree and min-
imal cut set approach are utilized to represent the system op-
Cap Reactive compensating capacitors. erational behavior and to deduce the appropriate system failure
Trn Transformer. modes. Statistics show that the availability of HVDC systems in
operation is over 90%. This completely depends on the failure
Vlv Valves.
rate and mean time to repair of the equipment associated with
SR Smoothing reactor. such dc systems. In [4], the effects of different components on
DCTL DC transmission line. the overall reliability of HVDC systems are investigated. How-
ever, the effects of generation, DCTLs, and load level on system
DCSW DC switch. reliability are not covered by [4]. An energy-based approach is
Subsys Subsystem. presented in [5] to evaluate the performance of HVDC transmis-
sion systems. The impact of group connected generators versus
a single generator on the reliability evaluation of an HVDC
Manuscript received November 03, 2009; revised December 24, 2009. First system was also investigated in [5]. Well-being analysis of an
published April 12, 2010; current version published June 23, 2010. Paper no.
TPWRD-00816-2009.
HVDC transmission system is presented in [6] and [7]. A hybrid
S. Zadkhast, M. Fotuhi-Firuzabad, and F. Aminifar are with the Centre of Ex- operation of static synchronous series compensators (SSSC) and
cellence in Power System Control and Management, Electrical Engineering De- an HVDC line is analyzed in [8]. It was revealed in [8] that
partment, Sharif University of Technology, Tehran 11365-11155, Iran (e-mail:
[email protected]; [email protected]; [email protected]).
changes in failure rate, repair time, and spare components in-
R. Billinton and S. O. Faried are with the Power Systems Research Group, stallation time can considerably affect the system reliability in-
University of Saskatchewan, Saskatoon, SK S7N 5A9, Canada (e-mail: roy. dices.
billinton @usask.ca; [email protected]). Only two terminal HVDC transmission systems were studied
A.-A. (A). Edris is with Siemens Energy Co., Mountain View, CA 94043-
1338 USA (e-mail: [email protected]). in the noted literature. In some applications, the use of multi-
Digital Object Identifier 10.1109/TPWRD.2010.2042470 terminal HVDC systems could be required. These systems can
0885-8977/$26.00 © 2010 IEEE
ZADKHAST et al.: RELIABILITY EVALUATION OF AN HVDC TRANSMISSION SYSTEM 1963

be supplied from more than one point and can supply loads to analysis procedure is also discussed in this Section. Numerical
more than one point. For example, these lines could pass through studies of the proposed models along with sensitivity analyses
regions which have low cost renewable energy resources such are presented in Section IV. Section V summarizes the con-
as wind power generation. It is also possible that there may be cluding remarks.
some small load centers where expanding a specific ac trans-
mission system to supply them cannot be economically or tech- II. HVDC SYSTEM WITH A VSC TAPPING STATION
nically justified. Accordingly, the operation of multiterminal Different HVDC transmission systems have different compo-
HVDC lines to deliver power to distant loads or transfer power nents. Some components, however, are common in all HVDC
from cost-effective generators will be highly desirable in the transmission systems. These components are converter stations,
near future. Reliability analysis of multiterminal HVDC sys- DCTLs as well as ACFs, DCFs, SRs, caps, etc. Also, each
tems and systems with tapping stations is essential and required HVDC system has some poles in its structure which work in
in both power system long-term planning and short-term oper- parallel. Each pole works independently and the system can
ation studies. operate without one or some of its poles but at derated ca-
Reliability evaluation of multiterminal HVDC systems has pacity. Fig. 1 shows a hypothetical 500 kV HVDC transmission
received very little attention. Reference [9] proposed a relia- system with a VSC tapping station. This system has three main
bility modeling technique for a multiterminal HVDC system segments: 1) the sending end; 2) the receiving end; and 3) the
in a sub-transmission level. A recursive method for calculating tapping station.
load point indices is presented in [9]. However, this paper did The following elements are present at the sending and re-
not investigate the effect of different converter failures on the ceiving ends:
system reliability indices. Virtually, no research work has been
cap these capacitors compensate reactive power
reported on the reliability evaluation of HVDC transmission
used by the HVDC system;
systems with tapping stations.
Tapping bulk transmission HVDC has been seen as being ACF ACFs filter harmonics produced by Vlvs;
quite costly. The reason for this is that using tapping stations of Trn transformer changes the voltage level to acquire
the same type as the main transmission line commutated con- proper dc voltage level and also act as a low
verter will cost almost the same as a full size converter. More- pass filter;
over, if the tapping station feeds a weak network, a strong reac-
tive power support (e.g., a synchronous condenser), is required Vlv power-electronic switches convert ac voltage
to avoid commutation failure problems. However, much smaller to dc;
stations can be built and the influence on the main transmission SR smoothing reactors eliminate high frequency
will be very small if the voltage-sourced converter (VSC) tech- current ripples;
nology is used in the tapping stations as VSC operation does
DCF DCFs reduce ripple in the dc voltage.
not rely on the strength of the connected ac system. The main
benefit of using a VSC tapping station is the ability to provide Each pole of the system is usually composed of Brk, Trn, Vlv,
on route power supplies to rural and urban communities in close and SR. ACFs, caps, and DCFs can be added to these poles or
proximity to an HVDC transmission line [10]. can be considered separately.
In this paper, the incorporation of a VSC tapping station reli- Fig. 2 shows elements of the VSC tapping station. This
ability model in the overall HVDC system model is presented. system has almost the same components as the sending end
The reported study examines the effect of this station on dif- except DCSW, which disconnect the tapping station from
ferent reliability indices. First, the system under study is intro- HVDC system.
duced and divided into three main segments, the sending end, Technical and economical justification of a tapping station re-
the receiving end, and the tapping station. In each segment, par- quires taking many aspects into account. In the following, some
allel and series components are combined to form subsystems. crucial points are discussed in summary. The points affecting
For each subsystem, an appropriate reliability model is con- the reliability modeling issue are discussed with more detail in
structed and these models are combined to make the reliability Section III.
block diagram of the corresponding segment. Reliability models In the system shown in Fig. 1, the available generation
of the three noted segments are combined using the equivalent capacity at the sending end is normally higher than the trans-
assisting unit approach to obtain the entire system reliability mission system capability. The situation is enhanced by adding
block diagram and the reliability indices are then calculated. a tapping station. In such cases, all the power at the assisting
The proposed model is numerically analyzed and the calcula- system is not transferred through the transmission system. The
tion procedure is demonstrated with thorough discussions. The assistance power capability is limited by both the assisting
effects of the load level and location of the tapping station are system capacity and the transmission system limit. For ex-
investigated. ample, if the transmission system capacity is equal to 1 p.u.,
The rest of the paper is organized as follows. Section II in- the equalized subsystem which represents the generation side
troduces the HVDC system with a VSC tapping station. In Sec- and the transmission system cannot have a state with capacity
tion III, the system is divided into three segments and after ex- of more than 1 p.u. When the total available generation at the
tracting associated reliability models of the segments, the re- sending end and the tapping station is more than 1 p.u., this
liability model of the whole system is devised. The reliability creates congestion in the transmission system. In this paper,
1964 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 25, NO. 3, JULY 2010

Fig. 1. HVDC transmission system with a VSC tapping station.

the performance of each element in the system. An accurate re-

liability analysis of an HVDC system needs a precise reliability
model. The first objective of the paper is to construct an accu-
rate and comprehensive reliability model. This model is based
on the configuration of the HVDC system and its subsystems.
A reliability model for an HVDC line with a solid link be-
tween the sending and receiving ends was presented in [3]. In
the study described here, the sending and receiving ends do not
Fig. 2. Configuration of a VSC. have a solid link and, therefore, each part should be considered
separately.
In order to develop the reliability model of the system, two
it is assumed that sending end generators have higher priority issues are taken into account as follows.
than the tapping station. If congestion occurs, the output of the 1) The total system consists of a number of components con-
tapping station is first reduced until the congestion is eliminated nected in a series/parallel structure. Failure of a component
and if this is not sufficient, the sending end power is decreased. in a parallel structure can result in a decrease in the max-
Installation of a tapping station and converting the energy imum transferable power by the system while a failure in
from ac to dc or vice-versa impose some costs on the power a series structure causes failure of the whole system.
system (investment and installation costs associated with Trn, 2) Some parts are common and can be shared by all poles.
Vlvs, SR, ACF, etc.). These costs increase by enhancing the This means that they belong to the entire system and not to
power and voltage level of the station. Such a converter station a specific pole. Some components are however dedicated
could improve the reliability indices, which could in turn de- to a specific pole. For example, in Fig. 1, ACFs and caps
crease the expected interruption cost of the system. There also are common and can be used by all poles while SRs are
might be congestion in the HVDC lines in some hours of day designed to be used by only one specific pole.
and the tapping station may not be used in those hours. Thus, Fig. 1 shows the system under study, which consists of two
considering congestion, economical, and technical aspects, one poles and a tapping station. The entire system is large and hard
should decide what converter capacity is the most beneficial for to analyze and is therefore divided into three major segments,
the overall system. namely, the sending end, the receiving end, and the tapping sta-
tion. Also, each segment is divided into a number of subsystems
III. RELIABILITY MODELING OF AN HVDC SYSTEM to facilitate the modeling process. The procedure used to model
WITH A VSC TAPPING STATION the system is as follows:
The first step in the determination of the reliability model for 1) The reliability model associated with each subsystem is
any system is to understand the function of the system, the con- developed.
straints under which it operates, and the root cause of its failure. 2) The reliability model of each segment is then developed by
The availability and reliability of an HVDC system depends on combining its subsystem models.
ZADKHAST et al.: RELIABILITY EVALUATION OF AN HVDC TRANSMISSION SYSTEM 1965

Fig. 3. Reliability model of a pole. Fig. 4. Reliability model for the VSC.

3) The reliability model of the whole system is obtained by the complete capacity outage probability table (COPT) [12].
combining the reliability models of the three segments. The complete COPT includes the probabilities and upward and
4) The reliability indices are calculated using the final relia- downward departure rates associated with each state. Using
bility model and the operational considerations. these data, other important parameters, such as the frequency
These steps are further discussed in the following of each state and the duration for which it is likely to exist, are
subsections. computed.

A. Sending End B. Receiving End

The sending end can be categorized into four subsystems. The From a converter connection point of view, the receiving end
first subsystem consists of ACFs and caps. The second one con- is analogous to the sending end, with relevant component reli-
sists of Vlvs, Brk, SR, and Trn. The DCFs make up the third sub- ability data and transmission line length. The reliability models
system and the fourth subsystem is composed of two DCTLs. for the subsystem components in the receiving end are therefore
Each one of these four subsystems is represented by an equiv- similar to those associated with the sending end. The reliability
alent reliability model with up/down and one or more derated models of Subsystems 5, 6, 7, and 8 are similar to those of Sub-
states. Each group of elements can therefore be considered as a systems 4, 3, 2, and 1, respectively.
single element in further modeling steps.
Subsystem 1) Caps and ACFs are common among all poles C. Tapping Station
and work in parallel with each other. Experiments have Analyzing the tapping station involves two different stages.
shown that there is usually no problem in the system, if The first one is the reliability modeling of the components. The
only one of these elements fails. A failure in more than one second one is based on the operational analysis related to the
component will result in a reduction in the maximum trans- various operating conditions such as generation, load, or both
ferrable power. Different combinations of available ACFs generation and load in the tapping station. These two stages are
and caps can be considered and each combination pro- discussed here.
duces a different level of sending end power. These ACFs 1) Modeling: The T-connection in the middle of the trans-
and caps are represented by a single equivalent reliability mission line is actually an ac/dc or dc/ac converter. If there is a
model with some derated states. load on the ac side of the converter, the converter will act as an
Subsystem 2) The next parallel components are poles. Any inverter and transfer energy from the dc side to the ac side. If,
problem in a pole brings out the pole from the system but however, a generator is installed on the ac side of the converter,
does not necessarily interrupt load supply. Overlapping the HVDC system receives energy from the tapping station. The
failures between two poles isolate the sending end. An- VSC tapping station connects the DCTLs to the tapping bus. The
alyzing overlapping failures between different poles and VSC configuration, as shown in Fig. 2, is composed of Vlvs,
common components is very involved and requires con- DCSWs, Brk, Trn, ACF, cap, and DCF. Since a failure of any
siderable computational effort. It is, therefore, preferable one of these components results in the failure of the entire tap-
to create an equivalent reliability model of each pole. This ping station, the reliability model is composed of series blocks,
model is shown in Fig. 3 and consists of series elements as shown in Fig. 4. As noted earlier, series elements can be com-
where failure of one of them causes the corresponding pole bined using the approximate equivalent series method [11].
to be unavailable. These series elements can be combined 2) Analysis of the Operational Modes: Three different oper-
using the approximated equivalent series method [11]. ational modes designated as generation, load, and a combination
Subsystem 3) The DCFs are common and can be used of generation and load modes are considered for the tapping sta-
by all poles and therefore, they are modeled separately. tion. These modes are described as follows.
Failure in each DCF reduces the maximum transferrable Generation mode: When there is a generation system on
power. The DCFs are therefore represented by a multistate the ac side of the tapping station, this generation can as-
model. This equivalent model is in series with the models sist the HVDC system up to the point where there is no
associated with Subsystems 1 and 2. congestion in the transmission line, and the load is sup-
Subsystem 4) This subsystem includes the transmission plied through two sources. Failure of the sending end or
lines between the sending end and the tapping station. of the tapping station will not isolate the load. A receiving
Since the outage of each line reduces the sending end end failure, however, can either reduce the maximum trans-
capability to 0.5 p.u., these parts can be represented by a ferrable power or completely disconnect the load from the
model with up, down, and a derated state. This model is in system.
series with the equivalent models of previous subsystems. Load mode: There may be a load at the tapping station and
The reliability model of the whole sending end segment is the sending end should supply that load as well as the load
developed using the reliability models of the subsystems. This on the receiving end. Calculating the reliability indices for
model includes several derated states which are inserted in different loads in the system depends on the policy of the
1966 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 25, NO. 3, JULY 2010

Fig. 6. Compact reliability model of the whole system.

Fig. 7. Final equivalent reliability model of the whole system.

and the load is not considered further. If the load exceeds

the generation, the remaining load is considered as a load
mode case.

Fig. 5. Comprehensive reliability model of the whole system. D. Reliability Model of the Whole System and the Analysis
Procedure
The reliability models associated with the sending end, the
system to supply the loads (i.e., what priority each load receiving end, and the tapping station are combined to obtain the
has in the system). The location of failure should also be reliability model of the entire system. Fig. 5 shows the detailed
determined for each state. In Fig. 1, failure can happen in reliability model of the HVDC system of Fig. 1. Figs. 6 and 7
the sending end, in the receiving end, or in the tapping illustrate more compact models equivalent to the comprehensive
station. Obviously, different failure locations have different model in Fig. 5. In Fig. 6, the components in each subsystem are
effects on the reliability indices. A failure in the tapping combined and represented by an equivalent reliability model. In
station interrupts the load at its load bus and does not affect Fig. 7, all of the subsystems at the sending or the receiving ends
the load at the receiving end. A failure in the components at are replaced with a single equivalent reliability model. This final
the receiving end only affects the load at the receiving bus. model is used to determine the reliability indices of the system
A failure in the sending end, however, can result in some located at the receiving end.
load curtailment. How much each load (at the receiving end In order to analyze the final reliability model, different
and the tapping station) is curtailed is dependent upon the methods, such as the COPT-based risk model, network reduc-
system operating policy. tion, minimal cut sets, and Monte Carlo simulation could be
Generation and load modes: When the tapping station is used [11]. The COPT-based technique is widely used for the
connected to a bus with load and generation, the problem reliability evaluation of an electric power generating system. An
can be converted to either the load or generation mode. In extension of this technique tailored for interconnected systems
such a case, it is reasonable to assume that the generation is referred to the equivalent assisting unit approach [12]. This
first supplies its load. If there is extra generating capacity, method was adopted to analyze the resulting reliability model.
it is considered as excess generation at the tapping station The following is a brief explanation of the overall procedure
ZADKHAST et al.: RELIABILITY EVALUATION OF AN HVDC TRANSMISSION SYSTEM 1967

and more detail and numerical evidences are contained in TABLE I

Section IV. RELIABILITY MODELS ASSOCIATED WITH SUBSYSTEMS 1 TO 8
The generating system at the sending end in Fig. 1 is consid-
ered as the assisting system and its additional generating poten-
tial is modeled by a multistate capacity assistance table. This
table has different capacity levels and the associated probabil-
ities and departure rates. The obtained COPT model is then
convolved with the reliability model representing the HVDC
sending end segment. The equivalent COPT of the tapping sta-
tion equipment and the associated generating station is then
added to the equivalent COPT of the assisting system along with
the sending end segment. At the next stage, the resulting COPT
is combined with the reliability model of the receiving end seg-
ment. Note that the capacity of dc transmission lines might re-
strict the assistance from the assisting system and the tapping
station. Convolution of the load model with the final COPT pro-
vides the risk model of the HVDC system under study. The re-
quired reliability indices are calculated using the obtained risk
model.

IV. CASE STUDY

The system shown in Fig. 1 is used to illustrate the operation
of an HVDC transmission system with a VSC tapping station. TABLE II
The transmission capacity of the 500 kV transmission system EQUIVALENT RELIABILITY MODEL OF SUBSYSTEMS 2 AND 4
and the generating capacity at the sending end are assumed to be
1200 MW. The generation capacity on the ac side of the tapping
station is 240 MW and the load at the receiving end is assumed
to be constant and equal to 960 MW.
The per-unit system is used in the studies with a base of 1200
MVA. The transmission line capacity is equal to 1 p.u. and the
TABLE III
transmitted power on the line can vary between 0 and 1 p.u. EQUIVALENT RELIABILITY MODEL OF SUBSYSTEMS 2, 3, AND 4
The total generation (1.2 p.u.) is more than the transmission line
capacity (1 p.u.) and the tapping station can assist the system
when there is no congestion on the transmission line. The length
of the transmission line is 500 km and the tapping station is
installed at the middle of the line (i.e., 250 km from each end).
The system load at the receiving end is 0.8 p.u. It is assumed
that the generating system at the sending side is fully reliable.
The system reliability data are provided in Appendix. Different TABLE IV
case studies are presented in the following subsections. RELIABILITY MODEL OF THE SENDING END SEGMENT

A. Base Case
The overall system reliability assessment problem is quan-
tified using the models developed in the previous section. The
analysis procedure is based on the assisting unit approach. In the
first stage, the reliability models of Subsystems 1 to 8 are devel-
oped as shown in Table I. This table shows the probability of
each state and the associated transition rates to the higher and
lower available capacity levels. At the next stage, the models
associated with Subsystems 2 and 4 are combined as shown in
Table II. This model is then convolved with the model of Sub-
system 3 as shown in Table III. The final stage on the modeling
of the sending end segment involves adding the model of Sub- the middle of the transmission system before adding the assis-
system 1 to the equivalent model of Subsystems 2 to 4. The tance of the tapping station.
resulting model which represents the sending end segment is The COPT of the VSC tapping station is given in Table V as-
shown in Table IV. The generation capacity at the sending end suming 0.2 p.u. fully reliable generation capacity at the tapping
is assumed to be equal to 1 p.u. and fully reliable, and therefore station. The next stage involves adding the assistance of the tap-
the model shown in Table IV is the COPT of assisting system at ping station to the COPT of the assisting system (i.e., Table IV).
1968 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 25, NO. 3, JULY 2010

TABLE V TABLE VII

COPT OF THE TAPPING STATION FINAL COPT OF THE SYSTEM

TABLE VI
COPT OF THE ASSISTING SYSTEM AND THE TAPPING
STATION AT THE MIDDLE OF THE TRANSMISSION SYSTEM

TABLE VIII
RELIABILITY INDICES WITHOUT AND WITH THE TAPPING STATION

probability of failure;
The resulting COPT is shown in Table VI. Note that this table frequency of failure (occ/yr);
includes some states with the assistance capacity level of more
than 1 p.u. expected energy not served (in megawatt
The final stage is to combine the COPT shown in Table VI hours per year);
with the reliability model of the receiving end segment. The expected duration of load curtailment (in
line capacity of this segment will restrict the available assistance hours per year);
level.
The same procedure used to that of Subsystems 1 to 4 is con- expected load curtailed (in megawatts per
ducted to equalize the models of Subsystems 5 to 8. Since the year).
structure of the receiving end segment is the same as that of the Table VIII includes the results of the base case. In order to
sending end segment and the reliability data of corresponding illustrate the impact of the VSC tapping station on the load point
components are also identical, the model of the receiving end reliability indices, another case excluding the tapping station
segment is the same as the model presented in Table IV. Adding is examined. The results of this simulation are also presented
this model to the COPT shown in Table VI leads to the final in Table VIII (i.e., the second column of this table). Note that
COPT which is given in Table VII. In this table, the maximum assuming a load mode tapping station and assigning a higher
capacity of the states is capped at 1 p.u. which is the line ca- priority to supply the load at the receiving end, the load at the
pacity. In this table, the frequency of each state is calculated tapping station would be supplied just during the periods that the
as follows: generating and transmission capacities exceed the receiving end
load. Accordingly, the existence of the tapping station would
(1) have no effect on the reliability indices of the receiving end and
the numerical results would be the same as the case excluding
where is the probability of occurrence of state , and the tapping station.
and are the upward and downward capacity departure rates, The last column of Table VIII shows the percentage improve-
respectively. ment in the reliability indices by adding the tapping station to
The reliability indices for the load point at the receiving end the HVDC system. It can be observed that all indices are im-
are calculated by adding the load model to the COPT shown in proved. There is an 11.7% decrease in and the when
Table VII. The resulting model is referred to as the system risk the tapping station is added. Improvements in the indices related
model. The following reliability indices are calculated: to the load and energy are more significant. There is more than a
ZADKHAST et al.: RELIABILITY EVALUATION OF AN HVDC TRANSMISSION SYSTEM 1969

TABLE IX
EFFECT OF LOAD LEVEL ON SYSTEM INDICES

Fig. 8. Q and EENS versus load variation.

34% decrease in the and the with the installation

of a tapping station in the middle of the transmission line. The
different reliability indices, value of lost load , and the
associated investment costs can be used to decide whether the
installation of the tapping station is beneficial or not, or what ca-
pacity of the tapping station would be the optimum level from a
cost/benefit point of view. In the base case, the assumed auxil-
iary generation at the middle station is much smaller than the
main generation at the sending end, which is more likely to
occur in practice. Auxiliary generation with a large capacity will
have more influence on the system reliability.

B. Effect of Load Level

In the base case study, the system load level is assumed to
be equal to 0.8 p.u. Here, this parameter is varied from 0.3 to
Fig. 9. Variation in EENS versus load level and variation in the installation
0.95 p.u in steps. The resulting system indices are presented in location of the tapping station.
Table IX. As expected, all the indices deteriorated as the load
level is increased. Fig. 8 shows the changes in and
versus the variation of the load level. As it can be seen, there are and it does not change with the location of the tapping station.
two sudden jumps in at load levels of 0.5 and 0.7 p.u. This At load levels greater than 0.5 p.u., the reliability index becomes
observation can be justified by referring to the probability values sensitive to the location of the tapping station. The de-
in the final COPT model of the system presented in Table VII. It creases as the location of the tapping station moves closer to the
can be seen that the probability associated with the state of 0.5 receiving end. At the load level of 0.6 p.u., the value of
p.u. available capacity is much higher than that of the state with increases more than 150% as the location of the tapping station
lower capacity (i.e., state of 0.3 p.u. available capacity). Such is changed from 400 to 100 km from the sending end. This in-
a justification could be raised for the state of 0.7 p.u. capacity crease is about 27% at the load level of 0.95 p.u.
level. Based on this behavior of index , one can conclude that
either load level of 0.5 or 0.7 p.u. could be appropriate operating V. CONCLUSION
points of the system since higher load levels result in a rapid The paper presented an evaluation methodology for the reli-
increase in ; but, lower ones do not significantly improve the ability of an HVDC transmission system with a VSC tapping
reliability. Fig. 8 also reveals that the EENS does not have such station. The operation of a VSC converter has no commutation
sudden jumps as and its variation is much smoother than that failure problems compared with line commutated converters
of . which can experience commutation failure when connected to
a weak ac system. The effects of a VSC tapping station on the
C. Effect of Location of the Tapping Station reliability indices were presented. A comprehensive reliability
It is expected that the installation location of the tapping sta- model associated with an HVDC system with a tapping station
tion will affect the reliability indices at the load point. Changing was developed taking into account the impacts of different
the location of the tapping station is modeled via the variation of system components. The proposed model was simplified by
failure rate values associated with DCTL in Subsystems 4 and using basic reliability engineering concepts. The results of the
5. The simulations in this section are conducted at different load numerical analyses show that a tapping station can significantly
levels considering different installation locations for the tapping improve the reliability indices at the load point. The indices
station. Fig. 9 presents the obtained results for the index related to the energy and power ( and ) are af-
pictorially. The distances indicated in this figure are between the fected more than the other indices related to the probability of
installation location and the sending end. Fig. 9 shows that at failure. Simulations at different load levels show that has
load levels less than 0.5 p.u., the is approximately zero nonsmooth variations but increases constantly with the
1970 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 25, NO. 3, JULY 2010

TABLE X [2] P. R. S. Kuruganty and D. A. Woodford, “A reliability cost-benefit

RELIABILITY DATA OF COMPONENTS analysis for HVDC transmission expansion planning,” IEEE Trans.
Power Del., vol. 3, no. 3, pp. 1241–1248, Jul. 1988.
[3] E. N. Dialynas and N. G. Koskolos, “Reliability modeling and evalua-
tion of HVDC power transmission systems,” IEEE Trans. Power Del.,
vol. 9, no. 2, pp. 872–878, Apr. 1994.
[4] S. Kuruganty, “Effect of HVDC component enhancement on the overall
system reliability enhancement,” IEEE Trans. Power Del., vol. 9, no.
1, pp. 343–349, Jan. 1994.
[5] S. Kuruganty, “Comparison of reliability performance of group con-
nected and conventional HVDC transmission systems,” IEEE Trans.
Power Del., vol. 10, no. 4, pp. 1889–1895, Oct. 1995.
[6] R. Billinton, S. Aboreshaid, and M. Fotuhi-Firuzabad, “Well-being
analysis for HVDC transmission system,” IEEE Trans. Power Syst.,
vol. 12, no. 2, pp. 913–918, May 1997.
[7] R. Billinton, M. Fotuhi-Firuzabad, and S. Aboreshaid, “Diagnosing
the health of bulk generation and HVDC transmission system,” IEEE
Trans. Power Syst., vol. 12, no. 4, pp. 1740–1745, Nov. 1997.
[8] R. Billinton, M. Fotuhi-Firuzabad, S. O. Faried, and S. Aboreshaid,
“Composite system reliability evaluation incorporating an HVDC link
and a static synchronous series compensators,” in Proc. IEEE Can.
Conf. Elec. Comp. Engin., May 2002, vol. 1, pp. 42–47.
[9] R. Billinton, M. Fotuhi-Firuzabad, and S. O. Faried, “Reliability eval-
uation of hybrid multi-terminal HVDC subtransmission systems,” in
Proc. Inst. Elect. Eng., Gen. Transm. Distrib., Sep. 2002, vol. 149, pp.
571–577.
[10] G. Asplund and A. Williamson, “A novel approach to providing on
TABLE XI route power supplies to rural and urban communities in close proximity
DERATED STATES FOR ACFS AND CAPS to the extra high voltage DC transmission line,” in Proc. IEEE Power
Eng. Soc. Power Africa Conf. Expo., Jul. 2007, pp. 1–6.
[11] R. Billinton and R. N. Allan, Reliability Evaluation of Engineering
Systems, 2nd ed. , New York: Plenum, 1994.
[12] R. Billinton and R. N. Allan, Reliability Evaluation of Power Systems,
2nd ed. , New York: Plenum, 1996.
[13] M. Fotuhi-Firuzabad, R. Billinton, and S. O. Faried, “Incorporating a
DC link in composite system reliability evaluation,” in Proc. Power
Eng. Soc. Winter Meeting, Jan. 2000, vol. 3, pp. 1760–1765.
[14] R. Billinton, M. Fotuhi-Firuzabad, and S. O. Faried, “Adequacy evalua-
tion of interconnected systems with an AC-DC link using an equivalent
approach,” in Proc. Can. Conf. Elect. Comput. Eng., Jan. 2000, vol. 3,
pp. 1760–1765.
TABLE XII
DERATED STATES FOR DCFS Sajjad Zadkhast is currently pursuing the Ph.D. degree at Sharif University of
Technology, Tehran, Iran.

Mahmud Fotuhi-Firuzabad is a Professor and Head of the Department of

Electrical Engineering, Sharif University of Technology, Tehran, Iran.

increase of load level. The studies show that the effect of the
tapping station location on the reliability indices plays a more Farrokh Aminifar (S’07) is currently pursuing the Ph.D. degree at Sharif Uni-
important role at load levels above 0.5 p.u. versity of Technology, Tehran, Iran.

APPENDIX
INPUT DATA FOR THE SYSTEM UNDER STUDY Roy Billinton (S’59–M’64–SM’73–F’78–LF’01) is an Emeritus Professor in
the Department of Electrical and Computer Engineering at the University of
Table X shows the failure rates, repair times, unavailabili- Saskatchewan, where he has been since 1964.
ties, and installation times for the system components. Table XI
shows the effect of outages in ACFs and caps on the maximum
transferrable power. Table XII presents the capacity table for Sherif Omar Faried (SM’00) is a Professor of Electrical Engineering in the
DCFs. It should be noted that the data used for simulations lie Department of Electrical and Computer Engineering, Power Systems Research
within reasonable ranges [6], [9], [13], [14]. Group, University of Saskatchewan, Saskatoon, SK, Canada.

REFERENCES
[1] R. Billinton and M. S. Sachdev, “Direct current transmission system Abdel-Aty (Aty) Edris (SM’88) is a Principal Consultant with Siemens Energy
reliability evaluation,” Trans. Can. Elect. Assoc., vol. 7, no. 3, 1968. Co., Mountain View, CA