IEEE TRANSACTIONS ON POWER SYSTEMS, VOL. 23, NO.

1, FEBRUARY 2008 47

Introduction to Multidimensional
Financial Transmission Rights
V. Sarkar and S. A. Khaparde, Senior Member, IEEE

Abstract—This paper introduces the concept of multidi- th element of .

mensional financial transmission right (MDFTR) which is a
Number of all possible active-inactive combinations
generalized group financial transmission right (FTR) concept.
An MDFTR allows its owner to hold a total FTR amount over a of option FTRs.
path group along with multiple choices to distribute this grouped Dual variable (i.e., Lagrangian multiplier) cor-
amount among the individual grouping paths. This added flexi- responding to the th simultaneous feasibility
bility provides forward-market participants with a useful support constraint.
to overcome the congestion price risk that originates from the
uncertainty in their transaction paths. All the fundamental issues Coefficient of the th multidimensional FTR term in
regarding the implementation of MDFTRs are addressed in detail the th simultaneous feasibility constraint.
in this paper. The practical utility of MDFTRs is judged by Column vector containing the MW amounts of the
analyzing their costs. It is shown with illustration that an MDFTR base nongenerated option FTRs.
may be more cost-effective than any given portfolio of individual
FTRs that has the ability to give the same risk-hedging benefit. Column vector, containing the MW amounts of the
base generated option FTRs, corresponding to the
Index Terms—Auction, financial transmission right, locational
marginal price, multidimensional financial transmission right,
th combination of MW-path distribution vectors
obligation, option, secondary trading. among the base and requested MDFTRs.
Column vector containing the MW amounts of the
base option MDFTRs.
NOMENCLATURE: Column vector, containing the MW amounts of the
Set of MW-path distribution vectors for the th mul- base generated obligation FTRs, corresponding to
tidimensional FTR. the th combination of MW-path distribution vec-
Price of the th multidimensional FTR. tors among the base and requested MDFTRs.
Column vector containing the MW amounts of the
Total number of simultaneous feasibility constraints. base obligation MDFTRs.
Path-branch sensitivity factor (or power transfer dis- Number of elements in .
tribution factor), relating a certain line to the path of
Number of elements in .
the th requested nongenerated obligation FTR, for
a particular network topology. MW amount of the th multidimensional FTR.
Path-branch sensitivity factor, relating a certain line Path-branch sensitivity factor, relating a certain line
to the path of the th requested generated obligation to the th path of the th multidimensional FTR, for
FTR, for a particular network topology. a particular network topology.
Path-branch sensitivity factor, relating a certain line Set of paths over which the th multidimensional
to the path of the th base generated obligation FTR, FTR is defined.
for a particular network topology. Total number of MW-path distribution vector com-
Path-branch sensitivity factor, relating a certain line binations among the MDFTRs.
to the path of the th nongenerated option FTR, for Column vector containing the variable terms that
a particular network topology. signify the awarded amounts towards the nongener-
Path-branch sensitivity factor, relating a certain line ated obligation FTR requests.
to the path of the th generated option FTR, for a Column vector, containing the variable terms that
particular network topology. signify the awarded amounts towards the gener-
Flow caused by the physical equivalents of base non- ated obligation FTR requests, corresponding to the
generated obligation FTRs on a certain line under a th combination of MW-path distribution vectors
particular network topology. among the base and requested MDFTRs.
Column vector containing the variable terms that
signify the awarded amounts towards the obligation
Manuscript received August 18, 2006; revised May 29, 2007. Paper no.
TPWRS-00539-2006. MDFTR requests.
The authors are with Department of Electrical Engineering, Indian Institute Column vector containing the variable terms that
of Technology Bombay, Mumbai 400076, India (email: [email protected].
in;[email protected]).
signify the awarded amounts towards the nongener-
Digital Object Identifier 10.1109/TPWRS.2007.913199 ated option FTR requests.

0885-8950/$25.00 © 2008 IEEE

Authorized licensed use limited to: GOVERNMENT COLLEGE OF ENGINEERING. Downloaded on January 15, 2010 at 03:11 from IEEE Xplore. Restrictions apply.
 48 IEEE TRANSACTIONS ON POWER SYSTEMS, VOL. 23, NO. 1, FEBRUARY 2008

Column vector, containing the variable terms that The concept of FGRs was developed basically to enhance the
signify the awarded amounts towards the generated decentralized trading of transmission rights [11]. However, as
option FTR requests, corresponding to the th com- discussed in [12], the FGR approach may encounter difficulty
bination of MW-path distribution vectors among the due to the following:
base and requested MDFTRs. 1) the presence of a significant number of contingent
Column vector containing the variable terms that flowgates;
signify the awarded amounts towards the option 2) the varying values of power transfer distribution factors.
MDFTR requests. As a result, the applicability of FGRs is limited to those systems
Flow limit of a certain line. where the above two problems are not too severe. By compar-
Total number of requested and base multidimen- ison, the FTR mechanism is much less affected by these prob-
sional FTRs. lems. Finally, it should be mentioned again that many of the
th element of . successfully running LMP-based power markets have found the
Zero vector of dimension 1. FTR mechanism to be more suitable for their systems as com-
Number of rows in the column vector (.) pared to the FGR mechanism.
In this paper, we introduce the concept of a multidimensional
Sum of all elements of the column vector (.).
financial transmission right (MDFTR). With the addition of
MDFTRs, forward-market participants get a useful support
for overcoming the congestion price risk that is caused by the
I. INTRODUCTION
uncertainty in their transaction paths. However, in a true sense,
a multidimensional FTR is not in itself a transmission right;
INANCIAL transmission rights (FTRs) are effective risk-
F hedging tools, designed with an aim to minimize conges-
tion price risk for the forward contracts under locational mar-
rather, it is a generator of transmission rights. An MDFTR
owner basically holds a group FTR over a set of paths, along
with multiple choices to distribute this grouped amount among
ginal pricing environment. This concept was first introduced by
the individual grouping paths. Therefore, he can choose from
Hogan in 1992 [1], [2]. FTRs are now successfully implemented
among multiple hedge alternatives after watching the actual
in many power markets like PJM, New England, New York, and
delivery pattern of his transaction.
others [3]–[7].
A point-to-point FTR has multiple specifications. The basic An elementary discussion on such group FTR concept was
parameters defining an FTR are: already presented in [13] in the form of a contingent transmis-
1) a source and a sink; sion right. Moreover, in [14], a contingent transmission right
2) a validity period; was recognized as a future FTR product. However, in this work,
3) a MW amount. we have generalized the idea of a contingent transmission right
Each FTR is assigned a monetary value for each hour depending into the complete form of a multidimensional financial trans-
upon the day-ahead locational marginal price (LMP) outcomes mission right. The purpose behind this generalization is to build
for that hour. The FTR owners are paid according to the hourly a complete framework for the implementation of group FTRs so
values of their FTRs. However, the settlement can be over mul- as to enhance the current FTR mechanism. We have elaborately
tiple hours at a time. discussed several issues that have not been addressed in depth
With regard to hedging adjustability, all the currently exer- in the previous literature. The main contributions of this paper
cised FTRs can be classified into two categories: options and are:
obligations. They differ in the sense that an obligation FTR may 1) general parameterization of group FTRs;
incur a negative value at certain hours, whereas the value of an 2) compact formulation for the clearing and pricing of group
option FTR is always nonnegative. When awarded simultane- FTRs;
ously, the price of an option FTR will always be greater than or 3) secondary trading rules for group FTRs;
equal to that of an obligation FTR on the same path. Note that, 4) permissible adjustments in the patterns of group FTRs;
in FTR context, a path is defined simply by a source-sink pair
5) assessment of the practical utility of group FTRs (by cost
(for example, from Node 1 to Node 2) [5] rather than an alter-
analysis).
nating series of nodes and lines. A generalized formulation for
The rest of this paper is organized as follows: Section II de-
the auction of obligation FTRs can be found in [8], whereas [9]
and [10] explain how to modify this formulation when option scribes the principle of multidimensional FTRs. Formulations
FTRs are also included in the auction. regarding the auction clearing and pricing of MDFTRs are dis-
An alternative to the financial transmission right mechanism cussed in Section III. Section IV presents an illustrative example
is the flowgate right (FGR) mechanism. Each FGR is defined on to help understand the auction formulation. The cost-effective-
a commercially significant flowgate in a specific direction. By ness of MDFTRs is illustrated in Section V. In Section VI, rules
definition, each FGR is an option. The hourly value of an FGR is for the secondary trading of MDFTRs are elaborately discussed.
determined by the shadow price of the relevant flowgate and the Facilities available for the contraction and expansion of the dis-
direction of congestion. A point-to-point physical transaction tribution vector set of an MDFTR are explained in Section VII.
can be fully hedged by a certain portfolio of FGRs. Finally, this paper is concluded in Section VIII.

Authorized licensed use limited to: GOVERNMENT COLLEGE OF ENGINEERING. Downloaded on January 15, 2010 at 03:11 from IEEE Xplore. Restrictions apply.
SARKAR AND KHAPARDE: INTRODUCTION TO MULTIDIMENSIONAL FINANCIAL TRANSMISSION RIGHTS 49

II. PRINCIPLE OF MULTIDIMENSIONAL FTRS III. AUCTION MODEL WITH MULTIDIMENSIONAL FTRS

A. Structure and Rule Multidimensional financial transmission rights can be is-

sued through FTR auctions. The objective of an auction
Structurally, a multidimensional FTR is defined by three
problem is to minimize the negated value of the quote-based
parameters:
sum of the cleared amounts towards the FTR and MDFTR
1) a MW amount ;
bids. The cleared amount towards each bid must lie within
2) a set of paths ;
its lower (which is zero) and upper (which is the requested
3) a set of MW-path distribution vectors (DVs)
amount) limits. To ensure revenue adequacy, simultaneous
where
feasibility condition must be satisfied while issuing the FTRs
and MDFTRs. The statement of simultaneous feasibility con-
dition, when only individual FTRs are involved, can be found
in [10]. However, this statement has to be slightly modified
for to to take multidimensional FTRs into account. This is because
a multidimensional FTR is essentially an FTR generator that
for to
can be broken down into multiple combinations of individual
for to FTRs depending upon its DV set. Therefore, the modified
simultaneous feasibility condition can be stated as: For each
The owner of the MDFTR can choose any distribution vector distribution vector combination among the MDFTRs, the phys-
from the DV set at any hour. Upon the selection of distribu- ical equivalent of each possible active-inactive combination of
tion vector , the MDFTR is divided into individual FTRs individual FTRs must result in a feasible power flow condition
over its constituent paths with MW on path . A multi- under each possible topological scenario of the network. Here,
dimensional FTR can be an obligation or an option. If it is an for the sake of simplicity, we are assuming that all the FTRs
obligation, each of the individual FTRs generated from it would and MDFTRs are being issued for a same set of hours. Note
be considered as an obligation FTR. Similarly, if it is an option, that we have to consider different active-inactive combinations
each of the individual FTRs generated from it would be consid- of individual FTRs. This is because an option FTR becomes
ered as an option FTR. inactive in case congestion occurs in the reverse direction.
Similarly, different topological conditions of the network have
B. How it Works to be considered in order to take “ ” contingencies into
The basic functionality of MDFTRs is simple. For example, account.
a market participant has two generating plants at two different
locations, say, at Locations A and B. Each plant consists of only
one unit, and the capacity of each unit is 50 MW. This market A. Formulation of Simultaneous Feasibility Constraints
player has signed a one-year bilateral contract for 40 MW with
a load-serving entity at Location C. Now, for maintenance, Unit As mentioned earlier, a multidimensional FTR has to be
A will be out of service for a certain time period; similarly, Unit broken down into individual FTRs (by selecting a suitable
B will be out of service for another time period. Under such distribution vector) before cashing it. Therefore, all the FTRs
a situation, this market player can purchase an MDFTR of 40 that the market participants hold for a certain hour can be
MW on Paths A-C and B-C with distribution vectors categorized into two groups.
and . In case power is supplied from Unit B, the player Generated: An individual FTR is said to be generated if it
can choose the first distribution vector to hedge this transac- is originated from a multidimensional FTR.
tion. Similarly, if power is supplied from Unit A, the player can Nongenerated: An individual FTR is said to be nongener-
choose the second distribution vector to hedge the transaction. ated if it is not originated from a multidimensional FTR.
However, if the player is a speculator of locational marginal However, any combination of distribution vectors is possible
prices, he will always go for that distribution vector which is among the MDFTRs. Now, for the th combination of distribu-
the most valuable one for the estimated LMP values. tion vectors, the flow limit constraints for a particular line under
Now, as our proposal suggests, the DV of an MDFTR must a certain network topology can be written as follows:
be frozen by the owner himself (for each hour of a day) before
the day-ahead market begins. In an alternative approach, it may
be the ISO who would decide the DV according to which one
is most valuable for the calculated LMP values. Both the ap-
proaches are fundamentally similar in the sense that the MDFTR
can take any distribution vector at any hour. Only, the entity that
makes the decision for the selection of distribution vector and (1)
the specific time (before or after LMP calculation) when this
decision is made are different in these two approaches. Which
of these two policies is a better one may be a matter for further
research. for to

Authorized licensed use limited to: GOVERNMENT COLLEGE OF ENGINEERING. Downloaded on January 15, 2010 at 03:11 from IEEE Xplore. Restrictions apply.
 50 IEEE TRANSACTIONS ON POWER SYSTEMS, VOL. 23, NO. 1, FEBRUARY 2008

where Now, if the th MDFTR has its th distribution vector in the th

DV combination

for to if MDFTR is an option

if MDFTR is an obligation

if MDFTR is an option

For the th active-inactive combination of option FTRs

if MDFTR is an obligation
if nongenerated option FTR is active
if nongenerated option FTR is inactive Similarly, for each of the remaining DV combinations, a sep-
arate constraint pair has to be formed. Therefore, at this stage,
if generated option FTR is active
there are in total constraints for this line/topology pair. Now,
if generated option FTR is inactive as any combination of distribution vectors is possible among the
multidimensional FTRs, both and exist. Hence, by
Note that, here, we are using the dc power flow model, which is applying directional breaking theorem 1, these constraints
rather the more commonly employed power flow model. Now, can finally be reduced to the following couple of constraints:
after directional breaking [10] (directional breaking theorem 1
is also described in Appendix I), the constraint set (1) can be
reduced to the following two constraints:

(4)

(2)

or

where

(5)

After expressing each generated FTR in terms of its originator

MDFTR, the constraint set (2) can be written as follows:
where

(3)

The above formulation suggests that there is no difficulty in

where dealing with MDFTRs in an auction. The auction is still a linear
programming problem. Furthermore, MDFTRs do not create
any overburden of an excessive number of constraints. Even
with the inclusion of MDFTRs, the auction problem has only
two constraints per line per network topology. Finally, it should

Authorized licensed use limited to: GOVERNMENT COLLEGE OF ENGINEERING. Downloaded on January 15, 2010 at 03:11 from IEEE Xplore. Restrictions apply.
SARKAR AND KHAPARDE: INTRODUCTION TO MULTIDIMENSIONAL FINANCIAL TRANSMISSION RIGHTS 51

TABLE III
SENSITIVITIES OF THE FLOW ON LINE 1–2 TO THE RELEVANT PATHS

the step-by-step procedure for formulating the simultaneous

feasibility constraints for Line 1–2 under the base network
topology. Line 1–2 has a power carrying capacity of 500 MW.
Fig. 1. Sample four-bus network. For the sake of simplicity, it is assumed that there is no base
MDFTR or nongenerated FTR. The path-branch sensitivity
TABLE I factors between Line 1–2 and the relevant paths are shown in
PATHS AND TYPES OF INDIVIDUAL FTR REQUESTS Table III.
Here, we assume that both the MDFTR bidders are
requesting for the same set of distribution vectors,
. Therefore, four combinations of
distribution vectors are possible. These combinations are
TABLE II
PATHS AND TYPES OF MDFTR REQUESTS
Combination

Combination

Combination

Combination
be mentioned that the above formulation of simultaneous fea- Now, for the first combination of distribution vectors, the actual
sibility constraints is irrespective of any particular settlement set of simultaneous feasibility constraints can be written as
policy (the different settlement policies have been discussed in
Section II-B) as both the policies are fundamentally similar.

B. Auction Pricing
Multidimensional FTRs are to be priced according to the (6)
same marginal pricing rule [5], [9], [13] as followed for the
pricing of individual FTRs. Therefore, the quantities to be for to
considered for pricing an MDFTR are:
1) shadow prices of the simultaneous feasibility constraints; where
2) impacts of the MDFTR on simultaneous feasibility con- ; ; ;
straints, i.e., sensitivity factors. ; ; ;
The price of the th MDFTR can be compactly written as ; ;
; ; ;
.
Here, the variable term corresponds to the MW amount
of the th generated FTR from the th MDFTR for the th
combination of distribution vectors. After directional breaking,
The necessary form of the Lagrangian can be found in [15] and
the constraint set (6) can be reduced to the following two
[16].
constraints:
IV. NUMERICAL EXAMPLE
In this section, we will illustrate the auction formulation
process with the help of a simple example. Fig. 1 shows the
sample power system where the line impedances are shown (7)
in per unit. The paths and types of individual FTR requests
are listed in Table I. Table II lists the types and paths of the
MDFTR requests. The fourth columns of Tables I and II con- where
tain the names of the variables that correspond to the awarded ; ; ;
amounts towards the FTR and MDFTR requests, respectively. .
The bracketed numbers in the second column of Table II are After replacing the generated FTR terms by the respective
path index numbers within the MDFTRs. We will demonstrate MDFTR terms (i.e., , ,

Authorized licensed use limited to: GOVERNMENT COLLEGE OF ENGINEERING. Downloaded on January 15, 2010 at 03:11 from IEEE Xplore. Restrictions apply.
 52 IEEE TRANSACTIONS ON POWER SYSTEMS, VOL. 23, NO. 1, FEBRUARY 2008

, ), the constraint set (7) can be reformed TABLE IV

as TEST MDFTRS

(8)

where
; ; ;
.
TABLE V
Similarly, for the remaining DV combinations, there are six REFERENCE FTR BIDS
constraints with the following coefficient vectors:
; ;
; ;
; .
Therefore, at this stage, the number of constraints is eight.
After directional breaking, these eight constraints can be re-
duced to a single pair of constraints as shown in the following:

(9)
TABLE VI
REFERENCE MDFTR BIDS

where
& .
In the above example, we have numerically illustrated all
the actual steps that are involved in the formulation of simul-
taneous feasibility constraints for a certain line under a partic-
ular network topology. Finally, there are only two constraints different paths, say, Path-1 and Path-2, with distribution vec-
per line per network topology. However, for the practical im- tors and . If this MDFTR is an obliga-
plementation of such an auction, all the intermediate steps de- tion, its cheapest alternative is a portfolio of 50 MW obligation
scribed in the above example are redundant. The values of the and 50 MW option FTRs on Path-1 and 50 MW option FTR on
coefficients in the final set of constraints can be directly cal- Path-2. Similarly, if this MDFTR is an option, its cheapest al-
culated by means of the generalized formulation presented in ternative is a portfolio of a 100 MW option FTR on Path-1 and
Section III. In addition, no major modification in formulation a 50 MW option FTR on Path-2. However, procurement of an
is required for including the base FTR or MDFTR terms. Base MDFTR may be a less costly affair than the procurement of its
FTRs and MDFTRs are those entities whose quantity values are alternative. In this section, this particular fact will be illustrated
considered as constants during the auction calculation. As an numerically.
example, the FTR awarded to a market participant in an annual In this study, the IEEE standard 30-bus system is adopted
auction is to be considered as a base FTR in the subsequent as the sample power system. The line capacities are those as
monthly auctions. Similarly, a self-scheduled (self-scheduling given in File “case30.m” in the MATPOWER software package
means showing insensitivity to price) FTR or MDFTR is a base [17]. Only base network constraints are considered for testing
entity in an auction. FTR and MDFTR surrenders are also to simultaneous feasibility (i.e., no contingency case is consid-
be considered through base case modeling. However, base case ered). Table IV presents a list of ten test MDFTRs. The option
entities can be accounted for simply by representing them as version cost as well as the obligation version cost of each of
constant MW values rather than as variables. these test MDFTRs is to be calculated with reference to the FTR
and MDFTR bids shown in Tables V and VI, respectively. The
V. COST-EFFECTIVENESS OF MDFTRS acronyms “Ob” and “Op” in tables stand for obligation and op-
The job of MDFTRs is to alleviate the congestion price risk tion, respectively. To calculate the cost of a test MDFTR (in any
associated with the uncertainty in the delivery patterns of phys- version), this MDFTR is considered as self-scheduled and an
ical transactions. However, MDFTRs are not the only option auction is conducted with the reference FTR and MDFTR bids.
to fulfill this requirement. The same risk-hedging benefit can The costs of the alternatives of these MDFTRs are calculated
alternatively be obtained from a portfolio of individual FTRs. in a similar way. The DV set of each (test as well as reference)
For example, consider an MDFTR of 100 MW, grouped on two MDFTR is taken as .

Authorized licensed use limited to: GOVERNMENT COLLEGE OF ENGINEERING. Downloaded on January 15, 2010 at 03:11 from IEEE Xplore. Restrictions apply.
SARKAR AND KHAPARDE: INTRODUCTION TO MULTIDIMENSIONAL FINANCIAL TRANSMISSION RIGHTS 53

TABLE VII MDFTRs or individual FTRs that will load some simultaneous
COSTS OF THE TEST MDFTRS feasibility constraints more than the original MDFTR does.
This is the same requirement as for the secondary trading of
individual FTRs. In case an MDFTR is to be traded in Mode
1, its distribution vector must be frozen at the very beginning
in order to permanently fix the FTR amount on each of its
constituent paths. Next, each of these individual FTRs is to be
traded separately while obeying the following set of rules.
1) Rule 1: Option and obligation FTRs must be traded as op-
tions and obligations, respectively.
2) Rule 2: Each of the derived FTRs must be defined on the
same path as that of the original FTR.
3) Rule 3: The validity periods of any two derived FTRs must
TABLE VIII not overlap partially.
COSTS OF THE ALTERNATIVE BUNDLES OF INDIVIDUAL FTRS
4) Rule 4: The sum of the MW amounts of all the derived
FTRs with same validity period must be equal to the MW
amount of the original FTR.
5) Rule 5: The validity period of any derived FTR must not
exceed the validity period of the original FTR.
However, the rules for the secondary trading of individual
FTRs are not sufficient for a valid Mode 2 secondary trading
of an MDFTR. The complete set of rules that must be obeyed
while trading an MDFTR in Mode 2 is described below.
1) Rule 1: Option and obligation MDFTRs must be traded as
options and obligations, respectively.
2) Rule 2: Each of the derived MDFTRs must be defined on
Table VII presents the auction outcomes for the costs of the the same path group as that of the original MDFTR.
test MDFTRs. The auction outcomes for the costs of the alter- 3) Rule 3: The DV set of each derived MDFTR must be a
natives of these test MDFTRs are presented in Table VIII. A subset of the DV set of original MDFTR.
comparison between Column 2 of Table VII and Column 2 of 4) Rule 4: The validity periods of any two derived MDFTRs
Table VIII reveals that the obligation version cost of each of the must not overlap partially.
test MDFTRs is less than the cost of the corresponding alterna- 5) Rule 5: The sum of the MW amounts of all the derived
tive. Similarly, the comparison between Column 3 of Table VII MDFTRs with same validity period must be equal to the
and Column 3 of Table VIII reveals that the option version cost MW amount of the original MDFTR.
of each of these MDFTRs is also either less than or equal to that 6) Rule 6: The validity period of any derived MDFTR must
of the corresponding alternative. not exceed the validity period of the original MDFTR.
It is clear from these results that an MDFTR may be more It can be observed that the secondary trading of an MDFTR
cost effective than its alternative. In fact, we studied many other following the above set of rules may give birth to some resul-
test auctions. In each case, the relevant MDFTR appeared to tant distribution vectors (with reference to the original MDFTR)
be cheaper or at most equally costly. It can be proven that (see that are not members of the DV set of the original MDFTR. For
Appendix B) the loading effect of an MDFTR on any simulta- example, suppose an MDFTR with the structure defined by the
neous feasibility constraint can never be higher than that of its triplet is broken into two MDFTRs, and
alternative. We can correlate the above results with this physical . Now, if two different distribution vectors, and
reality. This is similar to a very common phenomenon that an , are chosen for the above two derived MDFTRs, the resul-
obligation FTR can be no less cost effective than an option FTR tant distribution vector (i.e., ) may not be a
of the same MW on the same path. Finally, it should be men- member of . However, it can be easily verified from the discus-
tioned that the viability of the alternative of an MDFTR relies sion in Section VII-B that this kind of phenomena never creates
upon the availability of option rights in the market. an extra load on any simultaneous feasibility constraint.

VI. SECONDARY TRADING OF MDFTRS VII. ADJUSTMENT OF THE DV SET

Multidimensional FTRs are tradable in the secondary market. In this section, we will discuss two different facilities that are
An MDFTR can be traded in any of the two modes. available for adjusting the DV set of an MDFTR.
1) Mode 1: In this mode, a portfolio of individual FTRs is
derived from the original MDFTR. A. Surrendering a Portion of the Current DV Set
2) Mode 2: In this mode, a portfolio of MDFTRs is derived At a certain point of time, the owner of an MDFTR may find
from the original MDFTR. the DV set of his MDFTR to be excessive. In such a situation, he
Whatever mode is adopted for its secondary trading, the can surrender the unnecessary portion of the DV set in the sub-
MDFTR cannot be converted into any such collection of sequent auction(s) (within the validity period of the MDFTR).

Authorized licensed use limited to: GOVERNMENT COLLEGE OF ENGINEERING. Downloaded on January 15, 2010 at 03:11 from IEEE Xplore. Restrictions apply.
 54 IEEE TRANSACTIONS ON POWER SYSTEMS, VOL. 23, NO. 1, FEBRUARY 2008

Simultaneous feasibility does not get threatened due to this kind APPENDIX A
of activity. By surrendering a portion of the DV set, a player in DIRECTIONAL BREAKING THEOREM 1
effect sells his original MDFTR and, at the same time, buys a
Theorem 1: Let an optimization problem with variable vec-
new MDFTR of same MW amount as that of the original one,
tors and be subjected to number of
on the same path group, but with a reduced set of distribution
constraints of the following form:
vectors. The player always receives a net nonnegative payment
which is the difference between the amount of money that he
earns by selling the original MDFTR and the amount that he
has to spend for buying the new MDFTR.
(10)
B. Expanding the Current DV Set
Let the current DV set of an MDFTR be expanded for to
to a new DV set , where
The solution space provided by constraint set (10) is the same
as that provided by the following constraint set:

(11)

where
i.e., each of the new distribution vectors is a convex combina-
tion of the original distribution vectors. It can be shown (see
Appendix C for proof) that the loading effects of this new DV and
set are the same as those of the original DV set. This particular
characteristic of MDFTRs provides an owner with some oppor-
tunity to expand the DV set of his MDFTR without making any
additional payment. He is allowed to request ISO at any point
of time for such an expansion. Constant column vectors
The concept of DV set expansion has physical significance in Constant value
two ways. In case an MDFTR owner has to conduct his trans- Scalar functions of and
action with such a pattern that is not pre-decided, he may re-
quire a new distribution vector to hedge this transaction. If the
MDFTR owner is not a speculator of LMP values, the DV set
expansion facility mentioned above may be useful in fulfilling
his additional DV requirement. Second, an MDFTR owner may
need a new DV for trading his MDFTR in Mode 1. The DV Proof: We will first prove that any point in the solution
set expansion facility may also be useful under such a situation. space provided by constraint set (10) also lies in the solution
Moreover, Mode 2 secondary trading of MDFTRs is implicitly space provided by constraint set (11). Next, we will prove that
based on this particular concept of DV set expansion. any point in the solution space provided by constraint set (11)
also lies in the solution space provided by constraint set (10).
Part 1: Let be a point in the solution space pro-
VIII. CONCLUSION
vided by constraint set (10)
MDFTRs are FTR generators that provide forward-market
participants with a useful support to overcome the congestion
price risk that arises due to the uncertainty in the delivery pat-
terns of their physical transactions. An MDFTR owner has the
flexibility to distribute a grouped FTR amount in multiple ways
over the individual grouping paths. An MDFTR may be either
an option or an obligation. An obligation MDFTR generates
obligation rights. Similarly, an option MDFTR generates op-
tion rights. In this paper, we have presented all the fundamental
details regarding the structure, working, issuance, and pricing
of MDFTRs. The practical utility of MDFTRs has been eval-
uated by cost analysis. Other issues like secondary trading of as
MDFTRs and adjustment of the DV set of an MDFTR have also
been elaborately discussed. The proof of the first part is complete.

Authorized licensed use limited to: GOVERNMENT COLLEGE OF ENGINEERING. Downloaded on January 15, 2010 at 03:11 from IEEE Xplore. Restrictions apply.
SARKAR AND KHAPARDE: INTRODUCTION TO MULTIDIMENSIONAL FINANCIAL TRANSMISSION RIGHTS 55

Part 2: Let be a point in the solution space pro- where

vided by constraint set (11)

However, as , ,
The loading effects of the alternative of this obligation MDFTR
on those forward and reverse flow limit constraints ( and
, respectively) can be written as

and

where

The proof of the second part is complete.

However, the applicability of directional breaking theorem
1 depends on the existence of and . Both of these From the above expressions of , , , and ,
vectors can exist simultaneously only for a particular type of it can be easily verified that and ,
constraint sets [like constraint set (1)]. Directional breaking the- for any
orem 1 is applicable only to those cases.
and
APPENDIX B Now, let the alternative of this obligation MDFTR be simulta-
LOADING EFFECTS OF AN MDFTR ARE NOT HIGHER neously feasible with another set of FTRs and MDFTRs whose
THAN THOSE OF ITS ALTERNATIVE loading effects on forward and reverse simultaneous feasibility
There are two simultaneous feasibility constraints of the form constraints are and , respectively. Therefore
(5) per line per network topology. The first one in this constraint
set is the forward flow limit constraint, and the second one is the and
reverse flow limit constraint. and
Now, consider an MDFTR of MW, defined over
the path set ( ) with DV set Clearly, the MDFTR also maintains simultaneous feasibility.
( ). The sensitivity between the th However, as and , simultaneous
path of this MDFTR and a certain line is . For its obligation feasibility of this MDFTR does not ensure simultaneous feasi-
version, the cheapest alternative of this MDFTR is a portfolio bility of its alternative.
of option and obligation FTRs with MW obligation Similarly, the loading effects of the above MDFTR, in option
and MW option on path . Similarly, for version, on a forward flow limit and on the corresponding re-
its option version, the cheapest alternative of this MDFTR is a verse flow limit constraints ( and , respectively) can be
portfolio of option FTRs with MW on path . Here written as

and

where
For its obligation version, the loading effects of the MDFTR
on a forward flow limit and on the corresponding reverse flow
limit constraints ( and , respectively) can be written as

and

Authorized licensed use limited to: GOVERNMENT COLLEGE OF ENGINEERING. Downloaded on January 15, 2010 at 03:11 from IEEE Xplore. Restrictions apply.
 56 IEEE TRANSACTIONS ON POWER SYSTEMS, VOL. 23, NO. 1, FEBRUARY 2008

The loading effects of the alternative of this option MDFTR on It can be easily verified that
those forward and reverse flow limit constraints ( and
, respectively) can be written as

and

where
Note that

From the above expressions of , , , and , and

it can be easily verified that and ,
for any

and

As before, it can be further proven that the simultaneous fea- On the same line, it can be proven that
sibility of the above bundle of option rights also indicates the
simultaneous feasibility of the relevant MDFTR, although the
reverse may not hold true.

APPENDIX C where and are the new loading effects of the

CONVEX EXPANSION OF THE DV SET DOES NOT ALTER MDFTR on forward and reverse flow limit constraints, respec-
THE LOADING EFFECTS OF AN MDFTR tively, in option version.
Consider the same MDFTR again as in Appendix B. Let the
DV set of this MDFTR be expanded with a new distribution REFERENCES
vector which is a convex combination of original distri- [1] W. W. Hogan, “Contract networks for electric power transmission,” J.
bution vectors, i.e., Regulat. Econ., vol. 4, pp. 211–242, 1992.
[2] W. W. Hogan, Financial Transmission Right Formulations. [Online].
Available: http://www.whogan.com.
[3] X. Ma, D. I. Sun, G. W. Rosenwald, and A. L. Ott, “Advanced financial
transmission rights in the PJM market,” in Proc. IEEE/Power Eng. Soc.
General Meeting, Jul. 13–17, 2003, vol. 2, pp. 1031–1038.
[4] X. Ma, D. I. Sun, and A. Ott, “Implementation of the PJM financial
transmission rights auction market system,” in Proc. IEEE/Power Eng.
Soc. Summer Meeting, Jul. 21–24, 2002, vol. 3, pp. 1360–1365.
[5] PJM Manual on Financial Transmission Rights, PJM market. [Online].
Available: http://www.pjm.com.
[6] A. Bykhovsky, D. A. James, and C. A. Hanson, “The introduction of
option financial transmission rights into the new England market,” in
Proc. IEEE/Power Eng. Soc. General Meeting, Jun. 12–16, 2005, pp.
Let 284–289.
[7] T. Kristiansen, Markets for Financial Transmission Rights. [Online].
Available: http://www.ksg.harvard.edu/hepg/Papers/Kristiansen.mkts.
for.ftrs.Oct.03.updated.May.04.pdf.
[8] M. I. Alomoush and S. M. Shahidehpour, “Generalized model for fixed
transmission rights auction,” Elect. Power Syst. Res., vol. 54, no. 3, pp.
207–220, Jun. 2000.
[9] R. P. O’Neill, U. Helman, B. F. Hobbs, W. R. Stewart Jr., and M. H.
Rothkopf, “A joint energy and transmission rights auction: Proposal
and properties,” IEEE Trans. Power Syst., vol. 17, no. 4, pp. 1058–1067,
Nov. 2002.
[10] V. Sarkar and S. A. Khaparde, “A robust mathematical framework for
managing simultaneous feasibility condition in financial transmission
rights auction,” in Proc. IEEE/Power Eng. Soc. General Meeting, Jun.
and 18–22, 2006.
[11] H. Chao, S. Peck, S. Oren, and R. Wilson, “Flow-based transmission
rights and congestion management,” Electricity J., vol. 13, no. 8, pp.
38–58, Oct. 2000.
[12] L. E. Ruff, “Flowgates, contingency-constrained dispatch, and trans-
mission rights,” Electricity J., vol. 14, no. 1, pp. 34–55, Jan.-Feb. 2001.
where and are the new loading effects of the [13] R. P. O’Neil, U. Helman, R. Baldick, W. R. Stewart Jr., and M. H.
Rothkopf, “Contingent transmission rights in the standard market de-
MDFTR on forward and reverse flow limit constraints, respec- sign,” IEEE Trans. Power Syst., vol. 18, no. 4, pp. 1331–1337, Nov.
tively, in obligation version. 2003.

Authorized licensed use limited to: GOVERNMENT COLLEGE OF ENGINEERING. Downloaded on January 15, 2010 at 03:11 from IEEE Xplore. Restrictions apply.
SARKAR AND KHAPARDE: INTRODUCTION TO MULTIDIMENSIONAL FINANCIAL TRANSMISSION RIGHTS 57

[14] O. Alsaç, J. M. Bright, S. Brignone, M. Prais, C. Silva, B. Stott, and S. A. Khaparde (M’87–SM’91) received the Ph.D. degree in 1981 from the
N. Vempati, “The rights to fight price volatility,” IEEE Power Energy Indian Institute of Technology, Kharagpur, India.
Mag., vol. 2, no. 4, pp. 47–57, Jul.-Aug. 2004. He is a Professor of electrical engineering with the Indian Institute of Tech-
[15] S. Boyd and L. Vandenberghe, Convex Optimization. Cambridge, nology, Bombay, India. He has authored several research papers and has coau-
U.K.: Cambridge Univ. Press, 2006. thored two books. He is a member of the advisory committee of Maharastra
[16] A. J. Wood and B. F. Wollenberg, Power Generation, Operation, and Electricity Regulatory Commission, India. His current research activities are in
Control. New York: Wiley, 2003. the areas of power system restructuring.
[17] MATPOWER, A Matlab System Simulation Package. [Online]. Avail- Dr. Khaparde is on the editorial board of International Journal of Emerging
able: http://www.pserc.cornell.edu/matpower/. Electric Power Systems.

V. Sarkar received the B.E. degree in electrical engineering from Burdwan Uni-
versity, Bardhaman, India, in 2002 and the M.E. degree in electrical engineering
from the former Bengal Engineering College (currently Bengal Engineering and
Science University), Kolkata, India, in 2004. He is currently pursuing the Ph.D.
degree in the Department of Electrical Engineering at the Indian Institute of
Technology, Bombay, India.
His current research involves power system restructuring issues.

Authorized licensed use limited to: GOVERNMENT COLLEGE OF ENGINEERING. Downloaded on January 15, 2010 at 03:11 from IEEE Xplore. Restrictions apply.