IEEE Transactions on Poner S!stems, Vol. 13, No.

4, November 1998

1226

Optimal Real Time Pricing of Real and Reactive Powers

Joon Young Choi,

Seong-Hwang Rim

Member, IEEE
Dept. of Electrical Engineering
Jeonju University
Chonju, Chonbuk, 560-759 Korea

Jong-Keun Park,
Member, IEEE

Power System Planning Dept.

KEPCO
Seoul, 135-090 Korea

Abstract: This paper presents a theory and simulation results of

real time pricing of real and reactive powers that maximizes social
benefit. Conventional studies represented the responses of
consumers as demand functions, and the objective of optimization
was minimization of operating cost of power system. In this paper.
the responses of consumers are represented as the inverse of the
demand functions and power flow equations are incorporated in the
constraints. This makes it possible to incorporate constant power
factor demand consumer as a participant. That also makes it
possible to maximize the objective function of optimization that is
social benefit of participating in the electricity market and to
determine the price of electricity at each bus. Not only the real
time price of real power but also that of reactive power is
determined simultaneously by the incorporation of power factor of
loads into the constraint of optimization.

Keywords: Optimization, Electricity Pricing. Demand Function,

Benefit Function. Power Factor

I. INTRODUCTION

A concept of electricity energy market, which reflects

the generation cost on the price of electricity, is suggested to
replace conventional load management strategy [ I ] . This
market takes charge of transmission and distribution of
electricity and decides the price of electricity so as to
balance the demand and supply of electricity. Participants
participate in this market by selling and buying electricity.
Caramanis et a1 suggested a theory that maximizes the
benefit of participants and discussed the practical problems
that occurred when a new market was built. Also they made
it clear that the price that maximizes the social benefit is
marginal cost of electricity [2]. Subsequently, studies to
determine not only marginal cost of electricity but also
optimal investment on facility are carried out 13-51, But in
these studies, only the marginal cost of total generation is
PE-037-PWRS-0-10-1997 A paper recommended and approved by
the IEEE Power System Analysis, Computing and Economics
Committee of the IEEE Power Engineering Society for publication in the
IEEE Transactions on Power Systems. Manuscript submitted April 3,
1997; made available for printing October 21, 1997.

0885-8950/98/$10.00

School of Electrical Engineering

Seoul National University
Seoul, 15 1-742 Korea

calculated, not the marginal cost at each bus. Afterwards,

the responses of consumers and producers of electricity in
the real time pricing situation are studied [6-91, and the
demand forecasting studies are begun when the consumers
actively respond under the real time pricing situation
c10,111.
Baughman and Siddiqi [12] decided the price of
electricity at each bus of power system by solving the
bi-level optimization problem consisting of two parts. One
part is the first order condition of optimal power flow
problem that minimizes generation cost and the other part is
demand functions. These two parts should be solved
simultaneously to give the real time prices of real and
reactive powers, demands, generations, voltage magnitudes
and voltage angles of each bus, and the currents of each line.
However. this formulation hardly represent uniform power
factor demand. which is a common consumers' behavior,
because this bi-level optimization problem of optimal power
flow and demand function will not be solved simultaneously
if the demand function represents constant power factor
demand.
This paper presents a theory of social benefit
maximizing real time price of real and reactive powers and
its simulation results. An optimization problem whose
objective is maximizing the social benefit and whose
constraints are set of power flow equations, voltage limit of
buses, and transmission capacity limits of transmission lines
is formulated and solved. In this formulation, the demand
functions that represent the consumer are used in their
inverse form to give the objective function, the social
benefit. The real time price of reactive power is also
determined by incorporating the power factor of loads into
the constrains.

11. FORMULATION
It is assumed that there is an electricity energy market.
There exist three types of participants in the market. One
type is consumers of electric energy, another type is
producers of electricity and the other type is a regulator or
price setter who owns the transmission and distribution lines
system. The regulator is assumed to be able to set and
Each
communicate the price of electricity instantly.
participant represents itself by its characteristic function.
Producer's characteristic function is the cost function C(x),

0 1997 IEEE

1227

and consumer's characteristic function is the benefit function

B(x). Different types of consumers such as home, buildings,
and factories have different benefit functions. Moreover,
benefit functions vary according to the time of a day. We
assume that those benefit functions are independent from
each other.
The following notations, N , G and C represent set of all
buses in the system, set of all buses having generating capacity and set of all buses having consuming capacity,
respectively, and will be used in the formulation.

BENEFITFUNCTIONS
The demand function in (1) is assumed to be log linear
form exhibiting constant price elasticity. This demand
functions may be mathematically expressed in the following
form:

(5)

The price of elelctricity at one bus is the marginal value of

objective function in this maximization problem.
CONSTRAINTS
Power flow equations

A set of equations that characterizes the flow of real

and reactive powers through a system is given below;

P, - cyvJ1
cos(el/ + 6, - SI) = 0
J EN

(7)

which are the real and reactive power flow equations,

respectively, for all i, j E N . P, and Q are real and reactive
where D is the constant parameter and E is the elasticity of
the consumption. The inverse of this demand functions can
be expressed as follows.

powers of bus i,

and VJ are the magnitudes of the

voltages of bus i and j , 6, and 6, are the voltage angles of

bus i andj, and

and 6,,are the magnitude and angle of

, respectively.
ijth element of the admittance matrix YBu2;
This inverse demand function represents the marginal benefit
of consuming unit quantity of electricity. The integral of this
hnction becomes the benefit function B ( x ) .
I

(3)

Generation limits
The generating plants have a maiximum generating
capacity, above which is not feasible to generate due to
technical or elconomic reasons. Generation limits are
important in determining the operating point and marginal
costs of generation. Generating limits are usually expressed
as maximum or minimum real and reactive power outputs,

OBJECTIVE
The objective of pricing policy is to maximize the
benefit of all the participants, that is, to maximize
consumers' and producers' surplus, subject to the operational
constraints. This is accomplished by setting the prices of
real and reactive powers at each bus at a particular time
equal to the marginal values of supplying and consuming
real and reactive power, at the same bus and at the same
time, where the marginal values are determined by
maximizing the total surpluses of utilities and consumers,
subject to the operational constraints. The total surplus, the
sum of benefits of participating in the electricity market, is
as follows.
CB1(x,)- CC,(x,)
I tr

(4)

I EG

C(x) is the cost function of production that represents the

producers, and B ( x ) is the benefit function of consumption

at quantity x that represents the consumers. Maximizing this
sum means maximizing the social benefit and this is the
objective of pricing policy.

where P,,,i, and P,,,, are the minimum and maximum real
power outputs and

and Q,,,,

are the minimum and

maximum reactive power outputs, respectively, of the

generating plant at bus g, for all g E G .
Consumption limits
Consumers also have their capacity limits of
consumption. Consumption limits are expressed as follows
in this formulation,

1228

where

e,,,, and e,,

power consumption and

are the minimum and maximum real

and

a,,,, are the minimum

and maximum reactive power consumption, respectively.

p f , is the power factor of the consumers at bus c, for all
c E C . Consumption limits have their importance not in the
consumption of real power but in that of reactive power.
The ratio between real and reactive powers called power
factor is important in determining the price of reactive
power.

TO BE OPTI\flZED
THE LAGRANGIAN

The Lagrangian of this optimization problem is as

follows.

Transmission limits
Transmission limits refer to the maximum power that
given transmission line is capable of transmitting under
given conditions. These limits can be based on thermal
considerations or stability considerations. Thermal limits are
usually dominate for shorter lines. Dynamic stability limits
dominate for longer line behavior. These limits affect the
marginal costs of operation. Thermal transmission limits of
thermal consideration are expressed here in terms of the
maximum real power flow through lines,
<,,nun

41

'?j,max

(13)

I.;v/q

cos(bj, + 6, - 6,) - y 2 qCOSQ,, , assuming

where 6, =
that the shunt admittance is negligible, where
and

e,,,,, are the minimum and maximum real power flows,

THEPRICE OF ELECTRICITY

respectively, through the line connecting bus i to b u s j , for i

# j , and all i, j E N. Stability limits, which are dominant for
longer lines, of transmission lines between bus i and j are
expressed as follows,

The prices of real and reactive powers at bus i are the

Lagrangian multiplier value of the equality constraints, that
is, the real and reactive power flow equation of bus i. The
Lagrangian multiplier values are calculated by solving the
first order condition of the Lagrangian, partial derivatives of
the Lagrangian with respect to every variables concerned.
Taking the partial derivative of the Lagrangian with respect
we
to the quantity of real and reactive powers, 4 and

41,mtn

5 4 - 8,5 4i.ma.x

(14)

where 8, and 6, are the voltage angle of bus i and j , and

eJ,,,

e,,,,

and
are the minimum and maximum values of
the difference in the voltage angles of bus i and j ,
respectively.

e,,

Voltage limits

Voltage limits refer to the requirement for the system

bus voltages to remain within a narrow range of levels.
Since voltages are affected primarily by reactive power
flows, the marginal cost of reactive power at a bus is directly
dependent on the voltage level requirement at that bus.

Voltage limits can be expressed by the following constraints,

min

Y,max

(15)

where I.;,,,," and V,,ax are the minimum and maximum

voltage levels, respectively, that are acceptable at bus i, for
all i E N .

and we get the relation between the Lagrangian multipliers,

the prices of real and reactive powers at bus i, A,, and & .
and other multipliers such as price or premium of capacity
limit P p , m m and Pcp,max

1229

Max{B( p, ) - C(

e)}= Max(80000 - ~ . O , P , - ~ .-~ -21 4 2 )

A =-

de

Constraints
Equality constraints of this problem are load flow
equations of this power system and power factor constraint
of load bus.

111. SIMULATION

Q, + 17.7:lsin(-1.25) + 16.76V2sin(1.89 +8,)

- p2 - 16.76V2cos(1.89 - 6,) - 15.78Vi cOs(-1.25)
- Q2 + 16.761V,sin(1.89 - 8,)+ 15.7SVl sin(-1.25)

Two Bus CASE

V=1.06,

4 - 17.73~0~(-1.25)
- 16.76V2 COS(^ 89 + 8,)

6=o.o

~=5-J15
line charging y = - 116.667

Q, = 0.488
Fig. 1. Two bus power system

A simple two bus power system is used in the

simulation to illustrate the effect of real time pricing of
electricity at marginal costs. A consumer and a producer are
connected to each bus, and the buses are connected by a
transmission line. This simple model includes every
component of electricity energy market, and shows the effect
of the real time pricing.

('p$0.9)

Inequality constraints are the operational limits of the system

and expressed a a follows.

0.0 I 6 5 3.0
0.0 5 P2 23.0
-2.5s 62S2.5
0.95 5 V, S 1.05

-3.0 I Q, 53.0
-3.0 I Q, I 3.0
-0.785 I Si, 5 0.785

Objective function
The marginal cost function of a producer and the
inverse demand function of a consumer are assumed to be
simple forms as follows. The cost function and benefit
function are the integrals of the marginal cost function and
inverse demand function, respectively.

Real time price

eFlat price
PI

1
2

07

C(4) = jMC(q)dq = - P i
0

0 6 / ;

:
4

12

16

20

24

Time [HR]

Fig 2 Comparison of real powers at real time price vs flat price

= 80000 - 8 P2-4O

The range of this integral, the benefit function, is from c to

4 . If the starting point of the integral is 0, the benefit
function goes infinity, so the small constant, compared to
4 , is used as a starting value of the integral. E is the
elasticity of the consumption and set to be -0.2 in this
simulation. The objective of this optimization problem is
maximizing the social benefit, the difference of benefit
function and cost hnction.

Fig. 2 is a comparison of real power demands with two

different pricing policies. The line with black square shows
the real power demand of real time pricing, and the line with
white square is the real power demand <ofconventional flat
pricing. The peak demand of real time pricing is less than
that of flat pricing, and the bottom demand of real time
pricing is higher than that of flat pricing. So, the fluctuation
of the real time pricing demand is less than that of the flat
pricing demand. This means real time pricing smoothes the
peak demand and improves load rate.

1230

Fig. 5 is a graph of charges of real and reactive powers

and its ratio when the power factor of load is varying from
0.70 to 0.99. The charge of power is defined as the product
of quantity and price of power. When the power factor of
load is 0.70, the amount of reactive power charge is half of
the real power charge. This demonstrates that the marginal
value of reactive power is noticeable when the power factor
of load has low value, and the users of reactive power should
be charged fairly for reactive power usage.

T
1.15

lo0l
105

0 95

0 85

td

delivered

+generated

FIVEBUS C.GE

Fig. 3. Prices of real power of two buses

Fig. 3 is a graph of the prices of real power at each bus.

The prices of buses vary according to the cost function or
benefit function of bus. The difference in prices between
two buses is due to the transmission loss.

A simulation is carried out using a 5 bus system of

Table 1. Bus 1 and 2 are generator buses and bus 3. 4, 5 are
load buses. The power factors of loads are set to be 0.9, and
the power demands of loads are fixed constant. The change
in the prices of reactive power at loads is investigated when
the voltage limit of bus 5 becomes strict. Fig. 6 shows the
change in the price of reactive power at different nodes. The
prices of reactive powers start to rise when the voltage limit
of bus 5 reaches 0.96 p.u. The real time price of reactive
power rises at bus 5 but changes a little at other load buses
with the increasingly tight constraint on the voltage of bus 5 .
TABLE I. IMPEDANCE DATA OF 5 BUS SYSTEM

Bus l o .

2
3

Fig. 4 is a graph of the prices and quantities of real and

reactive powers at Load bus when the power factor of load is
varying from 0.70 to 0.99. The real power quantity grows
and reactive power quantity dwindles as the power factor of
load grows. The prices are dependent upon the quantities of
the powers.
8

Real power

0 06

2
2
2

3
4
>

0 04

0 01
0 08

0 05
0 04
0 04

0 18

0 12
0 03

0 03
0 02
0 05

0 24

12
1 .

+ 06

Reactive P o w e r

- 05

C h a r g e Ratio

Line Charging y'pq

._
f 04 -

i0 7

X
0 06
0 24
0 18

0 08
0 06
0 06

Fig. 4. Prices of real and reactive power at load

0 02

0 4
W

2 4

0 95

096

0 97

0 98

0 99

1 00

Bus Voltage Limit [P.u.]

tn

03

Fig 6 Price change of reactive powers at load buses

02

0 1

0
010

075

080

085

090

IV. CONCLUSION

095

Power Factor of Load

Fig. 5. Charges ofreal and reactive power and its ratio

In this paper, a theory of social benefit maximizing the

real time price of real and reactive powers and its simulation

1231

results are presented. The objective function, which is the

sum of participants' social benefit of participating in the
electricity energy market, is the sum of benefit functions and
cost functions of participants. The Lagrangian multipliers of
this maximization problem, subject to operational and
capacity constraints, are social benefit maximizing the real
time price of electrical power. The calculation of marginal
social benefit of buses is possible by the inclusion of the
power flow equation into the constraints of optimization.
And the calculation of marginal social benefit of reactive
power of buses is possible by the inclusion of the power
factor of load buses into the constraints of optimization.
This real time price of real and reactive powers smoothes the
peak power demand and fills the valley of power demand.
This peak clipping could reduce the investment of power
system equipment and both peak clipping and valley filling
effect improve load rate.

V. REFERENCES

F. C. Schweppe, R. D. Tabors, J. L. Kirtley, Jr., H. R.

Outhred, F. H. Pickel, and A. J. Cox, "Homeostatic
Utility Control," IEEE Trans. Power Apparatus and
Systems, vol. PAS-99, no. 3, MayIJune 1980, pp. 1151
- 1163.
MI. C. Caramanis, R. E. Bohn, and F. C. Schweppe,
"Optimal Spot Pricing: Practice and Theory," IEEE
Trans. Power Apparatus and Systems, vol. PAS-101,
no. 9, September 1982, pp. 3234 - 3245.
Y. Kaya, "Dynamic Analysis of Time-of-day Pricing in
Electrical Power Systems. A Theoretical Approach,"
Trans. ZEEJ, vol. 105-C, no. 10, October 1985, pp. 193
- 200.
Y. Kaya, and H. Asano, "Dynamic Analysis of Timeof-day Pricing in Electrical Power Systems. A
Numerical Analysis," Trans. IEEJ, vol. 106-C, no. 4,
April 1986, pp. 57 - 64.
Y. Kaya, and H. Asano, "A Theory of Time-of-Use
Rate for the Power System with Multi-type Power
Plants," Trans. IEEJ, vol. 108-B, no. 9, September
1988, pp. 399 - 406.
E. V. Garcia and J. E. Runnels, "The Utility
Perspective of Spot Pricing," IEEE Trans. Power
Apparatus and Systems, vol. PAS-104, no. 6, June
1985, pp. 1391 - 1393.
F. C. Schweppe, M. C. Caramanis, and R. D. Tabors,
"Evaluation of Spot Price Based Electricity Rates,"
IEEE Trans. Power Apparatus and Systems, vol. PAS104, no. 7, July 1985, pp. 1644 - 1655.
R. D. Tabors, F. C. Schweppe, and M. C. Caramanis,
"Utility Experience with Real Time Rates," IEEE
Trans. Power Systems, vol. 4, no. 2, May 1989, pp.
463 - 471.

A. K. David, "Optimal Consuimer Response for

Electricity Spot Pricing," IEE Proceedings, vol. 135,
pt. C, no. 5, September 1988, pp. 3'78 - 384.
A. K. David, "Load Forecasting under Spot Pricing,"
IEE Proceedings, vol. 135, gt. C, no. 5. September
1988, pp. 369 - 377.
A. K. David, and Y. C. Lee, "Dynamic Tariffs: Theory
of Utility - Customer Interaction," ZEEE Trans. Power
Systems, vol. 4, no. 3, August 1989, pp. 904 - 91 1.
M. L. Elaughman, and S. N. Siddiqi, "Real-Time
Pricing of Reactive Power: Theory and Case Study
Results," IEEE Trans. Power Systems, vol. 6 , no. 1,
February 1991, pp. 23 - 29.
M. A. Crew, and P. R. Kleindorfer, The Economics of
Public Utility Regulation, Macmillan Press, 1986, pp.
11.

VI. BIOGRAPHIES
Joon Young Choi (Member) was born in Naju, Korea on July 9, 1963. He
received his BS, MS and Ph.D. degrees in electricid engineering from Seoul
National University, Korea in 1986, 1988 and 1994, respectively. In 1994,
Dr. Choi worked as a research engineer at the LG Electronics, Inc. He is
currently a full time lecturer of Department of Electrical Engineering,
Jeonju University, Chonju, Chonbuk, Korea. HIis research interests are
power system economics and electricity pricing.
Seong-Hwang Rim was born in Pohang, Korea on June 2, 1959. He
received his BS and MS degrees in electrical engineering from Seoul
National University, Korea in 1987 and 1994, respectively. He is presently
with the Korea Electric Power Corporation (KEPCO), Seoul, Korea. His
research interests are power system economics and electricity pricing.
Jong-Keun Park. (Member) was born in Chung-nam province, Korea on
October 21, 1952. He received his BS degree in electrical engineering from
Seoul National University, Korea in 1973 and his MS and Ph.D. degrees in
electrical engineering from University of Tokyo, Japan in 1979 and 1982,
respectively. In 1982, Dr. Park worked as a researcher at the Toshiba Heavy
Apparatus Laboratory. He has been an associate professor of Seoul
National University since 1983. He is currently a professor of School of
Electrical Engineering, Seoul National University. In 1992 he attended as a
visiting professor at Technology and Policy Program and Laboratory for
Electromagnetic and Electronic Systems, M,%sachusetts Institute of
Technology. He 11sthe Director of Technoeconomics Program, College of
Engineering, Seoul National University, Korea. He is a member of the
Korean Institute of Electrical Engineers and Japan Institute of Electrical
Engineers.