Electrical Power and Energy Systems 58 (2014) 307318

Contents lists available at ScienceDirect

Electrical Power and Energy Systems

journal homepage: www.elsevier.com/locate/ijepes

A multi-objective solution algorithm for optimum utilization of Smart

Grid infrastructure towards social welfare
Sandip Chanda a,, Abhinandan De b
a
b

Department of Administration, Technique Polytechnic Institute, Hooghly, India

Department of Electrical Engineering, Bengal Engineering & Science University, Shibpur, India

a r t i c l e

i n f o

Article history:
Received 31 July 2012
Received in revised form 15 August 2013
Accepted 11 January 2014

Keywords:
Smart Grid
State space
Demand response
Generation surplus
Jacobian
Load curtailment

a b s t r a c t
This paper proposes an optimization model to maximize social welfare by standardizing the operating
conditions with an overall improvement of dynamic stability of power markets endowed with Smart Grid
communication technology. The state space based model developed along with the proposed methodology maximizes load catering and simultaneously minimizes the operating standard constrained generation cost to restore power market equilibrium even in the most inadvertent states of the Energy System
Network. For optimum utilization of smart metering facility, the model effectively involves resources like
demand response, generation surplus and an efcient methodology to optimize the Market Clearing Price
(MCP) as well as prot of the market participants by effective categorization. The power market dynamic
price equilibrium has been estimated by forming Jacobian of the sensitivity matrix to regulate the state
variables for the standardization of the quality of solution. A novel load curtailment strategy has also
been proposed to amalgam stability restoring shedding with prot retentive load cut. The model has been
tested in IEEE 30 bus system in comparison with standard curtailment based optimization technique to
produce encouraging results.
2014 Elsevier Ltd. All rights reserved.

1. Introduction
The synergism of power system with information network has
emerged as Smart Grid to undertake the modern power network
issues like demand side management, standardization of operating
conditions and integration of renewable energy sources. Moreover
the Smart Grid is inherently designed to be self-healing to improve
reliability and to respond to natural disaster or malicious sabotage
[1]. Efcient deployment of information network augmented with
Smart Grid can appear to be an invaluable resource to regulate the
operational condition of the system and to optimize the system
operation to a prolic solution [2,3]. In the quest of optimizing
the utilization of these new resources, researchers in the recent
past have been proposing indigenous methodologies and solution
algorithm. Though the power system planers heavily rely upon
the methodologies in [4,5] introduced a distinctive work where
operating conditions viz loss and voltage proles were optimized
with a coordination methodology of plug in vehicle charging. A hybrid method has been enunciated in [6] for effective utilization of
Smart Grid data viz synchrophasor to improve grid reliability from
Corresponding author. Tel.: +91 9836921589.
E-mail addresses: sandipee1978@gmail.com (S. Chanda), abhinandan.de@
gmail.com (A. De).
http://dx.doi.org/10.1016/j.ijepes.2014.01.029
0142-0615/ 2014 Elsevier Ltd. All rights reserved.

generation side. Ref. [7] portrayed a framework of future transmission grid while [810] identied the challenges associated with the
incorporation of Demand Response (DR) in distribution of the
existing grid. From this survey, it is quite evident that all the major
parts of the grid require extensive reformation for optimization of
Smart Grid resources summarized in [11]. All these alterations will
lead to a grid capable of monitoring and control and fast responsive
devices are to be installed to retaliate almost instantaneously [12]
to locate disturbance and to minimize the same. The grid under
consideration must possess at least these attributes to negotiate
matters like intermittent energy sources, up-gradation of operating
conditions and self-regularization. During the incorporation of
renewable energy sources the unprecedented intermittent nature
and cost curve pose immense difculty to system optimization.
The challenges and possible solutions have been enunciated for
the power system networks of Europe in [13]. Refs. [1416] proposed optimization methodologies to escalate the operational status of power grid subscribing a particular renewable energy source.
Ref. [17] depicted a novel algorithm to regulate the system parameters under multiple intermittent sources. The model optimized
the system operation by an energy hub concept but the implementation of the same will require efcient infrastructure which may
not be available and the model moreover does not incorporate issues like line ow management and load curtailment. Ref. [18]

308

S. Chanda, A. De / Electrical Power and Energy Systems 58 (2014) 307318

dealt with the integration of demand response with irregular energy sources. The multi objective algorithm proposed minimized
cost and curtailment but could not maximize load catering policy
of the system operator. In addition, issues like payment cost minimization by involving maximum number of consumers as suggested in [1921] have not been considered. Apart from
tribulations associated with intermittent renewable energy
sources, Smart Grid has to efciently employ demand side management technique. The methods proposed in [2224] effectively
utilize demand side bidding or auction strategy or multi-agent policy but the price responsiveness of other available parameters have
not been considered. All these efforts concentrated only one of the
price responsive parameter and their objective was to minimize
the market clearing price rather than maximizing social welfare
by catering optimum load at minimum cost. Refs. [25,26] introduced social welfare as an objective but apply less emphasis on
sustaining operating conditions of the system or the load shedding
technique. The contemporary load curtailment strategies have
been illustrated in [27,28] whereas in [29,30] some new strategies
have been introduced. Most of these techniques are operating condition constraint Optimal Load Curtailment (OLC) programs could
not assure customer a reliable supply with standard operating conditions. In view of this above survey, the need of an algorithm can
be felt which can ensure a standard parametric operational condition with an objective of minimizing the price of electricity with
optimal load catering without violating the price equilibrium of
the market. The algorithm is required to be supported by a pricing
model, which not only integrates the demand response and generation characteristics but also involves price sensitivity of voltage
prole, line loss, congestion and load curtailment. The existing
price forecasting models proposed in [3139] are optimistic in nature developed to offer solutions under specic operational constraints and these models are only fertile for single objective
domain. Moreover they do not consider the dynamic price equilibrium as stated in [40] hence cannot provide an insight of the power
market stability. The endeavor of the work presented in this paper
has been to develop a state space model of a power system network endowed with smart metering facility not only to forecast
price, but also to minimize the same without compromising social
welfare, power market stability and thus ensuring sustenance of
prolic operating conditions. The convex nature of solution algorithm with nonlinear working surface was compelling in the selection of a stochastic optimization technique like Particle Swarm
Optimization (PSO). For comparison of the solution obtained a
standard Optimal Power Flow (OPF) [41] has been adopted. The
simulations have been carried out in IEEE 30 bus system and the
obtained results looked quite promising.

(RTO), historical and forecasted data, Available Transmission

Capacity (ATC) margins etc.
Efcient employment with these new resources as wholesale
market product, Independent System Operator (ISO) will be able
to reach the furthest corners of the network from generation to
load end to maintain profound operating conditions under the
worst possible states of the system. In this context ISO will be able
to identify the state variable creating imbalance in the power market to de-standardize its operations. The price responsiveness of
the state variables of modern power markets has been elaborated
in the following section.
3. The proposed price responsive OPF model
3.1. Price sensitivity of demand
The price responsiveness of demand has become an incredible
input to the OPF algorithms for their load peak shaving capability
in high price conditions [3]. Effective deployment of this resource
may lead to the solutions of modern day power network tribulations like network congestion, voltage instability and perturbations
in dynamics in power market. The demand elasticity of price is,
hence as depicted in [42] an important parameter to be considered
for Optimal Power Flow. With the assistance of smart metering this
product can be incorporated in optimization to achieve distinction
in operating conditions and welfare of the market participants. As
shown in Fig. 2, a demand with a marginal benet above the marginal price will lead to an expansion in consumption until the equilibrium is reached. In Fig. 2a ABC represents the bid curve at a
particular hour, while XY its tangent. The price responsive demand curve (Fig. 2b) shows the nature of the consumer towards
price volatility corresponding to the bids. From the bid curve the
willingness to pay of the consumers can be determined as

willingness to pay tan h

where d1, d2. . .etc. are the power demands and the bid curve is
expressed as a function of demand such as f(di). The Market Clearing
Price (MCP) corresponding to each point of the bid curve have been
plotted in Fig. 2a and b. In this gure it has been assumed that
b1 = d1MCP1, b2 = d2MCP2, b3 = d3MCP3. Let us assume that the
two curves are tted with two different polynomials. k1x2 + l1x + m1
represent the bid curve while k2x2 + l2x + m2 represent the price
responsive curve where k1, l1, m1 are the coefcients of bid curve
and k2, l2, m2 are the coefcients of price responsive curve. Now
mapping willingness to pay into price responsiveness of demand

tan h
2. Market structure and functioning in Smart Grid

The Smart Grid is an eco-friendly optimization of the present

grid endeavored to achieve operational excellence with high degree of reliability. The functional architecture proposed [2] and
implemented [10] are based on some basic modication of the
present grid organization for proper management of distributed
generation with renewable energy sources, improvement of sustainability with self healing activities like congestion, power quality management and encouragement of price responsive demand
reduction. The profusion of these activities employs extensive bidirectional communication between wholesale markets/transmission operation and retail markets/distribution operations. Fig. 1 depicts one such architecture enabling the system operator to not
only utilize the generator information, but is also to make itself
capable of incorporating the demand response of consumers aggregated by local entities like Regional Transmission Organization

f d1  f d2 b1  b2

d1  d2
d1  d2

b2  b1 d2  MCP 2  d1  MCP1

d2  d1
d2  d1


2
2
d2 k2 d2 l2 d2 m2  d1 k2 d1 l2 d2 m2

d2  d1


2
2
k2 d2 d2  d1 d1 l2 d2 d1 m2

The willingness to pay is as sensitive to bid curve as is to price

responsive curve. Hence, inclusion of demand response or price
responsive characteristics into OPF not only incorporates the price
dependent consumption characteristics but also involves the willingness to pay of the consumers for a particular alteration in price.
In the present work load demand is also scheduled like generation,
the load curtailment becomes willingness to pay dependent and
involves the consumer more into OPF. In every hour (or a specied
period) the consumer with the assistance of smart metering, will
be able to modify his stand in the power market enabling Independent System Operator (ISO) to regulate curtailment depending on
willingness to pay of the consumers.

S. Chanda, A. De / Electrical Power and Energy Systems 58 (2014) 307318

309

Fig. 1. Interaction of power market entities of Smart Grid. Note: Conceptualized architecture of modern power grid or Smart Grid to ensure optimum utilization of generation,
transmission and distribution resources.

Fig. 2. A and B: Mapping the price responsiveness of demand from Bid curve. Note: Incorporation of this price responsive demand curve into optimal power ow ensures load
curtailment of consumers in accordance with their willingness to pay.

3.2. An alternative approach to price dependent load curtailment

In Demand Response (DR) scenario, the load demand is also
considered to be following a predetermined schedule just like

generation schedule, the willingness to pay of individual consumers are available to the system operator. The operator under peak
load condition, for security limit violation, can curtail a part of
the load according to consumers demand response curve but the

310

S. Chanda, A. De / Electrical Power and Energy Systems 58 (2014) 307318

process is entirely price dependant [28,29,43]. This process does

not concern the revenue loss suffered by the Generation Companies (GENCOs). A more rational way of curtailment should have
considered not only the market price but also the generation surplus of the GENCOs. The generation surplus can be dened as the
difference between maximum generation available and load dispatched. The present work proposes a novel curtailment strategy,
which not only adheres to consumer demand response curve, but
also equally shields the GENCOs against a foreseeable revenue loss.
In the proposed strategy, the ISO can set generation surplus Smax
for the GENCOs above which the net curtailment would be zero to
protect GENCOs from further revenue loss, irrespective of price
condition of the market. In demand response the minimum limit
of demand is set by the consumer. To maintain reliability of supply
and to build consumers condence for the least supply this minimum dispatch has to be assured for all the time.

Smax Sum of maximum possible generation

xh 1 f p Axh Buh
Is a function of price of electricity
where A is the sensitivity
(m + n + 4)  (m + n + 4).

A11
..
.

6
A6
4

 Sum of minimum limit of requested demand

When surplus is maximum i.e. Smax, the ISO should stick to zero
curtailment policy conversely to maintain price equilibrium ISO
must also set a maximum limit for curtailment Pmax when the surplus is minimum Smin.
The maximum curtailment limit Pmax can alternatively expressed in the form of

Pmax Sum of maximum limit of requested demand

 Sum of minimum limit of requested demand
For the rest of the cases ISO may choose the operating point
from the curtailment-surplus relationship, which can be approximated by a straight-line relationship

Pmax  S  Smin
P Pmax 
Smax  Smin

It is quite evident from the above discussion that the price of

electricity is dependent on these state variables.
And hence price of electricity p = f(x), as the market players
are connected with ISO through smart metering, on receipt of
the price information at the hth hour (a chosen time frame)
the participants either alter or persist with their characteristics.
With the objective of social welfare the ISO implements the
OPF and determines the price dependent state variables
x(h + 1) at the (h + 1)th hour based on the price of hth hour.
Mathematically

Amn41

...

matrix

of

5
the

order

of

A1mn4

7
7
5
. . . Amn4mm4
bi;1

bi;1

bi;1
with the elements Aij bj;1
bj;1n1 bj;1n , for linear approximation
n1
n
about hth hour.
B is the contingency or state modication matrix, which will
arise only when participant characteristics are altered by either
deliberately or inadvertently.
In an energy system or power system network the above modeling not only enables the ISO to determine stability of a particular
optimal solution in an hour but also makes the ISO capable of predicting the most feasible optimal generation pattern and load schedule for a specied operational standard.
The ISO benet function can be dened as

yh 1 Cxh
The demand cost benefit function

 limit violation constraint generation cost

where S represents generation surplus and P represent the corresponding upper limit of curtailment or ISO set limit of load curtailment. The proposed method utilizes Eq. (3) to determine the
maximum allowable load curtailment at any given price equilibrium point or ISO set limit of curtailment.
The proposed methodology with the assistance of PSO searches
for an optimal solution which can maximize the income of the
generators by reducing generation surplus to minimum while limiting load curtailment to an optimum value by adjusting the demand of the consumers.
The price of electricity in modern power markets not only relies
on demand response or load curtailment, but also includes generator cost characteristics, congestion, voltage and loss prole management costs [24,4446]. Optimum utilization of Smart Grid
infrastructure will be ensured only when all these inuencing factors are standardized with an objective to maximize the benets
for all the market participants.
4. State space modeling of the proposed system

Cd  Cglv
Social Welfare with desired operating conditions 6
where

Cd

nd
X
r j P 2dj sj Pdj demand cost benefit function

j1

d stand for demand, rj, sj are coefcients of cost benet function,

and
ng
X
Cglv
ai P2gi bi Pgi ci Pc  P 1 T L  P 2 V min  P3 Plmax  P4

i1

Limit violation constrained generation cost

8
glv stands for generation cost with limit violation constraint.
And C is a 1  (m + n + 4) matrix with the following pattern of
cofactors

C g1  C g2 . . .  C gm C L1 C L2 . . . C Ln  C Pc  C TL  C Vmin  C Plmax 
For m numbers of generators and n numbers of loads the price
sensitive state variables of a power system network may be dened as follows:

X P1 P2 . . . Pm L1 L2 . . . Ln Pc T L V min Plmax T

where Pi = Generation of ith bus for a particular schedule, Lj = Load

catered for jth bus for a particular load schedule Pc = load curtailed,
TL = Total line loss, Vmin = minimum bus voltage, Plmax = maximum
line ow.

where Cgi = Cofactors of the ith bus of generation, CLj = Cofactor to

jth bus of load, CPc = Cofactor of the load curtailed, CTL = Cofactor
of transmission loss, CVmin = Cofactor of the minimum bus voltage,
CPlmax = Cofactor of maximum line ow, ai, bi, ci = Cost coefcients
of the ith generator, Pgi = generation of the ith generator, Pdj = Load
catered to jth consumer, Vmin = minimum value of bus voltage, P1,
P2, P3 = Penalties for limit violation set by ISO.
Hence the objective function of the proposed optimization algorithm stands as

S. Chanda, A. De / Electrical Power and Energy Systems 58 (2014) 307318

Fig. 3. Traditional optimization technique. Note: Conventional optimization technique based on generation cost and Market Clearing Price (MCP).

Maximize

nd
X
r j P 2dj sj Pdj

j1

ng
X

ai P2gi

The objective function has been proposed in this literature as

the difference of consumer cost benet function and the limit violation constrained generation cost. Maximization of this result in
maximization of cost benet function or load catering as well as
minimization of generation cost and constraint violations like load
curtailment limit, transmission power loss limit, limit of minimum
bus voltage and maximum line ow or congestion limit. For the
collection of higher revenue, the generator companies would want
optimum utilization of available generation (also known as utility
maximization) by proper load distribution. On the other hand, consumers or distribution companies (DISCO) would want low price
with higher reliability of supply. In between the transmission companies (TRANSCO) would like to maintain standard operation condition in terms low transmission loss, high value of minimum bus
voltage and low level of line congestion. Thus the objective function causes welfare of all the stakeholders associated with energy
system under consideration.
In conventional power network optimization method, the
objective function is same as depicted in [47]. But in Smart Grid
scenario with the available communication facility, as enunciated
in [48], social welfare maximization should be the objective of
optimization. The function expressed in (9) is the outcome of that
contemplation.
For effective comparison of the proposed model, a standard OPF
model proposed in [28,40] has been utilized.
Equality or power balance constraints for the present OPF are
b
X
V j Gij cos hij Bij sin hij 0

PGi  PDi  V i

Q Gi  Q Di  V i

bi P gi ci Pc  P1 T L  P2 V min  P3 P lmax  P4

10

i1

i1

Instead of traditional cost optimization the system operators

should prefer formulation of pareto-optimal objectives for catering
maximum amount of load at standard operational constraints.

311

b
X

V j Gij cos hij  Bij sin hij 0

11

i1

where PGi = Active power injected in bus i, PDi = Active power demand on bus i, Vi = magnitude of voltage of buses i, Vj = magnitude

Fig. 4. The proposed optimization technique. Note: Proposed energy systems structure with smart metering facility which will assist the consumers to alter their
consumption level by directly participating in optimization with their price responsive demand characteristics.

312

S. Chanda, A. De / Electrical Power and Energy Systems 58 (2014) 307318

of voltage of bus j, Gij = Conductance of transmission line from bus i

to j, Bij = Susceptance of transmission line from bus i to j, b = number
of buses.
Inequality or generator output constraints

Pmin
 Pgi  Pmax
gi
gi

12

Q min
 Q gi  Q max
gi
gi

13

where, Pgi, Qgi = Active and reactive power of generator i respectively, Pmin
Q min
gi Lower limit of active and reactive power of the
gi ,
generators,

Pmax
gi ,

Q max
Upper limit of active and reactive power
gi

of the generators

Voltage constraint : V min

 V i  V max
i
i
V max
; V min
i
i

14

are upper and lower limits of Vi

Transmission constraint : Pijmax  Pij  Pijmin

15

Pijmax, Pijmin are the maximum and minimum line ow limits of

Pij

Curtailment Constraint : 0 < Pc < Pci

16

Pci is the ISO set limit of maximum allowable curtailment determined by Eq. (3).

Transmission Loss Constraint : 0 < T L < T LM

17

TLM is the maximum limit of power loss in MW.

5. Determination of price stability
In practice sensitivity matrix A can be applied to the small-signal stability analysis and the formulation of Jacobian matrix in
steady-state determination. From (5) and (6) the eigen values of
A can be used to determine the power market price equilibrium.
If all the eigen values lie to the left hand side of S plane, the system
will be stable. On the other hand, in the steady-state determination
method using Newtons formula, the Jacobian matrix J(r) in the rth
iteration can be expressed as follows:

J r

dDxn1 dxn1  xn dxn1 dxn


AI
dxn
dxn
dxn
dxn

18

In the proposed method this Jacobian and its eigen values are
determined to estimate stability. It may be mentioned that if the
price stability is not obtained for a given value of curtailment, it
may be necessary to alter curtailment P even beyond Pmax (As stated in Section 2). Obviously, the optimal operation will be hampered under such a situation but this may be inevitable to
maintain stability at the cost of losing optimality.
A comparison of the traditional and proposed optimization
techniques has been depicted in following Figs. 3 and 4
Table 1
Parameter setting of PSO.
Parameter

Setting used in simulation

Dimension of the problem

Maximum particle velocity
Maximum number of iterations
Population size
Acceleration constant 1
Acceleration constant 2
Initial inertia weight
Final inertia weight
Minimum global error gradient
PSO seed position

27
04
100
24
02
02
0.9
0.4
1  1025
0

A particular setting of PSO parameter is required to ensure desired solution space.

Fig. 5. Flowchart of the proposed algorithm. Note: The state space based model is a part
of this methodology which ensures social welfare by maximizing load catering and
minimizing generation cost maintaining standard operational conditions in the system.

313

S. Chanda, A. De / Electrical Power and Energy Systems 58 (2014) 307318

performing a controlled OLC based load curtailment at bus no. 5

and 2 (assumed to have least willingness to pay or the same has
gotten its turn in a rotational load curtailment strategy). In the
1st test bed active loading stress has been increased on selected
number of buses and the operating conditions have been kept under observation. Though the minimum bus voltage gets least affected with active power loading stress, Table 2 depicts a
signicant improvement in the line loss prole and line ow distribution compared to the traditional method. In the next step of
study, reactive loading stress has been applied on the weakest
bus of the system (bus no 30). As discussed earlier that excess
requirement of reactive power not only degrades the operating
conditions of the system especially the voltage prole but also affects the price of electricity. For the same MCP, a tantalizing
improvement in bus voltage prole, line loss and line congestion
can be observed in Table 2 for the proposed method. Line contingency is almost inevitable in modern and future grids working under heavy loading stress. Disruption resilience with self-healing
should be one of the most desirable attributes of the present
and future Smart Grid. It is imperative from Table 2 that the proposed algorithm working in a Smart Grid scenario would deliver
more excelling operating conditions especially in terms of congestion management under single or multiple contingencies as compared to traditional optimization algorithms working in the
environments which rely solely on forecasted demand with generation cost optimization objective.
The calculations carried out in this comparison, mainly determines the operating state of the system under consideration for a
particular generation pattern and load schedule (Optimal schedule
as obtained from the PSO based methodology developed) in specied solution space to maintain standards of operational condition
with high degree of reliability.

6. Implementation of the proposed method in IEEE 30 bus

system
The feasibility and effectiveness of the proposed algorithm has
been demonstrated in the modied IEEE 30 bus system shown in
Fig. B1 in the appendix. The summary of relevant data for the modied IEEE 30 bus system is presented in Tables B1B3 of the appendix. All the simulations have been performed in PSO environment.
A brief description of PSO based optimization technique has been
presented in the appendix. The parameter setting of the same is given in Table 1.
A owchart depicting implementation of the proposed method
in IEEE 30 bus system is presented in Fig. 5. In the proposed methodology the swarm intelligence based optimizer generates random
solution particles in the workspace bounded by the maximum and
minimum limit set by the generators and the consumers as described in Tables B2 and B3. In each step the optimizer generates
an optimum generation, load schedule with an objective of maximizing social welfare (described in (9)). The global best solution
is selected as optimal solution for the given set of operating
constraints.
6.1. Comparison of the proposed optimization technique with
traditional optimization
The examination of the feasibility and effectiveness of the algorithm proposed remains incomplete without testing its adoptability in the worst possible scenarios of the network. Four different
test-beds have been chosen to compare the proposed algorithm
with the traditional optimal load curtailment based OPF technique
[41]. In all the four cases the MCP of traditional algorithm has
been maintained at same level that of proposed method by

Table 2
Comparison of performance of traditional and proposed optimization technique with active, reactive loading stress and line contingency.
Observations

MCP ($/
MW h)

Traditional optimization

Proposed optimization

Line loss
(MW)

Max line ow
(MW)

Min bus voltage

(p.u)

Line loss
(MW)

Max line ow
(MW)

Min bus voltage

(p.u)

Under active loading stress

29%
3.3
32%
3.29
36%
3.29
40%
3.29

11.8
11.68
12.71
12.42

125.18
131.17
130.27
130.34

0.9742
0.9743
0.9740
0.9740

6.79
8.03
9.13
9.42

71.93
84.22
98.14
97.07

0.9782
0.9761
0.9800
0.9748

Under reactive loading stress

3.12%
3.22
7%
3.3
11%
3.28
42%
3.24

13.62
13.51
14.6
15.08

131.36
132.52
130.37
130.86

0.9440
0.9046
0.8606
0.8089

11.33
13
13.5
13.65

113
125
123
117.08

0.9600
0.9077
0.8667
0.8611

22.28
22
21.02
21.59

178.84
177.07
176.99
169.84

0.9687
0.9657
0.9660
0.9123

10.61
14.00
13.50
14.61

115.96
134.52
131.22
134.26

0.9776
0.9706
0.9723
0.9224

Under contingency
12
12, 24
12, 24, 2122
12, 24, 2122, 27
29

3.34
3.35
3.35
3.35

Test cases of comparison.

1. Active loading stress: The maximum loading is increased in different percentage of base case loading.
2. Reactive loading stress: The reactive loading stress is increased at different percentage of the base case reactive power loading at the weakest bus.
3. Under Contingency: (N  1), (N  2), (N  3), N  4) contingency cases have been studied for line outages.
Table 3
Forecasted percentage availability of intermittent sources.
Hour

Generator1 (%)

Generator2 (%)

Generator3 (%)

Generator4 (%)

Generator5 (%)

Generator6 (%)

h-1
h-2
h-3
h-4

100
100
100
100

100
100
100
100

100
60
40
20

100
90
90
90

100
100
100
100

100
50
30
10

Renewable energy sources may not be available in full capacity at every hour. Arbitrary shortages have been considered for different generating sources.

314

S. Chanda, A. De / Electrical Power and Energy Systems 58 (2014) 307318

Table 4
Performance comparison of traditional and proposed optimization technique during intermittency of generation.
Observations

MCP ($/MW h)

h-1
h-2
h-3
h-4

3.33
3.12
3.08
2.75

Traditional Optimization

Proposed Optimization

Line loss (MW)

Max line ow (MW)

Min bus voltage (p.u)

Line loss (MW)

Max line ow (MW)

Min bus voltage (p.u)

12.06
14.20
13.27
18.14

124.85
140.80
142.96
121.28

0.9740
0.9741
0.9746
0.9739

6.75
13.26
12.50
15.80

71.91
129
105.13
110.09

0.9763
0.9701
0.9767
0.9900

During intermittency of generation at different hours, the operating conditions have been found to improve by the application of the proposed optimization technique.

Fig. 6. Peak shaving of Demand during intermittency of Generation. Note: During

intermittency or shortage of generation the price increases. In the proposed
methodology the demand peak is effectively shaved in accord with the
requirement.

the present and future grids. However, their intermittent generation prole requires extensive load management for power market
equilibrium. With the intermittency of the generators as shown in
Table 3, the performance of the proposed algorithm has been observed to be remarkably superior in sustaining the operating conditions within safe limit (Table 4), owing to the fact that the
proposed algorithm not only manages generation for an optimal
cost (as in case of traditional method) but also efciently manages
loads to cater maximum demand at minimum cost in the most
optimal way maintaining all operating constraints.
The efciency of the proposed algorithm to self heal under the
alterations of operating conditions particularly during peak periods, when the available generation is less than requested demand
is depicted in Fig. 6, where the peak shaving attribute of the proposed algorithm has been demonstrated at selected buses.

6.2. Implementation of the proposed methodology with intermittent

energy sources

6.3. Conservative load curtailment attribute of the proposed

methodology

Promotion of distributed generation with renewable energy

sources as stated earlier is one of the nonnegotiable objectives of

Fig. 7 illustrates the comparison of load curtailment in the proposed method with reference to the ISO set limit for the case

Fig. 7. Comparison of curtailment by proposed method against their corresponding ISO set limit. Note: The ISO set limit is proposed to be determined by consumer price
responsiveness and maximum generation surplus. Load curtailment below this limit will benet all the stakeholders.

S. Chanda, A. De / Electrical Power and Energy Systems 58 (2014) 307318

315

Fig. 8. Convergence curve of maximization of the objective function.

studies conducted in Table 2 and 4. As discussed earlier, the curtailment should not always be price dependent but ISO should also
monitor the generation surplus in order to maintain a protable
curtailment operation in the power market. The ISO is constrained
to perform selective curtailment in order to keep the price of electricity constant or in other words to balance generation and demand at particular MCP. The results shown in Fig. 7AD depict
that, the proposed algorithm is capable of curtailing load at a more
conservative margin, well below the limits set by ISO.
The load curtailment limit has not fall below the ISO set curtailment limit in any of the test cases as in Fig. 7. The formula proposed in Eq. (3) is the optimal allowable load curtailment limit in
accordance with minimum generation surplus and curtailment
policy of ISO. This limit not only minimizes the generation surplus
for the benet of GENCOs but also retains the result within maximum curtailment limit for the welfare of consumers. As the results
are the outcome of Particle Swarm Optimization within the limit of
error gradient, the result has been tested to be optimal. Thus the
proposed methodology proves to maximize social welfare.
The claim of social welfare maximization may thus be assured
by the convergence of the objective function of optimization as
shown in Fig. 8.

7. Conclusion
The accessibility to smart metering infrastructure has brought
about radical changes in power grid operation and simultaneously
has thrown newer operational challenges for electric power system
operators, where a departure from traditional operations planning,
scheduling, and dispatch practices needs to be altered to take into
account tribulations like voltage instability, line congestion, line
loss intensication and payment cost minimization. In this pursuit
this paper presents a new state space based pricing model and a
methodology to illustrate optimal and efcient operations of Smart
Grid. The model effectively identies the state variables of MCP
and regulates the same utilizing PSO to reach prolic solutions
negotiating with generator characteristics, demand response, voltage stability, and congestion, curtailment and line loss limits. As

the market participants are aware of pricing signals, the proposed

model will enable them to regulate their consumption, production
or curtailment limits to cause an overall social welfare. The methodology determines the eigen values of the sensitivity matrix by
forming Jacobian to ensure power market stability and deploys a
novel curtailment technique to ensure the dispatch of minimum
power demand requested by the consumer for maintaining a high
degree of reliability. Simulation results convincingly demonstrate
that the combined operations of the proposed model and the
methodology mitigate network constraints while catering higher
demand levels and reducing the energy costs even under worst
possible states of the system operation. In comparison with the
conventional system operation which is based on only Genco
scheduling between maximum and minimum possible generation,
the proposed algorithm incorporates consumer scheduling with a
smart communication facility where not only GENCOs but also
consumers can participate to change their stand for their own
and social benet.
Appendix A. Overview of PSO
A brief overview of the PSO method is provided here which has
been used in this paper for solving the problem of maximization of
social welfare. Kennedy and Eberhart introduced the concept of
function Optimization by means of a particle swarm [49]. Suppose
the global optimum of an N-dimentional function is to be located.
The function may be mathematically represented as:

f x1 ; x2 ; x3 ; . . . xn f X

A1

where X is the search vector, which actually represents the set

of independent variables of the given function. The task is to nd
out such a X, that the function value f(X) is either a minimum or
a maximum denoted by f* in the search range. If the components
of X assume real values then the task is to locate a particular point
in the n-dimensional hyperspace which is a continuum of such
points.
PSO is a multi-agent parallel search technique. Particles are conceptual entities, which y through the multidimentional search

316

S. Chanda, A. De / Electrical Power and Energy Systems 58 (2014) 307318

space. At any particular instant, each particle has aposition and

velocity. The position vector of a particle with respect to the origin
of the search space represents a trial solution of the search problem. At the beginning a population of particles is initialised with
!

!

random positions marked by xi and random velocities v i . The population of such particles is called a Swarm S. A neighborhood
relation N is dened in the swarm. N determines for any two particles pi whether they are neighbors or not. Thus for any particle p,
neighborhood can be assigned as N(p), containing all the neighbors
of that particle. A popular version of PSO uses N = S for each particle. In this case, any particle has all the remaining particles in the
swarm in its neighborhood. Each particle has two state variables
viz., its current position and velocity as stated earlier. It is also
equiped with a small memory comprising its previous best position (One yielding the highest value of the tness function found
!
!
so far) pt that is personal best experience and the best pt of
!
p 2 Np : gt that is the best position so far in the neighborhood
!
of the particle. When we set N(p) = S, gt is referred to as the
globally best particle in the entire swarm. The PSO scheme has
the following algorithmic parameters:
!

 Vmax or maximum velocity which restricts v i within the interval

[Vmax, Vmax]
 An inertial weight factor x.
 Two uniformly distributed random numbers U1 and U2 that
!
!
respectively determine the inuence of pt and gt on the
velocity update formula
 Two constant multiplier terms c1 and c2 known as selfcondence and swarm condence, respectively.
!
!
! ! !
Initially the settings for pt and gt are p0 g0 x0 for
all particles. Once the particles are initialized, an iterative optimization process begins, where the positions and velocities of all
particles are altered by the following recursive equations. The
equations are presented for the dth dimension of the position
and velocity of the ith particle.

mid t 1 x  v id t c1  U1 pid t  xid t c2  U2 g id t  xid t

A2
xid t 1 xid t v id t 1

A3

The rst term in the velocity updating formula represents the

initial velocity of the particle. x is the inertia factor. Venter
and Sobeiski [50] termed c1 as self condence and c2 as swarm
condence. These terminologies provide an insight from a
sociological standpoint. Since the coefcient c1 has a contribution
towards self-exploration, we regard it as particles self-condence.
On the other hand the particle c2 has a contribution towards
motion of the particles towards a global direction, which takes into
account the motion of all the particles in the preceding program
iteration, naturally its denition, as Swarm Condence is apparent. U1 and U2 stand for a uniformly distributed random number
in the interval [0, 1] After having calculated the velocities and
position for the next step t + 1, the rst iteration of the algorithm
is completed.

Appendix B. Description of the system with generator and

consumer cost characteristics
Descriptions of the system with its generator and consumer
cost characteristics are given in the following tables and the Single
Line Diagram (SLD) is portrayed in Fig. B1 (see Tables B1B3).

Table B1
Description of IEEE 30 bus system.
SL No

1
2
3

IEEE 30 bus system

Variables

Adopted system

Branches
Generators
Total demand (MW)

41
6
365 (In the base case)

All the case studies have been executed in this system.

Fig. B1. SLD of IEEE 30 bus system. Note: Single Line Diagram of the test system with positions of generations and loads.

S. Chanda, A. De / Electrical Power and Energy Systems 58 (2014) 307318

Table B2
Generator cost coefcients.
Bus no

1
2
5
8
11
13

Real power output limit

in MW

Cost co-efcient

Min

Max

A (US$/MW2)

B (US$/MW)

C (US$)

50
20
15
10
10
12

200
80
50
35
30
40

0.00375
0.01750
0.06250
0.00834
0.02500
0.02500

2.00
1.75
1.00
3.25
3.00
3.00

0
0
0
0
0
0

This is the base case generator characteristics and relevant parameters.

Table B3
Coefcients of consumer cost benet function.
Bus no

2
3
4
5
7
8
10
12
14
15
16
17
18
19
20
21
23
24
26
29
30

Real power demand

limits in MW

Cost co-efcient

Min

Max

A (US$/MW2)

B (US$/MW)

C (US$)

21.7
2.4
7.6
94.2
22.8
30
5.8
11.2
6.2
8.2
3.5
9
3.2
9.5
2.2
17.5
3.2
8.7
3.5
2.4
10.6

31.7
5
10
110
40
45
10
15
12
10
10
15
5
12
5
20
6
10
5
4
15

0.1
0.15
0.25
0.2
0.075
0.1
0.2
0.1
0.15
0.15
0.15
0.05
0.1
0.25
0.25
0.2
0.1
0.2
0.3
0.4
0.2

87.2
90
85.5
87
70
90
70
50
80
60
70
65
90
68
90
80
60
80
90
95
95

0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0

Base case parameters of consumer cost benet function.

References
[1] Gang L, Debraj D, Wen-Zhan S. Smart Grid Lab: a laboratory-based smart grid
test bed. In: The 1st IEEE international conference on smart grid
communications, 2010; Gaithersburg, USA; 2010. p. 1438.
[2] Farrokh R, Ali I. Demand response as a market resource under the smart grid
paradigm. IEEE Trans Smart Grid 2010;1(1).
[3] Ali Ipakchi. Demand side and Distributed Resource Management a
transactive solution, IEEE OATI Power and Energy, July 22nd, 2011.
[4] Natarajan Ramsamy. Computer-aided system analysis. New York,
Basel: Marcel Dekkar INC.; 2002.
[5] Sara D, Amir S, Paul S, Moses S, Masoum Mohammad AS. Real-time
coordination of plug-in electric vehicle charging in smart grids to minimize
power losses and improve voltage prole. IEEE Trans Smart Grid 2011;2(3).
[6] Yunzhi C, Wei-Jen L, Shun-H, Huang John A. Dynamic parameter identication
of generators for smart grid development. In: Power and energy society
general meeting, IEEE, 2429 July, 2011, San Diego; 2011. p. 17.
[7] Fangxing L, Wei Q, Hongbin S, Hui W, Jianhui W, Yan X. Smart transmission
grid: vision and framework. IEEE Trans Smart Grid 2010;1(2).
[8] Jose M, Nelson M, Ilya R. Demand response and distribution grid operations:
opportunities and challenges. IEEE Trans Smart Grid 2010;1(2).
[9] Amir-Hamed M, Wong Vincent WS, Juri J, Robert S, Alberto L. Autonomous
demand-side management based on game-theoretic energy consumption
scheduling for the future smart grid. IEEE Trans Smart Grid 2010;1(3).
[10] Pouyan K, Hassan M, Hassan A. Load prole reformation through demand
response programs using smart grid modern electric power systems 2010,
Wroclaw, Poland MEPS10; 2010.
[11] David Sun, Laurent Schmitt. Advanced power system operations with smart
grid technologies, 2010 IEEE PES Panel Session; 2010.
[12] In-Ho C, Joung-Han Lee. Development of smart controller with demand
response for AMI connection. In: International conference on control,
automation and systems 2010 October, 2730, 2010, Korea; 2010. p. 7525.
[13] Hammons TJ. Integrating renewable energy sources into European grids. Int J
Electr Power Energy Syst 2008;30(8):46275.

317

[14] Yann R, Seddik B, Franck B, Stephane P. Optimal power ow management for

grid connected PV systems with batteries. IEEE Trans Sustain Energy
2011;2(3).
[15] Warsono, King DJ, zveren CS, Bradley DA. Economic load dispatch
optimization of renewable energy in power system using genetic algorithm,
Power Tech; 2007.
[16] Wong YS, Lai LL, KT, Gao Shuang, Chau. Stationary and mobile battery energy
storage systems for smart grids., Electric Utility Deregulation and
Restructuring and Power Technologies (DRPT). In: 4th international
conference on, 69 July, 2011, Weihai, Shandong; 2011. p. 16.
[17] Martin G, Gran A. Optimal power ow of multiple energy carriers. IEEE Trans
Power Syst 2007;22(1).
[18] Carlo C, Costantino C, Pierluigi S. Combined operations of renewable energy
systems and responsive demand in a smart grid. IEEE Trans Sustain Energy
2011;2(4).
[19] Kazem Zare, Sheikh-el-Eslami MK, Parsa Moghaddam Tarbiat M. Large
consumer decision making to cost reduction in real time power market. In:
42nd International conference on power engineering. UPEC; 2007. p. 8992.
46 September.
[20] Satish S, Rohit B, Padhy NP, Gupta HO. Payment cost minimization auction in
electricity markets power and energy society general meeting, IEEE, 2429
July, 2011, San Diego, CA; 2011. p. 16.
[21] Shahram J, Nahid A. Factoring the price elasticity of demand in the optimal
power ow. In: International power engineering conference, IPEC 2007,
Singapore; 207. p. 127.
[22] Undan Jason Alfons. Optimisation of demand-side bidding. In: 45th annual
conference, University of Auckland, Auckland, New Zealand, 2930 November,
2010. p. 7281.
[23] Bhuvaneswari R, Sanjeev KS, Chris S, David A. An intelligent auction scheme for
smart grid market using a hybrid immune algorithm. IEEE Trans Industr
Electron 2011;58(10).
[24] Centolella Paul, Ott Andrew. The integration of price responsive demand into
PJM
wholesale
power
markets
and
system
operations.
<www.hks.harvard.edu>.
[25] Arman K, Anuradha A. The effect of a smart meter on congestion and stability
in a power market. In: Proceedings of 49th IEEE conference on decision and
control, Atlanta, GA, USA, December 1517, 2010. p. 1949.
[26] Pedram S, Amir-Hamed M, Robert S, Wong Vincent WS, Jatskevich Juri.
Optimal real-time pricing algorithm based on utility maximization for smart
grid. In: 1st IEEE international conference on smart grid communication,
Gaithersburg, USA, 46 October, 2010.
[27] Sortomme E, El-Sharkawi. Optimal power ow for a system of micro-grids
with controllable loads and battery storage. In: Power systems conference and
exposition, PSCE 09. p. 15.
[28] Majid O, Gerd B, Hasan S, Mohammad S. Market-based transmission expansion
planning. IEEE Trans Power Syst 2004;19(4).
[29] Fernandes Thelma SP, Lenzi JR, Mikilita Miguel A. Load shedding strategies
using optimal load ow with relaxation of restrictions. IEEE Trans Power Syst
2008;23(2).
[30] Huang Garng M, Nair Nirmal-Kumar C. An OPF based algorithm to evaluate
load curtailment incorporating voltage stability margin criterion.
<www.pserc.wisc.edu>.
[31] Kresen Kjetil F, Egil H. A joint state-space model for electricity spot and
futures prices, Norwegian Computing Center; 2002.
[32] Mihaela M. Energy futures prices: term structure models with kalman lter
estimation. Taylor and Francis Group; 2010.
[33] Ofuji K, Kanemoto S. Price forecasting of japan electric power exchange using
time-varying AR model. In: Proceedings of the 14th international conference on
intelligent system applications to power systems, ISAP 2007, November 48.
[34] Shiva S, Mehdi P, Mohsen S. Short-term electricity price forecasting in
deregulated markets using articial neural network. In: Proceedings of
international conference on industrial engineering and operations
management, Kuala Lumpur, Malaysia, January 2224, 2011.
[35] Burgery Markus, Klarz Bernhard, Muller Alfred, Schindlmayry Gero. A spot
market model for pricing derivatives in electricity markets. J Quant Finance
2010;5(1).
[36] Alicia Mateo G, Antonio Muoz San R, Javier G. Modeling and forecasting
electricity prices with input/output hidden Markov models. IEEE Trans Power
Syst 2005;20(1).
[37] Ramteen S, Walter S. Evaluating the impacts of real-time pricing on the usage
of wind generation. IEEE Trans Power Syst 2009;24(2).
[38] Manuela B, Brockwell Anthony E, Duane J. A dynamic supply-demand model
for electricity prices. <www.stat.cmu.edu>.
[39] Aggarwal SK, Saini LM, Kumar A. Electricity price forecasting in deregulated
markets: a review and evaluation. Int J Electr Power Energy Syst
2009;31(1):1322.
[40] Florin C, Mevludin G, Damien E, Louis W. Interior-point based algorithms for
the solution of optimal power ow problems. Electric Power Syst Res
2007:50817. Elsevier.
[41] Rueda-Medina AC, Padilha-Feltrin A. Pricing of reactive power support
provided by distributed generators in transmission systems, 2011 IEEE
Power Tech.
[42] Bomparda Ettore, Maa Yuchao, Napolia Roberto, Abrateb Graziano, Ragazzib
Elena. The impacts of price responsiveness on strategic equilibrium in
competitive electricity markets. Int J Electr Power Energy Syst
2007;29(5):397407.

318

S. Chanda, A. De / Electrical Power and Energy Systems 58 (2014) 307318

[43] Losi Arturo. Trade curtailment schemes for security control of transmission
network in a deregulated environment. Int J Electr Power Energy Syst
2002;24(1):917.
[44] Juan Carlos M, Pablo Cuervo F. Transmission loss allocation through equivalent
bilateral exchanges and economical analysis. IEEE Trans Power Syst
2005;20(4).
[45] Tripathy M, Mishra S. Bacteria foraging-based solution to optimize both real
power loss and voltage stability limit. IEEE Trans Power Syst 2007;22(1).
[46] Sandip C, Abhinandan D. Improvement of economic aspect of power network
congestion management by swarm intelligence based multi-objective
algorithm. Int J Eng Sci Technol 2011;3(5).

[47] Coelho LdSa, Lee C-Sb. Solving economic load dispatch problems in power
systems using chaotic and Gaussian particle swarm optimisation approaches.
Int J Electr Power Energy Syst 2008;30(5):297307.
[48] Fanga Debin, Wua Jingfang, Tangb Dawei. A double auction model for
competitive generators and large consumers considering power transmission
cost. Int J Electr Power Energy Syst 2012;43(1):8808.
[49] Kennedy J, Eberhart R. Particle swarm optimisation. In: The proceedings of
IEEE international conference on neural networks; 1995. p. 19428.
[50] Venter G, Sobieszczanski-Sobieski J. Particle swarm optimization. AIAA J
2003;41(8):15839.