## Are you sure?

This action might not be possible to undo. Are you sure you want to continue?

CHAPTER 4

COMPUTATION OF AVAILABLE TRANSFER

CAPABILITY

4.1 INTRODUCTION

As discussed in the previous chapters, load flow study plays a vital

role in operation and control of modern power systems. It can also be

used for the computation of transfer capability. In a deregulated system,

the information about the transfer capability will help the energy

marketers in reserving the transmission services. For secured and

economic supply of electric power, long distance bulk power transfers are

essential, but the power transfer capability of a power system is limited.

To operate the power systems safely and to gain the advantages of bulk

power transfers, computations of transfer capability is essential. Transfer

capability plays a vital role in liberalized electricity market. All the

transmission lines are utilized significantly below their physical limits

due to various constraints. By increasing the transfer capability the

economic value of transmission lines can be improved and also there will

be an increase in overall efficiency as more energy trading can take place

between the competing regions operating with different price structures.

The power system should be planned and operated such that these

power transfers are within the limits of the system transfer capability.

112

Transfer capability of a power system is defined as the maximum power

that can be transferred from one area to another area.

4.2 GENERAL MOTIVATION

In open access transmission system, the transmission network

owners are required to provide unbundled services to support power

transactions and to maintain reliable operation of the networks. In a

liberalized electricity markets, to enforce the open access policy North

American Energy Reliability Council (NERC) in conjunction with Federal

Energy Regulatory Commission (FERC) defined the term available

transfer capability (ATC) to be posted in open access same time

information system (OASIS) to inform all the energy market participants

of a power system. This information is required to be made available on

hourly or daily basis. The two major challenges that make the task of

ATC calculation of a nonlinear power system more challenging are

computing speed and accuracy due to static and dynamic security

constraints.

Deregulation of power market has imposed great impact on the

utility industry. For smooth transaction of power between the areas or

paths new technologies and computation methods are urgently needed.

Transfer capability of a power system also indicates how much inter area

power transfer can be increased without system security violations. The

vital information required for the planning and operation of the large

power systems can be obtained from these transfer capability

113

calculations. These details provide system bottlenecks to the planners

and the limits of the power transfers to the system operators. The risk of

overloads, equipment damage and unexpected blackouts can be reduced

by repeated estimation of these transfer capabilities. These calculations

also help to determine the quantity of lost generation that can be

replaced by potential reserves and limiting constraints in each

circumstance.

Due to the deregulation of power systems the power transfers are

increasing which is necessary for a competitive market for electric power.

There is a strong economic incentive to improve the accuracy and

effectiveness of the transfer capability computations for the use by the

power system operators and the power markets.

Transfer capability can be computed using different methods and

these computations are evolving. The methods used at present are

oversimplified and they do not consider the effects such as

nonlinearities, system policies, interactions between the power transfers

and loop flows. Under open access environment of power system actual

evaluation of ATC is very much essential and a practical software

package for computing ATC will be an important tool for all transmission

providers. In recent years a significant progress has been made in

developing such a tool, one remaining major challenge is to determine

ATC accurately under varying load conditions considering static and

dynamic security limits. The main objective of this research is to improve

114

the accuracy and incorporate the pragmatism in transfer capability

computations.

4.3 DEFINITIONS

According to the report approved by NERC the definitions of ATC

are as follows. Fig. 4.1 [64] represents the various terms.

Total Transfer Capability (TTC): It is defined as the quantity of electric

power that can be transferred over the interconnected transmission path

reliably without violating the predefined set of conditions of the system.

Available Transfer Capability (ATC): It is a measure of the transfer

capability remaining in the physical transmission network for further

commercial activity over and above already committed uses.

Mathematically, ATC is defined as the Total Transfer Capability (TTC)

less the Transmission Reliability Margin (TRM), less the sum of existing

transmission commitments (which includes retail customer service) and

the Capacity Benefit Margin (CBM).

Transmission Reliability Margin (TRM): It is defined as that amount of

transmission transfer capability necessary to ensure that the

interconnected transmission network is secure under a reasonable range

of uncertainties in system conditions.

Capacity Benefit Margin (CBM) is defined as that amount of

transmission transfer capability reserved by load serving entities to

ensure access to generation from interconnected systems to meet

generation reliability requirements.

115

Fig. 4.1 Transfer capabilities and related terms

4.4 TOTAL TRANSFER CAPABILITY

Total Transfer Capability (TTC) is defined as the amount of

electric power that can be transferred over the interconnected

transmission network in a reliable manner while meeting all of a specific

set of defined pre- and post-contingency system conditions. The various

constraints that limit Total Transfer Capability may be the physical and

electrical characteristics of the systems including thermal, voltage, and

stability limits as shown in Figure 4.2.

TTC = Minimum of {Thermal Limit, Voltage Limit, Stability Limit}

TTC is an important parameter that indicates how much power

transfer can take place without compromising the system security. It

provides vital information for the planners, operators and marketers. The

accurate TTC calculation is essential to ensure that power system can

116

operate without reliability risks. A number of methods exist for

calculation of TTC and excessive conservative transfer capability may

limit the transfer unnecessarily and also lead to inefficient operation of

power system. In other words TTC is the maximum value of power

transfer between the paths without any limit violations, with or without

contingency. The objective can be defined as the determination of

maximum real power transfer between the utilities.

Fig. 4.2 Limits of total transfer capability

Transfer capability can be calculated as follows

- Establish a base case which is assumed to be a secured operating

condition so that all line flows and bus voltage magnitudes lie

within their operating limits.

- Specify a transfer which includes source and sink powers.

117

- Establish a solved case by increasing source and sink powers until

there is a limit violation.

- The transfer margin is the difference between the limiting case

transfer and the base case transfer.

4.4.1 PURPOSE OF TRANSFER CAPABILITY COMPUTATIONS

The need for transfer capability computations:

- Estimation of TTC can be used as a rough indicator of relative

system indicator.

- It can be used for comparing the relative merits of planned

transmission betterments.

- To improve reliability and economic efficiency of the power

markets.

- For providing a quantitative basis for assessing transmission

reservations to facilitate energy markets.

4.5 METHODS OF TRANSFER CAPABILITY CALCULATION

A number of methods have been proposed in the literature. These

methods are classified as i) continuation power flow (CP FLOW) [68]

based methods ii) optimal power flow (OPF) based methods and iii) Linear

approximation methods. Various methods of calculating transfer

capability are explained below.

118

- Continuation methods in which the transfer capability is computed

using a software model called continuation. This process requires a

series of load flow solutions to be solved and tested for limits.

- Optimal power flow approach is another method to formulate an

optimization problem. Equality constraints obtained from power

flow are used in this approach.

- Linear methods use PTDFs (power flow distribution factors) to

express the percentage of power transfer that occurs on a

transmission path.

4.5.1 CONTINUATION POWER FLOW APPROACH

Continuation power flow method is a comprehensive tool for tracing

the steady state behavior of the power system due to parametric variation

[84]. The parameters which are varied include bus real and/or reactive

loads, area real and /or reactive loads and real power generations at

generator or P-V buses. Continuation methods are also known as curve

tracing or path following which are used to trace solution curves for

general non-linear algebraic equations with a parametric variation. These

methods have four basic elements:

- Parameterization: This is a mathematical way of identifying each

solution for quantifying next solution or previous solution.

- Predictor: To find an approximate point for the next solution.

Tangent or secant method is used for this purpose.

119

- Corrector: To correct error in an approximation produced by the

predictor before it accumulates.

- Step size control: To adapt the step length for shaping the traced

solution curve.

This method is based on the static model of the power system.

Basically it calculates the successive equilibrium points for the Equation

4.1 assuming slow variation of the load parameter (λ) which represents

the increment in load demand and power supplied by the system

generators.

) , ( 0 ì y g = (4.1)

Rewriting the load flow equations the real and reactive power can

be represented by equation 4.2 and 4.3

( )

¿

e

+ = ÷

i j

ij ij ij j i Li Gi

ij Sin B Cos G V V P P ) ) ( ) ( u u ì ì (4.2)

( )

¿

e

+ = ÷

i j

ij ij ij j i Li Gi

ij Cos B Sin G V V Q Q ) ) ( ) ( u u ì ì (4.3)

The increments of the generator active power and demands are

given by

( )

Gi Gio Gi

K P P ì ì + = 1 ) ( (4.4)

( )

PLi Lio Li

K P P ì ì + = 1 ) ( (4.5)

( )

QLi Lio Li

K Q Q ì ì + = 1 ) ( (4.6)

In Equation 4.5 and Equation 4.6 PLio and QLio represent the base

case (λ = 0) active and reactive powers of i

th

bus.

120

PGio

in equation 4.5

is the base case active power supplied by the

generator of i

th

bus.

KPLi and KQLi are coefficients defining the load power factors of the

i

th

bus.

KGi is a coefficient defining the generator participation factor in the

i

th

bus for certain loading level λ.

Above equations can be written in a compact form as in equation

4.7.

0 . ) ( ) , ( = + = d y G y g ì ì (4.7)

In above equation 4.7 d represents a vector indicating the direction

of the active power increment supplied by the generators and reactive

power increment consumed by the loads.

Successive solution of the above equation for different values of

loading parameter λ is obtained by continuation power flow through a

predictor and corrector [65, 66, 68] as shown in Figure 4.3

121

Fig. 4.3 Predictor and corrector method of CP FLOW

4.5.2 OPTIMAL POWER FLOW APPROACH

The application of Optimal Power Flow (OPF) in power system

congestion management has been studied by some researchers

[69][70][71][72]. In the mean time, TTC calculation by OPF approach has

been proposed since 1999 [73][74][75]. The basic concept of OPF

approach is formulating the TTC calculation as an optimization problem,

with equity constraints of power flow, inequality constraints from basic

operation and equipment limits to more detailed approximation of

transient stability security requirements. The objective function,

obviously, is the maximum power flow on the specified transmission

route. To determine the total transfer capability the objective is to

122

maximize the power transfer between the two areas subjected to the

conditions that there is no voltage or thermal or stability limit violations.

Total transfer capability problem formulation can be explained as follows.

Maximize

¿

e

=

i j

kj i

P P (4.8)

Subjected to

0 ) ( = ÷ + ÷

¿

ei j

j i ij ij j i i

Cos Y V V P o o u (4.9)

0 ) ( = ÷ + ÷

¿

ei j

j i ij ij j i i

Sin Y V V Q o o u (4.10)

max min g g g

P P P s s (4.11)

max min g g g

Q Q Q s s (4.12)

max ij ij

S S s (4.13)

max min i i i

V V V s s (4.14)

4.5.3 REPEATED POWER FLOW APPROACH

Repeated power flow approach starts from a base case, and

repeatedly solves the power flow equations each time increasing the

power transfer by a small increment until an operation limit is reached

[76]. The advantage of this approach is its simple implementation and

the ease to take security constraints into consideration. In this

dissertation, this method is adopted to solve TTC problem.

123

4.6 ALGORITHM FOR REPEATED POWER FLOW METHOD

In this research work, it is proposed to utilize the repeated power flow

(RPF) method [67] for the calculation of transfer capabilities due to the

ease of implementation. This method involves the solution of a base case,

which is the initial system conditions, and then increasing the transfer.

After each increase, another load flow is solved and the security

constraints tested. Flow chart for the algorithm is given in Annexure IV.

The computational procedure of this approach is as follows:

i. Establish and solve for a base case

ii. Select a transfer case

iii. Solve for the transfer case

iv. Increase step size if transfer is successful

v. Decrease step size if transfer is unsuccessful

vi. Repeat the procedure until minimum step size reached

4.7 AVAILABLE TRANSFER CAPABILITY [ATC]

Available transfer capability computations are essential for

successful implementation of electric power deregulation where power

producers and customers share a common transmission network for

wheeling power from the point of generation to the point of consumption.

The available transfer capability indicates the amount which inter-area

bulk power transfers can be increased without compromising system

security. The value used for available transfer capability affects both

124

system security and the profits made in bulk power transactions.

Moreover, market participants can have conflicting interests in a higher

or a lower available transfer capability. Thus under deregulation, there is

increasing motivation for defensible calculations of available transfer

capability.

4.7.1 IMPORTANCE OF ATC ASSESSMENT

In this present open access or deregulated environment all

the participants (producers and buyers of electrical energy) desire to

produce or consume large amounts of energy and may force the

transmission system to operate beyond one or more transfer limits. This

kind of operation leads to congestion of the system. Therefore accurate

determination of available transfer capability is essential to ensure the

system security and reliability while serving a wide range of bilateral and

multilateral power transactions. The following reasons show the need of

ATC computations

- It gives the amount of maximum additional power transfer

between the specified interfaces.

- It ensures the secure operation of the system.

- On the basis of ATC computation firm and non-firm reservation

of transmission services can be made.

- In a deregulated open access market it can be used as tool for

transmission pricing.

125

- The limits or binding constraints for ATC can be used in power

system planning and network expansion.

4.8 METHODS OF CALCULATING ATC

As explained above the following methods are used for

determination of ATC

1. Load flow / continuous power flow (CPF) / Repeated Power Flow

(RPF) methods.

2. Optimization based methods

3. Network sensitivity method

Full AC power flow method is the most accurate method for

computation of ATC but its complexity can blot out relationships. The

following linear methods are used for calculation of ATC. Transfer

capabilities can be estimated with simple power system models such as

the DC load flow approximation. A DC model may be preferable to an AC

model particularly in circumstances where the extra data for an AC

model is unavailable or very uncertain, such as the case of very long time

frame analysis.

4.9 COMPUTATION OF ATC USING LINEAR METHODS

In this section linear methods are explained briefly for

determination of available transfer capability followed by the merits and

demerits of these approaches.

126

4.9.1 DC POWER FLOW METHOD

This model assumes that the voltage magnitudes are constant

and only the angles of the complex bus voltage vary. It is also assumed

that the transmission line has no resistance and therefore no losses. In

addition to the speed of computation this method has other useful

properties like linearity and super position.

The following Equation 4.15 explain DC power flow method

( )

j i

ij

ij

x

P u u ÷ =

1

(4.15)

Where

xij line inductive reactance

θi phase angle at bus i

θj phase angle at bus j

The total power flowing in to bus i is the algebraic sum of generation and

demand at the bus called bus power injection given by Equation 4.16.

( )

¿ ¿

÷ = =

j

j i

ij j

ij i

x

P P u u

1

(4.16)

This can be expressed in a matrix form by Equation 4.17

(

(

(

¸

(

¸

(

(

(

¸

(

¸

B =

(

(

(

¸

(

¸

n

x

n

P

P

u

u

. .

1 1

(4.17)

Where the elements of susceptance matrix Bx are functions of line

reactance xij

127

Phase angle of one of the buses is set to zero and eliminating the row and

column from Bx matrix, the reactance matrix can be obtained by

inversion.

The phase angles are obtained as a function of bus injections

as shown in Equation 4.18

(

(

(

¸

(

¸

(

(

(

¸

(

¸

X =

(

(

(

¸

(

¸

÷ ÷ 1

1

1

1

. .

n n

P

P

u

u

(4.18)

Line flows Pij are obtained from Equation 4.16 to compare with the line

MW limits.

4.9.2 POWER TRANSFER DISTRIBUTION FACTORS

According to power flow point of view power injected in to the

system at a point by generator is extracted by a load at another point

which is known as transaction. Transaction can be found from the linear

property of DC load flow model using sensitivity factor PTDFs [78, 79].

Power transfer distribution factor (PTDF) is defined as the coefficient of

the linear relationship between the amount of a transaction and the flow

on a line. As it relates the amount of one change i.e. transaction amount

to another change i.e. the lone flow. PTDF is the fraction of amount of a

transaction from one zone to another over a specified transmission line.

PTDFij,mn represents the fraction of a transaction from m zone to zone n

that flows over a transmission line connecting zone i to zone j.

128

The Power Transfer Distribution Factor (PTDF) is given by

Equation 4.19

ij

jn in jm im

mn ij

x

X X X X

PTDF

+ ÷ ÷

=

,

(4.19)

Where

xij transmission line reactance connecting zone i and

zone j

Xim entry in the i

th

row and m

th

column of the bus

reactance matrix X

The change in line flow associated with a new transaction is then given

by Equation 4.20.

New

mn mn ij

New

ij

P PTDF P · = A

,

(4.20)

4.9.3 LINE OUTAGE DISTRIBUTION FACTORS (LODF)

The simple but most inaccurate method used to calculate the

effect of line outage is DC power flow contingency. The speed of

computation of this method can be improved by using another sensitivity

called line outage distribution factor (LODF). When a line outage occurs

in a system the power flowing on that line is redistributed on to the

remaining lines in the system. This redistribution is measured by LODF

and the fraction of power flowing on the line from zone r to zone s before

it is removed which now flows over a line from zone i to zone j is given by

LODFij,rs.

129

The change in power flow is given by

rs rs ij rs ij

P LODF P · = A

, ,

(4.21)

( )

( ) | |

rs ss rr rs rs

js jr is ir

ij ij

rs rs

rs ij

X X X x N

X X X X

x N

x N

LODF

2 .

.

.

.

,

÷ + ÷

+ ÷ ÷

= (4.22)

Where

xij line reactance connecting zone i and j

Xir entry in the i

th

row and r

th

column of the bus reactance matrix

X

Nij number of circuits connecting zone i and zone j

4.10 MAJOR ADVANTAGES & DISADVANTAGES OF THE EARLIER

METHODS:

The advantages and disadvantages of various methods of transfer

capability computation are as follows:

4.10.1 DC APPROXIMATION

The DC approximation is preferred for these reasons:

- Fast computation - no iteration.

- Thermal limits, MW limits are considered.

- Network topology handled with simple and linear methods.

- Good approximation over large range of conditions.

Minimum data is required.

But DC approximation is poor for these reasons:

- It cannot identify voltage limits

130

- It is not accurate when VAR flow and when voltage

deviations are considerable.

- Over use of linear superposition increases errors

4.10.2 SENSITIVITY FACTORS

One of the earliest solutions proposed for Available Transfer

capability (ATC) is sensitivity analysis. These sensitivity factors are based

on linear incremental power flow, which are very simple to define and

calculate.

This approach has a major disadvantage that they do not

consider the nonlinear effects of voltage and reactive power. Moreover the

methods based on Power Transfer Distribution Factors (PTDFs) and Line

Outage Distribution Factors (LODFs) can be used to estimate the ATC of

the cases nearer to the base case from which these factors are derived.

4.10.3 OPTIMAL POWER FLOW METHOD

Optimal Power Flow (OPF) and Security Constrained Optimal

Power Flow (SCOPF) are widely used to determine ATC in power corridors

of the system. However these optimization methods are suitable in case

of open access system where there is a possibility of power transactions

occurring from any point to any point.

4.10.4 CONTINUATION POWER FLOW

As discussed earlier this method is initially used for finding

maximum loadability point, however its applications are extended to

131

determine TTC and ATC. The computational time increases due to its

complexity, when contingencies are included.

4.11 ATC COMPUTATION USING COMPLEX VALUED NEURAL

NETWORKS

The practical computations of transfer capability are evolving. The

computations presently being implemented are usually oversimplified

and in many cases do not take sufficient account of effects such as

interactions between power transfers, loop flows, non-linearities,

operating policies and most importantly voltage collapse blackouts. The

transfer capability estimation method proposed by X. Luo et. al. [77] uses

a Multi Layered Feed Forward Neural Networks which is capable of

reflecting variations in load levels and in the status of generation and

transmission lines. The transfer capability was estimated accurately of

between system areas with variations in load levels, in the status of

generation, and in the status of lines. In this work Quick prop algorithm

is used to train the neural network.

The goal of the methods described here is to improve the accuracy

and realism of transfer capability computations. The artificial neural

network approach reported in the methods proposed require a large

input vector for bilateral transaction, so it has oversimplified the

determination of ATC by limiting it to a special case of power transfer to

a single area from all of the remaining areas. Therefore, this method is

unable to track down the bus-to-bus transactions, which is the true

132

spirit of deregulation and nonlinearities of the systems are also not

considered.

4.11.1 ASSUMPTIONS

From the formulations discussed in section 3.4 a Complex Valued

Neural Network approach is implemented for ATC computation.

The following assumptions are made while calculating the ATC.

a) The base case power flow of the system is feasible and

corresponds to a stable operating point;

b) The load and generation patterns vary very slowly; and

c) TTC calculation is in the steady state analysis domain.

4.12 COMPUTATION OF ATC FOR A 9-BUS SYSTEM

For computing Available Transfer Capability using the proposed

approach a 9 bus system [120] is considered (Annexure – 1). It has 3

generator buses and the number of transmission lines is 9 as shown in

Figure 4.4.

This method involves the solution of a base case, which is the

initial system conditions, and then increasing the transfer. After each

increase, another load flow is solved and the security constraints tested.

Voltage constraint is taken

133

Fig. 4.4 Nine bus system.

The system divided in to two areas. Area 1 comprises of bus 1,bus

3,bus 4, bus 5 and bus 6 where as area 2 comprises of the remaining

buses as shown in Fig. 4.4 [84]. The tie lines between the two areas are

line 4-9 and line 6-7. These lines transmit the power from one

geographical area to other area.

134

4.12.1 LEARNING METHODOLOGY

Learning methodology uses complex back propagation algorithm

explained in Chapter 3. Repeated power flow method is utilized to

generate training patterns using Newton Raphson load flow method.

Repeated power flow (RPF) method is proposed for calculating ATC due to

the ease of implementation. In this method the available transfer

capability (ATC) from one bus/area/zone to another bus/area/zone can

be found by varying the amount of transaction until one or more line

flows in the transmission system considered or a bus voltage reaching

the limiting value. The proposed new methodology is based on the full Ac

load flow and strong generalization capability of complex neural network

offers a great potential for real time evaluation of ATC incorporating load

variation, effect of reactive power loss of the system.

For calculating ATC using Repeated Power Flow (RPF) method the

following choices are made:

Establish a secure, solved base case.

Specify a transfer including source and sink assumptions.

Identify the branch flows influencing the ATC of selected branch

appreciably.

Identify the line outages having significant influence on the above

said branch power flows.

135

Generate numerous training data sets involving above said power

flows and line outages.

The transfer margin is the difference between the transfer at the

base case and the limiting case.

The flow chart for the above algorithm is shown in Annexure –IV.

The calculation of ATC is done by using the Newton Raphson load flow

solution to compute the power flow of each transfer case. This method is

less prone to divergence with ill-conditioned problems. And also the

number of iterations required is independent of the system size. The

loads at bus number 7 and 9 are increased simultaneously and the

transfers from area 1 to area 2 are obtained. The total transfer capability

is the sum of transfers through the interconnecting lines i.e. line joining

buses 4 and 9, buses 6 and 7. The available transfer capability is given

by Equation 4.23

ATC = TTC – base case transfer (4.23)

Satisfying the following system operating conditions

¿

= ÷ + ÷

N j

i j ij ij j i

Cos Y V V Pi

c

o o u 0 ) ( (4.24)

¿

= ÷ + ÷

N j

i j ij ij j i

Sin Y V V Qi

c

o o u 0 ) ( (4.25)

max min g g g

P P P s s (4.26)

max min g g g

Q Q Q s s (4.27)

max , min , i i i

V V V s s (4.28)

136

0 50 100 150 200 250

0.65

0.7

0.75

0.8

0.85

0.9

0.95

1

1.05

Load in area 2 (MW)

V

o

l

t

a

g

e

Bus 7

Bus 9

Load at Bus 5 is 90+j30

Fig. 4.5 P-V curves with load 90+j30

The above Figure 4.5 shows the P-V curve obtained from

repeated power flow. The load in area 1 that is at bus 5 is maintained as

90+j30 where as the load in area 2 that is at bus 7 and bus 9 is varied

slowly. It is assumed that bus 7 and bus 9 are critical buses compared

to other buses and the voltage collapse points of these buses are shown

with respect to the total load in area 2. From the results obtained from

the repeated power flow it is observed that the voltage at these two buses

reached the nose point compared to other buses.

137

0 20 40 60 80 100 120 140 160 180 200

0.65

0.7

0.75

0.8

0.85

0.9

0.95

1

1.05

Load in area 2 (MW)

V

o

l

t

a

g

e

Bus 7

Bus 9

Load at Bus 5 is 120+j40

Fig. 4.6 P-V curves with load 120+j40

Figure 4.6 represents the P-V curve without considering the

contingencies. To show the effect of change of load in one area on the

voltage profile these curves are plotted at different loads. In this case the

load is taken as 120+j40 at bus 5 of area one.

138

0 20 40 60 80 100 120 140 160 180 200

0.65

0.7

0.75

0.8

0.85

0.9

0.95

1

1.05

Load in area 2 (MW)

V

o

l

a

t

g

e

Bus 7

Bus 9

Load at Bus 5 is 150+j50

Fig. 4.7 P-V curves with load 150+j50

In figure 4.7 variations in voltage is plotted at a different load

of 150+j50 in area 1. It is observed that at a lesser value of the

convergence is ceased as shown which will also affect the transfer

capability of the system. It is observed that the there is an

interaction between the power transfers and variation of load. It is

also observed that there is a certain change the load margin due to

which the transfer capability also changes.

139

0 20 40 60 80 100 120 140

0.7

0.75

0.8

0.85

0.9

0.95

1

1.05

Load in area 2 (MW)

V

o

l

t

a

g

e

Bus 7

Bus 9

Load at Bus 5 is 90+j30

Fig. 4.8 P-V curves with line outage 4-5

In Figure 4.8 the bus voltages at bus 7 and bus 9 are plotted using

RPF method considering a single contingency with line outage 4-5. The

load in area 1 is maintained at 90+j30. When compared with the Figure

4.7, for the same load in area 1 there is large change in load power

margin and hence the transfer capability also. The graphical

representation of change in ATC is shown in Fig. 4.10 and Fig. 4.11.

140

0 20 40 60 80 100 120 140 160 180

0.7

0.75

0.8

0.85

0.9

0.95

1

1.05

Load in area 2 (MW)

V

o

l

t

a

g

e

Bus 7

Bus 9

Load at Bus 5 is 90+j30

Fig. 4.9 P-V curves with line outage 5-6

The single contingencies considered in Figure 4.9 is outage of

line connecting bus 5 and bus 6 in area 1. It can be observed that the

voltage collapse at bus 7 and bus 9 at a lower value of load compared to

that of contingency free system.

141

80 100 120 140 160 180 200 220

200

225

250

275

300

325

350

375

400

425

450

Load in area 1 (MW)

A

T

C

(

M

W

)

Fig. 4.10 Effect of change in load on ATC without contingency

Figure 4.10 represents the variation of available transfer

capability with respect to the change in load in area 1. The load in area 1

is changed with constant power factor. As the load is increased the ATC

from are 1 to area 2 is decreased due to decrease in load margin. No

contingencies are considered here.

142

80 100 120 140 160 180 200 220

50

55

60

65

70

75

80

85

90

95

100

Load in area 1 (MW)

A

T

C

(

M

W

)

Fig. 4.11 Effect of change in load on ATC with contingency

The variations in ATC with respect to the changes in load of area 1

with contingency i.e. removing line 4-9 are shown in Figure 4.11.

The number of training samples is 40 which are used for training

the network. These training samples are obtained from repeated load flow

method at different load patterns and two single line outages. The

proposed network accepts the diagonal elements of bus admittance

matrix to map the contingencies and load in area 1 as inputs, the

available transfer capability in complex form is the output. This method

is proposed for better prediction how a realistic power system will react

143

over a wide range of operating conditions. The variation of error with

respect to number of iterations is shown in Figure 4.12.

0 1 2 3 4 5 6 7

x 10

4

0

0.05

0.1

0.15

0.2

0.25

0.3

Iterations

E

r

r

o

r

Figure 4.12 Error plot

4.12.2 RESULTS AND DISCUSSIONS

For the purpose of verifying the validity and correctness of the

proposed method a 9 bus system is selected to compute the real and

reactive power transfer from one area to another area. The system

consisting of 9 buses is divided in to two areas. The complex load levels

144

used to create data for training the proposed neural network in area 1

are varied from 100% to 250% of base case values using different line

outages. The available transfer capability (ATC) in MW and the reactive

power transfers in MVAR at different test cases are computed. The

comparison between the proposed CVNN method and the repeated power

flow (RPF) methods are shown in Tables 4.1 to 4.4.

Table 4.1: Power transfer and ATC without contingency

Load in

area 1

RPF CVNN

ATC

(MW)

90+j30 438+j276 441+j282 441

120+j40 416+j245 418+j244 418

150+j50 393+j221 385+j224 385

180+j60 365+j190 359+j180 359

210+j70 340+j175 332+j168 332

Table 4.2: Power transfer and ATC with Line 5-6 outage

Load in

area 1

RPF CVNN

ATC

(MW)

90+j30 357+j214 351+j211 351

120+j40 344+j203 338+j210 338

150+j50 323+j77 318+j95 318

180+j60 287+j130 284+j133 284

210+j70 210+j40 204+j42 204

145

Table 4.3: Power transfer and ATC with Line 6-7 outage

Load in

area 1

RPF CVNN

ATC

(MW)

90+j30 298+j192 290+j186 290

120+j40 293+j185 284+j182 284

150+j50 290+j186 281+j179 281

180+j60 282+j175 276+j172 276

210+j70 244+j260 239+j256 239

Table 4.4: Power transfer and ATC with Line 9-4 outage

Load in

area 1

RPF CVNN

ATC

(MW)

90+30j 90+j34 92+j35 92

120+j40 86+j38 86+j33 86

150+j50 80+j16 81+j20 81

180+j60 78+j20 76+j22 76

210+j70 73+j25 75+j21 75

It is observed that some outages cause a large change in

power transactions as shown in Table 4.3 and Table 4.4. This is due to

the voltage violation limits of the system. The results are compared and it

is found that transfer capability is reduced by 70% in case of

contingency.

146

4.13 COMPUTATION OF ATC FOR A 30 BUS SYSTEM

In this case IEEE 30 bus system Fig. 4.13 [77] (Annexure – 1)

is considered for calculation of Available Transfer Capability. There are

three areas which are interconnected through tie lines as shown in table

4.5 and 4.6.

Table 4.5: Areas

Table 4.6: Tie Lines

The power transferred between one area to other area is the sum of the

powers transferred through the tie lines connecting the two areas. The

available transfer capability between the areas can be found using the

Equation 4.29. Repeated Power flow method discussed in section 4.10.2

is used to obtain the voltage constrained transfer capability of the

system. The load in one area is varied in steps until Jacobian becomes

singular and a point of voltage collapse is reached. In this case Bus 18,

Bus 19 and Bus 20 are assumed to be critical buses according to voltage

profiles and the P-V curves obtained from RPF solution are plotted.

Area Bus Numbers

1 1,2,3,4,5,6,7,8,9,11 & 28

2 12,13,14,15,16,17,18,19,20 & 23

3 10,21,22,24,25,26,27,29 & 30

Areas Tie Lines

Area 1 – Area 2 Bus 4 – Bus 12

Area 1 – Area 3

Bus 6 – Bus 10

Bus 9 – Bus 10

Bus 28 – Bus 27

147

Fig. 4.13 IEEE 30 Bus system

148

75 80 85 90 95 100 105

0.65

0.7

0.75

0.8

0.85

0.9

0.95

1

Load in area 2 (MW) with p.f 0.8

V

o

l

t

a

g

e

Bus 20

Bus 19

Bus 18

Fig. 4.14 P-V curves

4.13.1 RESULTS AND DISCUSSIONS

The voltage constrained transfer capability is calculated using

repeated power flow (RPF). The voltages at bus 18, 19 and 20 are

considered to be critical and the voltage collapse points are shown in

Figure 4.14. The Load in area 3 is maintained at 48.5+j25 where as the

load in area 2 is varied in small steps while the power factor is

maintained constant at 0.8.

149

76 78 80 82 84 86 88 90 92 94

0.7

0.75

0.8

0.85

0.9

0.95

1

Load in Area 2 (MW) with p.f 0.75

V

o

l

t

a

g

e

Bus 20

Bus 19

Bus 18

Fig. 4.15 P-V curves

In the above Figure 4.15 the power factor maintained at 0.75

where as the area 3 load is changed to 210+j157.88. It can be observed

that the voltage constrained total transfer capability is lesser compared

to that with lesser power transaction.

150

75 80 85 90 95 100 105

0.65

0.7

0.75

0.8

0.85

0.9

0.95

1

Load in area 2 (MW) with p.f 0.82

V

o

l

t

a

g

e

Bus 20

Bus 19

Bus 18

Fig. 4.16 P-V curves

The above Figure 4.16 shows the variation in bus voltages with the

change in load of area 2 keeping the Area 3 load as 138.5+j103.8.

In this case the number of input patterns is reduced to two. The

power factor of the load in area 2 and the load in area 3 are considered

as inputs to observe the effect of load power factor in one area and

change in real and reactive powers in another area.

151

The values of ATC estimated by using the load and power factor

reduce the number of input neurons and very much useful for large

systems. No contingencies are considered in this case. In case of

contingencies the method used in the previous section can be used.

Table 4.7: ATC with varying Load in area 3

P.F Load in Area 3 ATC1-2 (RPF) ATC1-2 (CVNN) ATC1-3 (RPF) ATC1-3 (CVNN)

0.8 138.5+j103.8 149 152 250 245

0.8 168.5+j126.3 151 150 260 258

0.8 198.5+j148.8 152 145 265 267

0.8 210.5+j157.8 151 144 268 270

In Table 4.7 The effect of change in load i.e. real and reactive power

in area 3 on the transfer capability is shown. It is observed that the

increase in load has a noticeable effect on the available transfer

capability from area 1 to 2 and also from area 1 to 3. In this case the

power factor is maintained constant while the load is varied in steps.

Table 4.8: ATC with constant Load in area 3 and variable P.F

P.F Load in Area 3 ATC1-2 (RPF) ATC1-2 (CVNN) ATC1-3 (RPF) ATC1-3 (CVNN)

0.78 210.5+j157.8 150 148 267 272

0.8 210.5+j157.8 152 150 271 274

0.82 210.5+j157.8 153 150 274 278

0.86 210.5+j157.8 153 152 277 278

152

Table 4.8 shows the variation of ATC with respect to power factor

and at a constant area load. There is a very small effect on ATC in this

case. The result obtained using the proposed method is compared with

the standard RPF method.

4.14 CONCLUSIONS

This chapter introduces the application of complex valued neural

network for ATC computations with and without contingencies. To

evaluate the performance a numerical example of 9 bus test system is

presented. The voltage limits of the buses and the line losses are well

considered in this method. It is observed that, using this method

available transfer capabilities between system areas can be estimated

accurately with variations in load levels, in the status of lines. The

simulation results show that the proposed method is very effective in

determining the ATC. The suitability of this method is also demonstrated

by taking IEEE 30 bus system where the number of inputs are reduced

to allow the on line computations of ATC for a larger system. In this

problem it is observed that as the number of inputs mapping the

nonlinear system output decreases the relative error increases. In a

deregulated environment, the power transaction between one area to

other area can occur only when there is adequate ATC available for that

interface to ensure system security. This available transfer capability

(ATC) information is to be continuously updated and made available to

153

all participants of the energy market through Open Access Same Time

Information System (OASIS). For such type of n line applications the

proposed method is suitable as it makes use of repeated power flow

method based on Newton-Raphson formulation and the good

generalization capability of the complex valued artificial neural networks.

The main conclusions of this work are:

The proposed CVNN method is effective in calculating the

ATC between different areas subject to system operating

limits;

This method can be adopted for computation of ATC with

constant load power factor at different power factors of the

load; and

The application of proposed method can also be extended to

determine the variations in ATC with respect to reactive

power incorporating FACTS devices.

112 Transfer capability of a power system is defined as the maximum power that can be transferred from one area to another area. 4.2 GENERAL MOTIVATION In open access transmission system, the transmission network owners are required to provide unbundled services to support power transactions and to maintain reliable operation of the networks. In a liberalized electricity markets, to enforce the open access policy North American Energy Reliability Council (NERC) in conjunction with Federal Energy Regulatory Commission (FERC) defined the term available transfer capability (ATC) to be posted in open access same time

information system (OASIS) to inform all the energy market participants of a power system. This information is required to be made available on hourly or daily basis. The two major challenges that make the task of ATC calculation of a nonlinear power system more challenging are computing speed and accuracy due to static and dynamic security constraints. Deregulation of power market has imposed great impact on the utility industry. For smooth transaction of power between the areas or paths new technologies and computation methods are urgently needed. Transfer capability of a power system also indicates how much inter area power transfer can be increased without system security violations. The vital information required for the planning and operation of the large power systems can be obtained from these transfer capability

113 calculations. These details provide system bottlenecks to the planners and the limits of the power transfers to the system operators. The risk of overloads, equipment damage and unexpected blackouts can be reduced by repeated estimation of these transfer capabilities. These calculations also help to determine the quantity of lost generation that can be replaced by potential reserves circumstance. Due to the deregulation of power systems the power transfers are increasing which is necessary for a competitive market for electric power. There is a strong economic incentive to improve the accuracy and effectiveness of the transfer capability computations for the use by the power system operators and the power markets. Transfer capability can be computed using different methods and these computations are evolving. The methods used at present are oversimplified and they do not consider the effects such as and limiting constraints in each

nonlinearities, system policies, interactions between the power transfers and loop flows. Under open access environment of power system actual evaluation of ATC is very much essential and a practical software package for computing ATC will be an important tool for all transmission providers. In recent years a significant progress has been made in developing such a tool, one remaining major challenge is to determine ATC accurately under varying load conditions considering static and dynamic security limits. The main objective of this research is to improve

**114 the accuracy and incorporate the pragmatism in transfer capability computations.
**

4.3 DEFINITIONS

According to the report approved by NERC the definitions of ATC are as follows. Fig. 4.1 [64] represents the various terms. Total Transfer Capability (TTC): It is defined as the quantity of electric power that can be transferred over the interconnected transmission path reliably without violating the predefined set of conditions of the system. Available Transfer Capability (ATC): It is a measure of the transfer capability remaining in the physical transmission network for further commercial activity over and above already committed uses.

Mathematically, ATC is defined as the Total Transfer Capability (TTC) less the Transmission Reliability Margin (TRM), less the sum of existing transmission commitments (which includes retail customer service) and the Capacity Benefit Margin (CBM). Transmission Reliability Margin (TRM): It is defined as that amount of transmission transfer capability necessary to ensure that the

interconnected transmission network is secure under a reasonable range of uncertainties in system conditions. Capacity Benefit Margin (CBM) is defined as that amount of transmission transfer capability reserved by load serving entities to ensure access to generation from interconnected systems to meet generation reliability requirements.

voltage. The various constraints that limit Total Transfer Capability may be the physical and electrical characteristics of the systems including thermal. and stability limits as shown in Figure 4. TTC = Minimum of {Thermal Limit.and post-contingency system conditions. Stability Limit} TTC is an important parameter that indicates how much power transfer can take place without compromising the system security.1 Transfer capabilities and related terms 4.115 Fig.2.4 TOTAL TRANSFER CAPABILITY Total Transfer Capability (TTC) is defined as the amount of electric power that can be transferred over the interconnected transmission network in a reliable manner while meeting all of a specific set of defined pre. Voltage Limit. operators and marketers. The accurate TTC calculation is essential to ensure that power system can . It provides vital information for the planners. 4.

The objective can be defined as the determination of maximum real power transfer between the utilities.116 operate without reliability risks. Specify a transfer which includes source and sink powers. A number of methods exist for calculation of TTC and excessive conservative transfer capability may limit the transfer unnecessarily and also lead to inefficient operation of power system. In other words TTC is the maximum value of power transfer between the paths without any limit violations. with or without contingency. . 4.2 Limits of total transfer capability Transfer capability can be calculated as follows Establish a base case which is assumed to be a secured operating condition so that all line flows and bus voltage magnitudes lie within their operating limits. Fig.

It can be used for comparing the relative merits of planned transmission betterments.117 Establish a solved case by increasing source and sink powers until there is a limit violation. 4.5 METHODS OF TRANSFER CAPABILITY CALCULATION A number of methods have been proposed in the literature. To improve reliability and economic efficiency of the power markets. For providing a quantitative basis for assessing transmission reservations to facilitate energy markets. Various methods of calculating transfer capability are explained below. The transfer margin is the difference between the limiting case transfer and the base case transfer. These methods are classified as i) continuation power flow (CP FLOW) [68] based methods ii) optimal power flow (OPF) based methods and iii) Linear approximation methods.4. 4. .1 PURPOSE OF TRANSFER CAPABILITY COMPUTATIONS The need for transfer capability computations: Estimation of TTC can be used as a rough indicator of relative system indicator.

Continuation methods are also known as curve tracing or path following which are used to trace solution curves for general non-linear algebraic equations with a parametric variation. area real and /or reactive loads and real power generations at generator or P-V buses.1 CONTINUATION POWER FLOW APPROACH Continuation power flow method is a comprehensive tool for tracing the steady state behavior of the power system due to parametric variation [84]. . Linear methods use PTDFs (power flow distribution factors) to express the percentage of power transfer that occurs on a transmission path. This process requires a series of load flow solutions to be solved and tested for limits. The parameters which are varied include bus real and/or reactive loads. Equality constraints obtained from power flow are used in this approach.5.118 Continuation methods in which the transfer capability is computed using a software model called continuation. These methods have four basic elements: Parameterization: This is a mathematical way of identifying each solution for quantifying next solution or previous solution. Predictor: To find an approximate point for the next solution. Tangent or secant method is used for this purpose. 4. Optimal power flow approach is another method to formulate an optimization problem.

Step size control: To adapt the step length for shaping the traced solution curve.1 assuming slow variation of the load parameter (λ) which represents the increment in load demand and power supplied by the system generators.2 and 4.119 Corrector: To correct error in an approximation produced by the predictor before it accumulates.5) QLi ( ) QLio 1 K QLi (4.5 and Equation 4.3) QGi ( ) QLi ( ) ViV j Gij Sin ij Bij Cosij ) ji The increments of the generator active power and demands are given by PGi ( ) PGio 1 K Gi PLi ( ) PLio 1 K PLi (4. Basically it calculates the successive equilibrium points for the Equation 4.6 PLio and QLio represent the base case (λ = 0) active and reactive powers of ith bus.6) In Equation 4.1) Rewriting the load flow equations the real and reactive power can be represented by equation 4.4) (4. . ) (4.3 PGi ( ) PLi ( ) ViV j Gij Cos ij Bij Sinij ) ji (4.2) (4. This method is based on the static model of the power system. 0 g ( y.

66. Successive solution of the above equation for different values of loading parameter λ is obtained by continuation power flow through a predictor and corrector [65. g ( y.3 . 68] as shown in Figure 4.d 0 (4.7) In above equation 4.7.5 is the base case active power supplied by the generator of ith bus. Above equations can be written in a compact form as in equation 4.7 d represents a vector indicating the direction of the active power increment supplied by the generators and reactive power increment consumed by the loads.120 PGio in equation 4. ) G( y) . KPLi and KQLi are coefficients defining the load power factors of the ith bus. KGi is a coefficient defining the generator participation factor in the ith bus for certain loading level λ.

To determine the total transfer capability the objective is to .5.121 Fig. is the maximum power flow on the specified transmission route. In the mean time. inequality constraints from basic operation and equipment limits to more detailed approximation of transient stability security requirements. 4. The objective function. obviously. TTC calculation by OPF approach has been proposed since 1999 [73][74][75]. with equity constraints of power flow.2 OPTIMAL POWER FLOW APPROACH The application of Optimal Power Flow (OPF) in power system congestion management has been studied by some researchers [69][70][71][72]. The basic concept of OPF approach is formulating the TTC calculation as an optimization problem.3 Predictor and corrector method of CP FLOW 4.

Maximize Pi Pkj ji (4. The advantage of this approach is its simple implementation and the ease to take security constraints into consideration.14) Qi ViV j Yij Sin ( ij i j ) 0 ji Pg min Pg Pg max Qg min Qg Qg max S ij S ij max Vi min Vi Vi max 4.5.13) (4. . and repeatedly solves the power flow equations each time increasing the power transfer by a small increment until an operation limit is reached [76]. In this dissertation.10) (4.8) Subjected to Pi ViV j Yij Cos( ij i j ) 0 ji (4. Total transfer capability problem formulation can be explained as follows.11) (4.12) (4. this method is adopted to solve TTC problem.122 maximize the power transfer between the two areas subjected to the conditions that there is no voltage or thermal or stability limit violations.9) (4.3 REPEATED POWER FLOW APPROACH Repeated power flow approach starts from a base case.

another load flow is solved and the security constraints tested.6 ALGORITHM FOR REPEATED POWER FLOW METHOD In this research work. and then increasing the transfer. Repeat the procedure until minimum step size reached 4. This method involves the solution of a base case.7 AVAILABLE TRANSFER CAPABILITY [ATC] Available transfer capability computations are essential for successful implementation of electric power deregulation where power producers and customers share a common transmission network for wheeling power from the point of generation to the point of consumption. After each increase. Establish and solve for a base case ii. Decrease step size if transfer is unsuccessful vi. Flow chart for the algorithm is given in Annexure IV. Increase step size if transfer is successful v. which is the initial system conditions. The available transfer capability indicates the amount which inter-area bulk power transfers can be increased without compromising system security. Solve for the transfer case iv. Select a transfer case iii. The value used for available transfer capability affects both .123 4. The computational procedure of this approach is as follows: i. it is proposed to utilize the repeated power flow (RPF) method [67] for the calculation of transfer capabilities due to the ease of implementation.

In a deregulated open access market it can be used as tool for transmission pricing. . there is increasing motivation for defensible calculations of available transfer capability.1 IMPORTANCE OF ATC ASSESSMENT In this present open access or deregulated environment all the participants (producers and buyers of electrical energy) desire to produce or consume large amounts of energy and may force the transmission system to operate beyond one or more transfer limits. 4. It ensures the secure operation of the system.124 system security and the profits made in bulk power transactions.7. Moreover. Therefore accurate determination of available transfer capability is essential to ensure the system security and reliability while serving a wide range of bilateral and multilateral power transactions. market participants can have conflicting interests in a higher or a lower available transfer capability. On the basis of ATC computation firm and non-firm reservation of transmission services can be made. Thus under deregulation. This kind of operation leads to congestion of the system. The following reasons show the need of ATC computations It gives the amount of maximum additional power transfer between the specified interfaces.

such as the case of very long time frame analysis.9 COMPUTATION OF ATC USING LINEAR METHODS In this section linear methods are explained briefly for determination of available transfer capability followed by the merits and demerits of these approaches. . 4.125 The limits or binding constraints for ATC can be used in power system planning and network expansion. A DC model may be preferable to an AC model particularly in circumstances where the extra data for an AC model is unavailable or very uncertain. The following linear methods are used for calculation of ATC.8 METHODS OF CALCULATING ATC As explained above the following methods are used for determination of ATC 1. Network sensitivity method Full AC power flow method is the most accurate method for computation of ATC but its complexity can blot out relationships. Load flow / continuous power flow (CPF) / Repeated Power Flow (RPF) methods. 4. Optimization based methods 3. 2. Transfer capabilities can be estimated with simple power system models such as the DC load flow approximation.

n (4. Pn x 1 .15 explain DC power flow method Pij 1 i j xij (4.9. Pi Pij j j 1 i j xij (4. The following Equation 4.17) Where the elements of susceptance matrix Bx are functions of line reactance xij .1 DC POWER FLOW METHOD This model assumes that the voltage magnitudes are constant and only the angles of the complex bus voltage vary.16) This can be expressed in a matrix form by Equation 4.15) Where xij θi θj line inductive reactance phase angle at bus i phase angle at bus j The total power flowing in to bus i is the algebraic sum of generation and demand at the bus called bus power injection given by Equation 4.126 4.17 P 1 . It is also assumed that the transmission line has no resistance and therefore no losses. In addition to the speed of computation this method has other useful properties like linearity and super position.16.

79]. n 1 P 1 . Power transfer distribution factor (PTDF) is defined as the coefficient of the linear relationship between the amount of a transaction and the flow on a line. the reactance matrix can be obtained by inversion.mn represents the fraction of a transaction from m zone to zone n that flows over a transmission line connecting zone i to zone j. transaction amount to another change i. The phase angles are obtained as a function of bus injections as shown in Equation 4.2 POWER TRANSFER DISTRIBUTION FACTORS According to power flow point of view power injected in to the system at a point by generator is extracted by a load at another point which is known as transaction. . Transaction can be found from the linear property of DC load flow model using sensitivity factor PTDFs [78.18) Line flows Pij are obtained from Equation 4.127 Phase angle of one of the buses is set to zero and eliminating the row and column from Bx matrix. PTDFij.e.e.18 1 . 4. As it relates the amount of one change i.9. PTDF is the fraction of amount of a transaction from one zone to another over a specified transmission line. the lone flow. Pn 1 (4.16 to compare with the line MW limits.

mn Where xij X im X jm X in X jn xij (4.20) 4. New PijNew PTDFij .9.19) transmission line reactance connecting zone i and zone j Xim entry in the ith row and mth column of the bus reactance matrix X The change in line flow associated with a new transaction is then given by Equation 4.rs. . This redistribution is measured by LODF and the fraction of power flowing on the line from zone r to zone s before it is removed which now flows over a line from zone i to zone j is given by LODFij. When a line outage occurs in a system the power flowing on that line is redistributed on to the remaining lines in the system. The speed of computation of this method can be improved by using another sensitivity called line outage distribution factor (LODF).128 The Power Transfer Distribution Factor (PTDF) is given by Equation 4.20.3 LINE OUTAGE DISTRIBUTION FACTORS (LODF) The simple but most inaccurate method used to calculate the effect of line outage is DC power flow contingency.19 PTDFij .mn Pmn (4.

xrs .rs Prs LODFij .rs LODFij .10.129 The change in power flow is given by Pij .1 DC APPROXIMATION The DC approximation is preferred for these reasons: Fast computation . . MW limits are considered. N ij .no iteration.10 MAJOR ADVANTAGES & DISADVANTAGES OF THE EARLIER METHODS: The advantages and disadvantages of various methods of transfer capability computation are as follows: 4. Network topology handled with simple and linear methods.xij N rs . Good approximation over large range of Minimum data is required. But DC approximation is poor for these reasons: It cannot identify voltage limits conditions. Thermal limits.22) X ir X is X jr X js N rs .rs Where xij Xir X Nij number of circuits connecting zone i and zone j line reactance connecting zone i and j (4.xrs X rr X ss 2 X rs entry in the ith row and rth column of the bus reactance matrix 4.21) (4.

130 It is not accurate when VAR flow and when voltage deviations are considerable.3 OPTIMAL POWER FLOW METHOD Optimal Power Flow (OPF) and Security Constrained Optimal Power Flow (SCOPF) are widely used to determine ATC in power corridors of the system. however its applications are extended to .10. This approach has a major disadvantage that they do not consider the nonlinear effects of voltage and reactive power.10. Over use of linear superposition increases errors 4. 4.2 SENSITIVITY FACTORS One of the earliest solutions proposed for Available Transfer capability (ATC) is sensitivity analysis. These sensitivity factors are based on linear incremental power flow. 4.10. Moreover the methods based on Power Transfer Distribution Factors (PTDFs) and Line Outage Distribution Factors (LODFs) can be used to estimate the ATC of the cases nearer to the base case from which these factors are derived. However these optimization methods are suitable in case of open access system where there is a possibility of power transactions occurring from any point to any point.4 CONTINUATION POWER FLOW As discussed earlier this method is initially used for finding maximum loadability point. which are very simple to define and calculate.

Luo et. which is the true . Therefore. The computational time increases due to its complexity. In this work Quick prop algorithm is used to train the neural network. al. this method is unable to track down the bus-to-bus transactions. The goal of the methods described here is to improve the accuracy and realism of transfer capability computations. operating policies and most importantly voltage collapse blackouts. and in the status of lines. The transfer capability was estimated accurately of between system areas with variations in load levels. The computations presently being implemented are usually oversimplified and in many cases do not take sufficient account of effects such as interactions between power transfers.131 determine TTC and ATC. when contingencies are included. The artificial neural network approach reported in the methods proposed require a large input vector for bilateral transaction. [77] uses a Multi Layered Feed Forward Neural Networks which is capable of reflecting variations in load levels and in the status of generation and transmission lines.11 ATC COMPUTATION USING COMPLEX VALUED NEURAL NETWORKS The practical computations of transfer capability are evolving. so it has oversimplified the determination of ATC by limiting it to a special case of power transfer to a single area from all of the remaining areas. The transfer capability estimation method proposed by X. in the status of generation. 4. loop flows. non-linearities.

4.4. This method involves the solution of a base case. b) c) The load and generation patterns vary very slowly. 4.4 a Complex Valued Neural Network approach is implemented for ATC computation. After each increase.12 COMPUTATION OF ATC FOR A 9-BUS SYSTEM For computing Available Transfer Capability using the proposed approach a 9 bus system [120] is considered (Annexure – 1). It has 3 generator buses and the number of transmission lines is 9 as shown in Figure 4. another load flow is solved and the security constraints tested.11. and TTC calculation is in the steady state analysis domain.132 spirit of deregulation and nonlinearities of the systems are also not considered. a) The base case power flow of the system is feasible and corresponds to a stable operating point.1 ASSUMPTIONS From the formulations discussed in section 3. The following assumptions are made while calculating the ATC. which is the initial system conditions. and then increasing the transfer. Voltage constraint is taken .

. Area 1 comprises of bus 1. bus 5 and bus 6 where as area 2 comprises of the remaining buses as shown in Fig. The system divided in to two areas. 4. These lines transmit the power from one geographical area to other area.4 Nine bus system.bus 4. 4. The tie lines between the two areas are line 4-9 and line 6-7.4 [84].bus 3.133 Fig.

Identify the line outages having significant influence on the above said branch power flows. Repeated power flow method is utilized to generate training patterns using Newton Raphson load flow method. The proposed new methodology is based on the full Ac load flow and strong generalization capability of complex neural network offers a great potential for real time evaluation of ATC incorporating load variation. In this method the available transfer capability (ATC) from one bus/area/zone to another bus/area/zone can be found by varying the amount of transaction until one or more line flows in the transmission system considered or a bus voltage reaching the limiting value. For calculating ATC using Repeated Power Flow (RPF) method the following choices are made: Establish a secure.1 LEARNING METHODOLOGY Learning methodology uses complex back propagation algorithm explained in Chapter 3.12. Repeated power flow (RPF) method is proposed for calculating ATC due to the ease of implementation. Identify the branch flows influencing the ATC of selected branch appreciably. effect of reactive power loss of the system. Specify a transfer including source and sink assumptions.134 4. . solved base case.

max (4.23) (4.25) (4. The loads at bus number 7 and 9 are increased simultaneously and the transfers from area 1 to area 2 are obtained. And also the number of iterations required is independent of the system size.28) .min Vi Vi .27) (4.24) (4. The flow chart for the above algorithm is shown in Annexure –IV. The available transfer capability is given by Equation 4.23 ATC = TTC – base case transfer Satisfying the following system operating conditions Pi ViV j Yij Cos( ij j i ) 0 jN (4. The transfer margin is the difference between the transfer at the base case and the limiting case.26) Qi ViV j Yij Sin ( ij j i ) 0 jN Pg min Pg Pg max Qg min Qg Qg max Vi .135 Generate numerous training data sets involving above said power flows and line outages. line joining buses 4 and 9. The calculation of ATC is done by using the Newton Raphson load flow solution to compute the power flow of each transfer case. The total transfer capability is the sum of transfers through the interconnecting lines i. buses 6 and 7.e. This method is less prone to divergence with ill-conditioned problems.

5 shows the P-V curve obtained from repeated power flow. The load in area 1 that is at bus 5 is maintained as 90+j30 where as the load in area 2 that is at bus 7 and bus 9 is varied slowly.5 P-V curves with load 90+j30 The above Figure 4. From the results obtained from the repeated power flow it is observed that the voltage at these two buses reached the nose point compared to other buses. It is assumed that bus 7 and bus 9 are critical buses compared to other buses and the voltage collapse points of these buses are shown with respect to the total load in area 2.8 0.75 0.95 Bus 7 0.9 Voltage 0.05 1 0. .136 1.7 Bus 9 Load at Bus 5 is 90+j30 0. 4.65 0 50 100 150 Load in area 2 (MW) 200 250 Fig.85 0.

85 0.6 represents the P-V curve without considering the contingencies.65 0 20 40 60 80 100 120 140 Load in area 2 (MW) 160 180 200 Fig. .6 P-V curves with load 120+j40 Figure 4.95 0.9 Voltage Bus 9 0.75 0. In this case the load is taken as 120+j40 at bus 5 of area one.05 1 Bus 7 0. To show the effect of change of load in one area on the voltage profile these curves are plotted at different loads.8 0. 4.7 Load at Bus 5 is 120+j40 0.137 1.

9 Volatge Bus 9 0.7 Load at Bus 5 is 150+j50 0.95 0. 4.85 0.138 1. It is also observed that there is a certain change the load margin due to which the transfer capability also changes. It is observed that the there is an interaction between the power transfers and variation of load.7 P-V curves with load 150+j50 In figure 4. It is observed that at a lesser value of the convergence is ceased as shown which will also affect the transfer capability of the system.7 variations in voltage is plotted at a different load of 150+j50 in area 1.8 0.05 1 Bus 7 0.65 0 20 40 60 80 100 120 Load in area 2 (MW) 140 160 180 200 Fig. .75 0.

11.10 and Fig. The graphical representation of change in ATC is shown in Fig.85 0. When compared with the Figure 4.8 P-V curves with line outage 4-5 In Figure 4.05 1 Bus 7 0. 4.8 the bus voltages at bus 7 and bus 9 are plotted using RPF method considering a single contingency with line outage 4-5.8 0.139 1.7. . The load in area 1 is maintained at 90+j30.95 Voltage 0. 4.7 0 20 40 60 80 Load in area 2 (MW) 100 120 140 Fig. 4. for the same load in area 1 there is large change in load power margin and hence the transfer capability also.9 Bus 9 0.75 Load at Bus 5 is 90+j30 0.

. It can be observed that the voltage collapse at bus 7 and bus 9 at a lower value of load compared to that of contingency free system.7 0 20 40 60 80 100 120 Load in area 2 (MW) 140 160 180 Fig.8 0.05 1 Bus 7 0.75 Load at Bus 5 is 90+j30 0.85 0.140 1. 4.9 Bus 9 0.9 is outage of line connecting bus 5 and bus 6 in area 1.9 P-V curves with line outage 5-6 The single contingencies considered in Figure 4.95 Voltage 0.

As the load is increased the ATC from are 1 to area 2 is decreased due to decrease in load margin.141 450 425 400 375 ATC ( MW) 350 325 300 275 250 225 200 80 100 120 140 160 Load in area 1 (MW) 180 200 220 Fig.10 Effect of change in load on ATC without contingency Figure 4.10 represents the variation of available transfer capability with respect to the change in load in area 1. . 4. The load in area 1 is changed with constant power factor. No contingencies are considered here.

This method is proposed for better prediction how a realistic power system will react .e. removing line 4-9 are shown in Figure 4. The number of training samples is 40 which are used for training the network.142 100 95 90 85 ATC (MW) 80 75 70 65 60 55 50 80 100 120 140 160 180 200 220 Load in area 1 (MW) Fig. These training samples are obtained from repeated load flow method at different load patterns and two single line outages. 4. the available transfer capability in complex form is the output.11 Effect of change in load on ATC with contingency The variations in ATC with respect to the changes in load of area 1 with contingency i. The proposed network accepts the diagonal elements of bus admittance matrix to map the contingencies and load in area 1 as inputs.11.

05 0 0 1 2 3 4 5 6 7 x 10 4 Iterations Figure 4. The variation of error with respect to number of iterations is shown in Figure 4.2 RESULTS AND DISCUSSIONS For the purpose of verifying the validity and correctness of the proposed method a 9 bus system is selected to compute the real and reactive power transfer from one area to another area.143 over a wide range of operating conditions. The complex load levels .12.2 Error 0.3 0.25 0. The system consisting of 9 buses is divided in to two areas.1 0.15 0. 0.12 Error plot 4.12.

Table 4.2: Power transfer and ATC with Line 5-6 outage Load in area 1 90+j30 120+j40 150+j50 180+j60 210+j70 RPF 357+j214 344+j203 323+j77 287+j130 210+j40 CVNN 351+j211 338+j210 318+j95 284+j133 204+j42 ATC (MW) 351 338 318 284 204 .4. The available transfer capability (ATC) in MW and the reactive power transfers in MVAR at different test cases are computed.1: Power transfer and ATC without contingency Load in area 1 90+j30 120+j40 150+j50 180+j60 210+j70 RPF 438+j276 416+j245 393+j221 365+j190 340+j175 CVNN 441+j282 418+j244 385+j224 359+j180 332+j168 ATC (MW) 441 418 385 359 332 Table 4. The comparison between the proposed CVNN method and the repeated power flow (RPF) methods are shown in Tables 4.144 used to create data for training the proposed neural network in area 1 are varied from 100% to 250% of base case values using different line outages.1 to 4.

.3: Load in area 1 90+j30 120+j40 150+j50 180+j60 210+j70 Power transfer and ATC with Line 6-7 outage RPF 298+j192 293+j185 290+j186 282+j175 244+j260 CVNN 290+j186 284+j182 281+j179 276+j172 239+j256 ATC (MW) 290 284 281 276 239 Table 4. This is due to the voltage violation limits of the system.145 Table 4.3 and Table 4. The results are compared and it is found that transfer capability is reduced by 70% in case of contingency.4.4: Power transfer and ATC with Line 9-4 outage Load in area 1 90+30j 120+j40 150+j50 180+j60 210+j70 RPF 90+j34 86+j38 80+j16 78+j20 73+j25 CVNN 92+j35 86+j33 81+j20 76+j22 75+j21 ATC (MW) 92 86 81 76 75 It is observed that some outages cause a large change in power transactions as shown in Table 4.

3.20 & 23 10.2 is used to obtain the voltage constrained transfer capability of the system.146 4.11 & 28 12.16.5 and 4. In this case Bus 18.24.21.22.5.6.6.13 COMPUTATION OF ATC FOR A 30 BUS SYSTEM In this case IEEE 30 bus system Fig.7. Table 4.13. Bus 19 and Bus 20 are assumed to be critical buses according to voltage profiles and the P-V curves obtained from RPF solution are plotted.8. . 4.14.4.26.5: Areas Area 1 2 3 Bus Numbers 1. The load in one area is varied in steps until Jacobian becomes singular and a point of voltage collapse is reached.19.6: Tie Lines Areas Area 1 – Area 2 Area 1 – Area 3 Tie Lines Bus 4 – Bus 12 Bus 6 – Bus 10 Bus 9 – Bus 10 Bus 28 – Bus 27 The power transferred between one area to other area is the sum of the powers transferred through the tie lines connecting the two areas.29.25. There are three areas which are interconnected through tie lines as shown in table 4.29 & 30 Table 4.10.9.2.15.17.13 [77] (Annexure – 1) is considered for calculation of Available Transfer Capability.18. The available transfer capability between the areas can be found using the Equation 4. Repeated Power flow method discussed in section 4.27.

4.147 Fig.13 IEEE 30 Bus system .

95 0.1 RESULTS AND DISCUSSIONS The voltage constrained transfer capability is calculated using repeated power flow (RPF).8 100 105 Fig.85 Voltage Bus 20 Bus 19 Bus 18 0.14.f 0.65 75 80 85 90 95 Load in area 2 (MW) with p.13. The Load in area 3 is maintained at 48. 4. 19 and 20 are considered to be critical and the voltage collapse points are shown in Figure 4.9 0.14 P-V curves 4. The voltages at bus 18.7 0.5+j25 where as the load in area 2 is varied in small steps while the power factor is maintained constant at 0.148 1 0.8 0.75 0. .8.

149 1 Bus 20 Bus 19 Bus 18 0.75 0.88.f 0.85 0.7 76 78 80 82 84 86 88 90 Load in Area 2 (MW) with p.9 Voltage 0.15 P-V curves 92 94 In the above Figure 4. .8 0.75 where as the area 3 load is changed to 210+j157. 4. It can be observed that the voltage constrained total transfer capability is lesser compared to that with lesser power transaction.75 Fig.95 0.15 the power factor maintained at 0.

The power factor of the load in area 2 and the load in area 3 are considered as inputs to observe the effect of load power factor in one area and change in real and reactive powers in another area. 4.65 75 80 85 90 95 Load in area 2 (MW) with p. In this case the number of input patterns is reduced to two.f 0.5+j103. .82 Fig.8 0.7 0.75 0.150 1 Bus 20 Bus 19 Bus 18 0.85 0.95 0.16 P-V curves 100 105 The above Figure 4.16 shows the variation in bus voltages with the change in load of area 2 keeping the Area 3 load as 138.9 Voltage 0.8.

5+j157.5+j148. Table 4.8 0.8 168.7: ATC with varying Load in area 3 P. real and reactive power in area 3 on the transfer capability is shown.78 0.8 210.86 Load in Area 3 ATC1-2 (RPF) ATC1-2 (CVNN) ATC1-3 (RPF) ATC1-3 (CVNN) 210.151 The values of ATC estimated by using the load and power factor reduce the number of input neurons and very much useful for large systems.5+j103.5+j157.8 210.8 210.F P.F 0.5+j157. Table 4.8 149 151 152 151 152 150 145 144 250 260 265 268 245 258 267 270 In Table 4.5+j126. It is observed that the increase in load has a noticeable effect on the available transfer capability from area 1 to 2 and also from area 1 to 3.e.3 198.8 0.82 0.8 Load in Area 3 ATC1-2 (RPF) ATC1-2 (CVNN) ATC1-3 (RPF) ATC1-3 (CVNN) 138.F 0.8: ATC with constant Load in area 3 and variable P.8 0.8 150 152 153 153 148 150 150 152 267 271 274 277 272 274 278 278 . In this case the power factor is maintained constant while the load is varied in steps.5+j157. In case of contingencies the method used in the previous section can be used.7 The effect of change in load i.5+j157.8 0. No contingencies are considered in this case.8 210.

in the status of lines. the power transaction between one area to other area can occur only when there is adequate ATC available for that interface to ensure system security. To evaluate the performance a numerical example of 9 bus test system is presented.8 shows the variation of ATC with respect to power factor and at a constant area load.152 Table 4. In a deregulated environment. The result obtained using the proposed method is compared with the standard RPF method. The simulation results show that the proposed method is very effective in determining the ATC. This available transfer capability (ATC) information is to be continuously updated and made available to . The suitability of this method is also demonstrated by taking IEEE 30 bus system where the number of inputs are reduced to allow the on line computations of ATC for a larger system. using this method available transfer capabilities between system areas can be estimated accurately with variations in load levels. In this problem it is observed that as the number of inputs mapping the nonlinear system output decreases the relative error increases. There is a very small effect on ATC in this case. 4. The voltage limits of the buses and the line losses are well considered in this method. It is observed that.14 CONCLUSIONS This chapter introduces the application of complex valued neural network for ATC computations with and without contingencies.

The main conclusions of this work are: The proposed CVNN method is effective in calculating the ATC between different areas subject to system operating limits. This method can be adopted for computation of ATC with constant load power factor at different power factors of the load. .153 all participants of the energy market through Open Access Same Time Information System (OASIS). For such type of n line applications the proposed method is suitable as it makes use of repeated power flow method based on Newton-Raphson formulation and the good generalization capability of the complex valued artificial neural networks. and The application of proposed method can also be extended to determine the variations in ATC with respect to reactive power incorporating FACTS devices.

- 00158835
- 10.1.1.70
- Ieee 14 Bus System Data
- 5 Vol 6 No 1
- IEEE Reliability Test System
- Deem-2011 Proceedings
- PSA12 Rep Complete HS12
- Power System Optimization
- power simulation1
- Powerflow Newton-raphson Example
- Contingency Thesis
- 06119996_002
- ECE4762011_Lect14
- Load Flow Analysis and GA Method
- CD 4201522527
- Optimal Power Flow Ieee
- date-57f1abf58dc212.04275283.pdf
- Induction furnace, Harmonics, THD, Simulation.
- DVR INtro
- h 044023944
- 10 Chapter 5
- IOSRJEN(www.iosrjen.org)IOSR Journal of Engineering
- newton,statcom1
- Trans Pricing NPTI 090811
- Power Flow Studies-chapter 5aaa
- Composite Plate Stiffness Multicriteria Optimization Using Lamination Parameters 2015 Composite Structures
- 10_chapter 5.pdf
- Power flow solutions
- SSModel_PowerFlow_PV_IEEE.pdf
- Loss Sharing Insts

- Comparative study of Conventional and Traditional Controller for Two Equal Area interconnected power plant for Automatic Generation Control
- UT Dallas Syllabus for opre7330.001.07f taught by Milind Dawande (milind)
- Emerging Trends of Ant Colony Optimization in Project Scheduling
- UT Dallas Syllabus for eco5302.501.07f taught by Chetan Dave (cxd053000)
- Surveying, Planning and Scheduling For A Hill Road Work at Kalrayan Hills by Adopting ISS Method
- tmp5056.tmp
- Weight Optimization of Composite Beams by Finite Element Method
- Comparative Analysis of Optimization Algorithms Based on Hybrid Soft Computing Algorithm
- The Optimizing Multiple Travelling Salesman Problem Using Genetic Algorithm
- tmpBFF4
- Analysis of CAS Using Improved PSO with Fuzzy Based System
- tmpA166.tmp
- UT Dallas Syllabus for opre7313.001.11s taught by Milind Dawande (milind)
- Determining the shortest path for Travelling Salesman Problem using Nearest Neighbor Algorithm
- Optimization of Water Distribution Network for Dharampeth Area
- A Review on optimization of Dry Electro Discharge Machining Process Parameters
- A Review on Various Approach for Process Parameter Optimization of Burnishing Process and TAGUCHI Approach for Optimization
- Planning of Supply when lead time uncertain
- Mathematics-III
- UT Dallas Syllabus for math6321.521 06u taught by Janos Turi (turi)
- tmp5DD5.tmp
- A Review on Image Segmentation using Clustering and Swarm Optimization Techniques
- UT Dallas Syllabus for opre7320.001.09s taught by Suresh Sethi (sethi)
- Analysis and Modeling of Single Point Cutting Tool with help of ANSYS for Optimization of Vibration Parameter
- Multi-objective Optimization for Ultrasonic Drilling of GFRP Using Grey Relational Analysis
- Recent Advancements in Scheduling Algorithms in Cloud Computing Environment
- Quantitative Management for Analysis
- Optimization of Shell and Tube Heat Exchanger
- ON Fuzzy Linear Programming Technique Application
- tmp82D3.tmp

- Improvement of Probabilistic small signal stability of power system with large scale wind integration
- Tao_HONG_DR
- Transformation of Random Variables
- 1-s2.0-S030626191200400X-main.pdf
- Structural Reliability under Incomplete Probability Information
- Impact of Increased Penetration of DFIG based wind turbine generators on transient and small signal stability of power systems
- 07848082
- Writing a Paper Guidelines
- li2013.pdf
- Feasibility study of segmenting large power system interconnections with ac link using energy storage technology
- Firefly Algorithm
- Impact of Battery Energy Storage System on Dynamic Properties of Isolated Power Systems
- 1308.3898.pdf
- Impact of wind generators considering cor-relation
- Probabilistic Analysis of Small-Signal Stability of Large-Scale Power Systems as Affected by Penetration of Wind Generation
- Power system assignment
- Power system practical assignment
- Pokhara feeder single line diagram
- IITB PhD application form
- Tower Design
- New Text Document
- Rural electrification
- 4th Year
- A detailed description on active filter
- ACSR conductor data
- Single Phase Z Source PWM AC AC Converters 11H
- Nepali Patro - Nepali Calendar - 2072
- Introduction to Electrical Engineering
- New Text Document
- Deregulated power system

Sign up to vote on this title

UsefulNot usefulRead Free for 30 Days

Cancel anytime.

Close Dialog## Are you sure?

This action might not be possible to undo. Are you sure you want to continue?

Close Dialog## This title now requires a credit

Use one of your book credits to continue reading from where you left off, or restart the preview.

Loading