World Academy of Science, Engineering and Technology

International Journal of Energy and Power Engineering

Vol:12, No:9, 2018

Comparison of Newton Raphson and Gauss Seidel

Methods for Power Flow Analysis
H. Abaali, T. Talbi, R. Skouri

 in a power system, powers are known rather than currents. So,

Abstract—This paper presents a comparative study of the Gauss the steady-state performances of an interconnected power
Seidel and Newton-Raphson polar coordinates methods for power system can be modeled from a polynomial system equation in
flow analysis. The effectiveness of these methods are evaluated and several variables called power flow equations become
tested through a different IEEE bus test system on the basis of
nonlinear. To approximate the solutions of these nonliner
number of iteration, computational time, tolerance value and
convergence. equations, the numerical methods should be used [2]. The
algorithms applied to solve the power flow equations can be
Keywords—Convergence time, Gauss-Seidel Method, Newton- classified into two broad categories: coordinate category
Raphson Method, number of iteration, power flow analysis. methods such as Gauss Seidel [1] and gradient category
methods such as Newton-Raphson [4]-[6]... The power flow
results give the bus voltage magnitude and phase angles and
Open Science Index, Energy and Power Engineering Vol:12, No:9, 2018 waset.org/Publication/10009525

NOMENCLATURE
Var volt-ampere reactive hence the power flow through the transmission lines, line
Y  Y  G  jB complex nodal admittance matrix losses and injection at all the buses.
Z nodal impedance matrix The power system is assumed to be operating under
J Jacobian matrix balanced condition and can be represented by a single-phase
G conductance network. The power system network contains hundreds of
B susceptance buses and branches with impedances specified in per-unit on a
N-bus number of buses in system common MVA base.
s slack bus index In this work, we present a comparative study of these two
NPV number of generator buses
NPQ number of load buses categories iterative solution techniques for power flow
analysis [7]. The performances of these methods are assessed
Vi  Vi  i complex nodal voltage at ith bus
from four aspects: Number of iterations, computational time,
Ii complex nodal injected current at ith bus
tolerance value and convergence. The two methods are tested
Pi active power at ith bus
Qi reactive power at ith bus on five IEEE standard 3-Bus, 5-Bus, 14-Bus, 30-Bus and 57-
Bus test systems. The simulation is carried out using
S i  Pi  jQi complex nodal injected power at ith bus
MATLAB version 7.8.0.347 (R2008a) in Intel® Atom ™
ΔPi = Pci -Pi active power mismatch
CPU N450 @1,66GHz.
ΔQi = Qci -Qi reactive power mismatch
This paper is organized as follows: After the introduction,
I. INTRODUCTION the modelling and power flow problem formulation of an
electric power system was briefly discussed and deals with the
T HE power flow studies, commonly referred to as load
flow, are the backbone of power system analysis and
design. They are necessary for planning future expansion,
steady-state analysis of interconnected power system during
normal operation in Section II. Section III describes the load
flow solution using Gauss Seidel and Newton Raphson polar
control, economic scheduling, and management of power coordinates methods. Section IV introduces the simulation
systems as well as in determining the best operation of results and discussion. Finally Section V presents the
existing systems. In addition, power flow analysis is required conclusion.
for many other analyses such as transient stability and
contingency studies. II. POWER FLOW PROBLEM FORMULATION
The network equations can be formulated systematically in
a variety of forms. However, the node-voltage method is A. Bus Classification
commonly used for power system analysis. The network In a power flow study, a bus is defined as the vertical line at
equations which are in the nodal admittance form result in which the several components such as generators, loads and
complex linear simultaneous algebraic equations in terms of transformers are connected. Each bus is associated with four
node currents [1]. When node currents are specified, the set of variables: Magnitude of voltage, phase angle of voltage, active
linear equations can be solved for the node voltages. However, and reactive power. Two of the four variables are specified
and the other two variables are unknown. The two unknown
variables are determined through the solution of the nonlinear
H. Abaali is with the Moulay Ismaïl University, Faculty of Sciences and
techniques, BP 509 Boutalamine, Errachidia, Morocco (phone : +212 5 35 57
power flow equations. The buses are classified into three
44 97; fax: +212 5 35 57 47 14; e-mail: [email protected]). categories:

International Scholarly and Scientific Research & Innovation 12(9) 2018 627 ISNI:0000000091950263
 World Academy of Science, Engineering and Technology
International Journal of Energy and Power Engineering
Vol:12, No:9, 2018

- Load (PQ bus): No generator is connected to the bus. The minimum number of simultaneous equations in terms of an
load drawn by these buses are defined by real power -PLi equal number of state variables (equal number of equations
and reactive power QLi in which the negative sign and unknowns). The solution of these equations provides the
accommodates for the power flowing out of the bus. The system state.
objective of the load flow is to find the bus voltage The power flow model consists of a set of nonlinear
magnitude Vi and its angle δi. algebraic equations. These equations formulate active and
- Generator bus: The input power PGi and the bus voltage Vi reactive line flows based on bus voltage magnitudes and phase
are kept constant. We have to find the unknown angle δi angles. In most power system studies, transmission lines are
of the bus voltage and the generated reactive power QGi represented by a π-equivalent circuit [8].
through the load flow solution. A systematic way of writing the power flow equations for
- Slack (swing) bus: Usually this generator bus is numbered any bus is given by (9), followed by the selection process. The
1 for the load flow studies. This bus sets the angular power flow equations can be developed with reference to Fig.
reference for all the other buses. Since it is the angle 1 illustrating a general circuit between any two buses i and j
difference between two voltage sources that dictates the buses.
real and reactive power flow between them, the particular In general, one or more circuits may be connected to a bus.
angle of the slack bus is not important. However, it sets In addition, an admittance yii may be also connected to a bus (a
the reference against which angles of all the other bus capacitor, a reactor, etc.). It is assumed that electric current Igi
voltages are measured. For this reason the angle of this is injected to bus i from the generators connected to this bus.
bus is usually chosen as δ1=0. Furthermore, it is assumed Also, electric current Idi is absorbed from the electric load
Open Science Index, Energy and Power Engineering Vol:12, No:9, 2018 waset.org/Publication/10009525

that the magnitude of the voltage V1 of this bus is known. connected to this bus. One or both of these currents may be
The bus classification is summarised in Table I. absent from a bus. The voltage of bus i is assumed to be Vi and
the voltage of bus j is assumed to be Vj.
B. Network Models
Power network can be operating under balanced or TABLE I
BUS CLASSIFICATION
unbalanced conditions. The normal procedure for a load flow
problem formulation is to assume balanced equations at the Bus Type Bus code Known variables Unknown variables
buses of the system and to use a single-phase representation Slack (Swing) bus 1 V∟ P and Q
Generator (PV) bus 2 P and V Q and 
equivalent to the positive sequence network and selecting a
Load (PQ) bus 3 P and Q V∟
proper subset of these equations which will provide the


Igi
bus i Iij yij bus j
Idi

Vi yii yjj Vj

Fig. 1 Symbolic representation of a line model of an electric power system

C. Bus Admittance Matrix N

The bus admittance matrix, or Y matrix, is a numerical I i   YijV j i  1, 2,...N (1)

j 1
model of the power system that characterizes the behaviour of
the network components such as power lines, transformers and
loads using the complex injected currents at the nodes and In matrix representation:
their relation to the complex node voltages, based on
Kirchoff’s current law [1]. [ I ]  [Y ][V ] (2)
Generally, for a network with N independent buses, we can
write the following N nodal equations for a power system where: [I ] : the bus current injection vector; [V ] : the bus
network applying the Kirchoff's current law to bus i will yield: voltage vector; [Y ] : the bus admittance matrix; Yii : the
diagonal elements of bus admittance matrix, are called the
self-admittance of bus i, which equals the sum of all branch

International Scholarly and Scientific Research & Innovation 12(9) 2018 628 ISNI:0000000091950263
 World Academy of Science, Engineering and Technology
International Journal of Energy and Power Engineering
Vol:12, No:9, 2018

admittances connecting to bus i. Yij : the off-diagonal two of these should be known for each bus (Table I).
Assuming the system consists of N buses. The 1th bus is the
elements, of the bus admittance matrix, are the negative of the
slack bus, buses 2 to M are PQ buses and buses (M+1) to N are
admittances buses i and j ( Yij  Y ji ). The off-diagonal
PV buses. The variables and equations are twice the number of
elements are equal to zero if there is no line between buses i the network buses, then 2(N-1)-(M-1) unknowns variables.
and j, obviously, the bus admittance matrix is a sparse matrix. If the bus voltage is expressed using the polar coordinate
D. Real and Reactive Power Injected in a Bus system, the complex voltage, and admittance elements can be
written as:
The complex power delivered to ith bus is given by:
Vi  Vi  i  Vi (cos  i  j sin  i ) (15)
S i  Vi I i* with i=1, 2, 3 … N (3)

Yij  Yij  ij with i, j =1, 2, 3 …N

and
S i*  S gi*  S di*  ( Pgi  Pdi )  j (Qgi  Qdi ) (4)
Yij  Yij (cos  ij  j sin  ij )  G ij  jBij (16)

where S gi : the complex power injection vector generator

where ij=i−j, which is the angle difference between buses i
connecting to bus i; S di : the complex power load vector and j.
connecting to bus i; Pgi: the real power output of the generator The power flow equations in polar coordinate are based on
Open Science Index, Energy and Power Engineering Vol:12, No:9, 2018 waset.org/Publication/10009525

connecting to bus i; Qgi: the reactive power output of the the nonlinear power flow equation gives by (9). The terms V,
generator connecting to bus i; Pdi: the real power load P and Q are represented in per-unit and  terms are
connecting to bus i; Qdi: the reactive power load connecting to represented in degrees. For each PV or PQ bus, we have the
bus i. following real and reactive powers mismatch equations given
The bus current can be represented by bus voltage and by (17) and (18):
power:
N
Pi  Pic  Pi  Pic  Vi V j [Gij cos  ij  Bij sin  ij ]  0 (17)
S S
*
gi
*
di ( Pgi  Pdi )  j (Q gi  Qdi ) j 1
Ii  *
 *
(5)
Vi Vi
N
Qi  Qic  Qi  Qic  Vi V j [Gij sin  ij  Bij cos ij ]  0 (18)
Substituting (5) into (1), we have: j 1

( Pgi  Pdi )  j (Q gi  Qdi ) N

where Pic, Qic are, respectively, the calculated bus real and
Ii    YijV j i  1, 2,...N (6)
Vi *
j 1
reactive power injections.
The power flow equations (17) and (18) are nonlinear and
In the power flow problem, the load demands are known cannot be explicitly inverted. To solve these nonlinear
variables. We define the following bus power injections as: equations the numerical iterative techniques are used [2].

Pi  Pgi  Pdi (7) III. LOAD FLOW SOLUTION

The mathematical recalls of the Gauss Seidel method and
Qi  Q gi  Qdi (8) Newton Raphson polar coordinates method are given in this
section.
Substituting the above two equations into (6), we can get the A. Gauss Seidel Method
general form of power flow equation as The Gauss-Seidel iterative method [3] is the simplest of all
the iterative methods. From (9), we get
N
Pi  jQi  Vi *  YijV j i  1, 2,...N (9)
 
j 1
1  Pi  jQi N
 i  1, 2,...N
Vi 
Yij  Vi *
 
j 1
Y V
ij j

(10)
If we divide (9) into real and imaginary parts, we can get  j i 
two equations for each bus with four variables, that are,
voltage amplitudes V and phases  at load PQ buses, reactive
According to the Gauss-Seidel method, the iteration
power Q and voltage phases V at generator PV buses and
formula (10) can be written as:
active and reactive power (P, Q) at the slack bus, as given in
Table I. The slack bus variables 1 and V1 are omitted, since
they are already known. To solve the power flow equations,

International Scholarly and Scientific Research & Innovation 12(9) 2018 629 ISNI:0000000091950263
 World Academy of Science, Engineering and Technology
International Journal of Energy and Power Engineering
Vol:12, No:9, 2018

1  Pi  jQi i 1 N  Pi
Vi k 1     YijV jk 1  Y V k
 (11) J1 is an (N-1)x(N-1) matrix, and its element is J 1ij 
Yij  Vi k*
j 1 j  i 1
ij j
  j
Pi
Equation (11) is the formulation for the iterative solution of J2 is an (N-1)xM matrix, and its element is J 2 ij  V j
power flow problem by Gauss Seidel method; for the PQ bus, V j
the real and reactive powers are known. Thus, if the initial bus Qi
voltage V̇ i0 is given, we can use (11) to perform the iteration J3 is an Mx(N-1) matrix, and its element is J 3ij 
 j
calculation.
For the PV bus, the bus real power and the magnitude of the Qi
J4 is an MxM matrix, and its element is J 4ij  V j
voltage are known. It is necessary to give the initial value for V j
bus reactive power. The bus reactive power will then be
computed by iterative calculation. That is: If i≠j the expressions for the elements in Jacobian matrix
are as:
  i 1 N

Qik  Im Vi k I ik *   Im Vi k   Yij*V jk 1*   Yij*V jk *   (12) J 1ij  ViV j [Gij sin  ij  Bij cos  ij ] (20)
  j 1 j  i 1  
J 2 ij  ViV j [Gij cos  ij  Bij sin  ij ] (21)
After the iteration is over, all bus real and reactive power, as
Open Science Index, Energy and Power Engineering Vol:12, No:9, 2018 waset.org/Publication/10009525

well as the voltage, are obtained. The power of the slack bus
J 3 ij  ViV j [Gij cos  ij  Bij sin  ij ] (22)
can be obtained by solving:

N J 4 ij  ViV j [Gij sin  ij  Bij cos  ij ] (23)

P1  jQ1  V1 Y
j 1
*
V j*
1j
(13)

If i=j, the expressions for the elements in the Jacobian

The line complex power flow can be also obtained as: matrix are as:

S ij  Pij  jQij  Vi I ij*  Vi 2Yi 0*  Vi (Vi *  V j* )Yij* (14) J 1 ii  Vi 2 Bii  Qi (24)

where Yij is the admittance of the branch i-j and Yi0 is the J 2 ii  Vi 2 Gii  Pi (25)
admittance of the ground branch at the end i.
B. Newton Raphson Method J 3 ii  Vi 2 Gii  Pi (26)
According to the Newton method [4]-[6], the power flow
equations (17) and (18) can be expanded into Taylor series and J 4 ii  Vi 2 Bii  Qi (27)
the following first-order approximation. The result is a linear
system of equations that can be expressed as: The steps for calculation of the Newton power flow
solutions are as follows [1], [2]:
 P   J 1 J 2     1 Given input data.
Q    J J 4  V / V 
(19)
2 Form bus admittance matrix.
   3
3 Assume the initial values of bus voltage.
4 Compute the power mismatch according to (17) and (18).
where:
Check whether the convergence conditions are satisfied.

V / V  V D1 V Max Pi ( k )   1 (28)

 P2   Q2    2   V 2 
 .   .  .  
P     Q          V     Max Qi( k )   2 (29)
PN  Q M   N  V M 
       
V  If (28) and (29) are met, stop the iteration, and calculate the
VD1   2
VM 
line flows and real and reactive powers of the slack bus. If not,
 go to next step.
5 Compute the elements in the Jacobian matrix (19)–(27).
The partial derivatives in each Jacobian matrix block, 6 Compute the corrected values of the bus voltage using
derived from (17) and (18) are given by: (30).
Then compute the bus voltage.

International Scholarly and Scientific Research & Innovation 12(9) 2018 630 ISNI:0000000091950263
 World Academy of Science, Engineering and Technology
International Journal of Energy and Power Engineering
Vol:12, No:9, 2018

Vi k 1  Vi k  Vi k (30) 1 0.02+j0.04pu 2 P =400MW

2

 ik 1   ik   ik (31)
~ Q2=250MVAR
0.01+j0.03pu 0.0125+j0.025pu
P1=?
Q1=? V2=1.050°
3 V3=1.04?pu
7 Return to Step (4) with new values of the bus voltage. V1=1.050°pu
P3=200MW
The Jacobian matrix gives the linearized relationship Q3=?
between small changes in angle and voltage magnitude with ~
the small changes in real and reactive power. This method
begins with initial guesses of all unknown variables (voltage Fig. 1 Standard IEEE 3-bus test system diagram
magnitude and angles at load buses and voltage angles at TABLE II
generator buses). The good initial guess is needed to start the STANDARD EEE 3-BUS TEST SYSTEM CIRCUIT DATA
iterative process [4]. Typically a flat start is an acceptable From
To bus
Resistance Reactance Susceptance Transformer
initial guess. The algorithm stops if the variable increments are bus° R pu X pu B/2 pu TAP
1 2 0.02 0.04 0.03 1
lower than a given tolerance or the number of iterations is
1 3 0.01 0.03 0.02 1
greater than a given limit. In the latter case, the algorithm has 3 0.0125 0.025 0.025 1
2
likely failed to converge. In this case it will be seen that J1=J4
and J2=-J3. Or, in other words the symmetry is restored. The TABLE III
number of elements to be calculated for an N-dimensional STANDARD IEEE 3-BUS TEST SYSTEM BUS DATA
Open Science Index, Energy and Power Engineering Vol:12, No:9, 2018 waset.org/Publication/10009525

Jacobian matrix are only N + N2/2 instead of N2, thus again Bus Power Load
saving computer time and storage. Bus Bus Voltage generated MW
N° type Pu deg MW MVAR MVAR
|V|  PGi QGi PLi QLi Qmin Qmax
IV. SIMULATION RESULTS AND DISCUSSION
1 1 1.05 0 0 0 0 0 0 0
A. Simulation Results 2 3 1.00 0 0 0 400 250 -40 50
The simulation results of the comparative study of Gauss 3 2 1.04 0 200 0 0 0 0 40
Seidel (GS) and Newton-Raphson (NR) methods for power
TABLE IV
flow analysis are presented for different standard IEEE bus NR LOAD FLOW ANALYSIS
test systems [7]. The performances of this comparative study Power Injection Power Generation Load
are assessed from number of iterations, computational time, Bus |V| 
MW MVAR MW MVAR MW MVAR
tolerance value and convergence. The simulation is carried out N° pu deg P Q P Q P Q
using MATLAB version 7.8.0.347 (R2008a) in Intel® Atom 1 1.05 0.00 219.85 236.17 219.85 236.17 0.00 0.00
™ CPU N450 @1,66GHz. 2 0.96 -2.51 -400 -250 0.00 0.00 400 250
The standard IEEE 3-bus test system diagram is given in 3 1.02 -0.67 200 39.18 200 39.18 0.00 0.00
Fig. 1. The circuit data and the bus data of the standard IEEE Total 19.85 25.35 419.85 275.35 400 250
3-bus test system are respectively given in Tables II and III.
TABLE V
The NR and GS comparative results of power flow analysis GS LOAD FLOW ANALYSIS
for a simple standard IEEE 3-bus test system are respectively Bus |V|  Power Injection Power Generation Load
given in Tables IV and V. The line flow and losses of a N° pu deg P Q P Q P Q
standard IEEE 3-bus test system are given in Table VI. MW MVAR MW MVAR MW MVAR
The comparison results of convergence, number of 1 1.05 0.00 219.82 235.28 219.82 235.28 0.00 0.00
iterations and computational time using NR and GS methods 2 0.96 -2.51 -399.99 -249.99 0.01 0.01 400 250
are given respectively in Figs. 2, 3 and 4. The summary of the 3 1.02 -0.70 199.99 40.00 199.99 40 0.00 0.00
obtained results are given in Table VII. Total 19.82 25.29 419.82 275.29 400 250

TABLE VI
LINE FLOW AND LOSSES OF STANDARD IEEE 3- BUS TEST SYSTEM (NR)
Line Loss
From Bus To Bus P(MW) Q (MVar) From Bus To Bus P(MW) Q(MVar)
P(MW) Q(MVar)
1 2 184.58 148.14 2 1 -174.42 -127.82 10.16 20.32
1 3 35.24 92.65 3 1 -34.35 -89.98 0.89 2.68
2 3 -225.57 -117.11 3 2 234.35 134.66 8.77 17.55
Total Loss 19.83 40.54

International Scholarly and Scientific Research & Innovation 12(9) 2018 631 ISNI:0000000091950263
 World Academy of Science, Engineering and Technology
International Journal of Energy and Power Engineering
Vol:12, No:9, 2018

-3
IEEE 3-Bus x 10 IEEE 5-Bus
0.025 14

12
0.02
10  GS: Iter=22 & Time=1.09s

Voltage Mismatch
Voltage Mismatch

0.015  GS: Iter=11 & Time=1.06s 8

6 NR: Iter=11 & Time=1.23s

0.01 NR: Iter=8 & Time=1.32s

 4

0.005
2

0
0
2 3 4 5 6 7 8 2 3 4 5 6 7 8
Iterations Number Iterations Number
IEEE 14-Bus IEEE 30-Bus
0.06

0.04

0.05
0.035
Open Science Index, Energy and Power Engineering Vol:12, No:9, 2018 waset.org/Publication/10009525

0.03
0.04
Voltage Mismatch

Voltage Mismatch

0.025

0.03  GS: Iter=199 & Time=1.54s

0.02  GS: Iter=77 & Time=1.09s

0.015
0.02

0.01
0.01
0.005
NR: Iter=7 & Time=1.45s NR: Iter=5 & Time=1.57s
0  
0
2 3 4 5 6 7 8 2 3 4 5 6 7 8
Iterations Number Iterations Number
IEEE 57-Bus
0.06

0.05

0.04  GS: Iter=308 & Time=2.93s

Voltage Mismatch

0.03

0.02

0.01
NR: Iter=11 & Time=2.17s


0
2 3 4 5 6 7 8
Iterations Number

Fig. 2 Comparison of convergence for different standard IEEE test bus system using NR method and GS method

International Scholarly and Scientific Research & Innovation 12(9) 2018 632 ISNI:0000000091950263
 World Academy of Science, Engineering and Technology
International Journal of Energy and Power Engineering
Vol:12, No:9, 2018

Iter=f(Nbus)
100 which has a quadratic convergence. Hence, the GS method
requires more number of iterations to get a converged solution
90
as compared to the NR method. In the GS method, the number
80  Gauss seidel of iterations increases directly as the size of the system
70 increases. In contrast, the number of iterations is relatively
constant in NR method. They require about 10 iterations for
Iterations Number

60
convergence in large systems.
50

40 V. CONCLUSION
30 This paper has described a comparison of NR and GS
20 methods for Power Flow Analysis. Five different standard
Newton Raphson
10
IEEE bus test systems are considered to investigate the

effectiveness of the proposed methods. The compared results
0
0 5 10 15 20 25 30 show that the NR is the most reliable method because it has
Number of Bus
the least number of iteration and converges faster.
Fig. 3 Comparison of number of iterations for different standard bus
test system using NR method and GS method REFERENCES
[1] A. Shahriari, H. Mokhlis, Ab Halim Abu Bakar “Critical Reviews Of
Time comput=f(Nbus)
Load Flow Methods For Well, Ill And Unsolvable Condition” Journal of
Open Science Index, Energy and Power Engineering Vol:12, No:9, 2018 waset.org/Publication/10009525

3
Electrical Engineering, Vol. 63, No. 3, , pp. 144–152, 2012
2.8 [2] G. Leoniopoulos “Efficient starting point of load-flow equations”
International Journal of Electrical Power & Energy Systems, Vol. 16,
2.6 Issue 6, pp. 419-422, December 1994
2.4 Gauss Seidel 
[3] H. S. Huang and C. N. Lu “Vector implementations of fast decoupled
and Gauss-Seidel load flows” Electrical Power & Energy Systems, Vol.
Computational time (s)

2.2 17, No. 5, pp. 355-358, 1995

[4] Abilio Manuel Variz, Vander Menengoy da Costa, José Luiz R. Pereira,
2 Nelson Martins “Improved representation of control adjustments into the
1.8
Newton–Raphson power flow” International Journal of Electrical Power
& Energy Systems, Vol. 25, Issue 7, pp. 501-513, September 2003
1.6 Newton Raphson [5] W. Hubbi and A. Refsum “Starting algorithm and modification for
 Newton-Raphson load-flow method” International Journal of Electrical
1.4 Power & Energy Systems, Vol. 5, Issue 3, ,pp. 166-172, July 1983
1.2
[6] S. Mallick, D.V. Rajan, S.S. Thakur, P. Acharjee, S.P. Ghoshal
“Development of a new algorithm for power flow analysis” International
1 Journal of Electrical Power and Energy Systems Vol. 33, pp. 1479–
0 10 20 30 40 50 60
1488, 2011
Number of Bus
[7] H.Abaal, E.Talbi and R.Skouri “A Comparison of Newton Raphson and
Fig. 4 Comparison computational time for different standard IEEE Gauss Seidel Methods for Power Flow Analysis” 1st International
Symposium on Computational Mathematics and Engineering Sciences.
test bus system using NR method and GS method Faculty of Sciences and Techniques Errachidia, Morocco, Mar 3-6, 2016
[8]
TABLE VII
PERFORMANCES OF THE COMPARED STUDY
IEEE Bus System test 3 5 14 30 57
GS
Number of Iterations 11 22 77 199 308
Computation Time (s) 1.06 1.09 1.09 1.54 2.93
NR
Number of Iterations 8 11 7 5 11
Computation Time (s) 1.32 1.23 1.45 1.57 2.17

B. Discussion
The GS method is relatively easy to program however
programming of NR method is more involved and becomes
more complicated mostly if the buses are randomly numbered
and the storage requirements are more (Jacobian elements).
The time per iteration in NR method is larger than in the GS
method. In NR method, the time per iteration increases
directly as the number of buses.
The number of iterations is determined by the convergence
characteristic of the method. The GS method exhibits a linear
convergence characteristic as compared to the NR method

International Scholarly and Scientific Research & Innovation 12(9) 2018 633 ISNI:0000000091950263