1

An Equivalent pi Network Model for

Power System State Estimation with
Network Parameter Errors
Amit Jain, Member, IEEE, and Sivaramakrishnan Raman

small in nature would generate inconsistencies in post-analysis

Abstract—With the role of state estimation in energy like security assessment.
management systems gaining weight every day, the demand for Network parameters are known to comprise line resistance,
its accountability is an imperative, to say the least. The reliability reactance, and line charging conductance and susceptance in
of the network database is thus a key factor. This paper employs
majority of the cases. Another parameter equally or even more
the two bus π equivalent model for every line in the power
network to approach the network parameter estimation problem. significant in the network is the position of the transformer
The network π model is known to combine the effect of fixed tap tap. The lack of communication between the substation and
transformers in the equivalent’s line impedance and line the control center often leads to wrong information of the
charging admittances. The method aims to correct the equivalent transformer tap. Thus, the transformer tap too is to be given
line parameters of the suspicious branch rather than evaluating due attention in this case. Parameter estimation may not boast
every network component individually based on whether there is
of abundant literature as it is still constantly evolving into a
a transformer or not. The suspicious branch is determined by
ranking the branches in order of their respective branch indices mature prospect. Still, the literature is not sparse either.
obtained in a particular way, from normalized residuals of Different methods have contributed towards its cause since
measurements involved. The technique employs the state vector days as early as when the concept of the generalized state
augmentation approach to correct and store the updated estimator was introduced [1] and they can be classified
equivalent parameters of the erroneous branch, which is broadly into two major groups [2], [3]. The concept of
sufficient for an accurate state estimation. Besides this, the
residual analysis [4]-[6] rules one of them and the other is
correct value of the parameters can always be estimated from the
equivalent network if needed. based on augmenting the existing state vector. The residual
Illustrations on the IEEE 14 and 30 bus systems have been analysis technique was detailed primarily in [4], wherein, the
furnished to validate the approach for different cases of measurement residuals provide information of existence of
parameter errors on any single branch. gross error. The residual analysis technique is used entirely for
the estimation of parameters in [4], [5]. Measurement
Index Terms— Branch, measurement, parameter error, π residuals are also used only to the extent of determining
equivalent, residual, state estimation, transformer tap. suspicious branches, but not for parameter estimation as such
[6]. The state vector augmentation approach [7]-[12] considers
I. INTRODUCTION
the conventional state variables and the suspicious parameters

S TATE estimation in power systems today holds key to

more than for what it was utilized in the initial stages. It
has become nothing short of a backbone to the field of energy
as the combined state vector. The state augmentation can be
carried out by either the normal equations method [7]-[10] or
the Kalman filter [11], [12], which is a recursive algorithm.
management studies. It caters to the basic requirement in the The objective in the present paper is to provide a common
system security domain, specifically the contingency analysis. platform for parameter estimation on branches with or without
In view of the stated, state estimation bears the heavy transformer taps by using the π equivalent for every branch in
responsibility of picturing the state of the system as best and the network. The π network’s parameters contain the effects of
accurate as possible. This in turn shifts focus onto the the fixed transformer tap, if any. The idea behind this is to
dependability of fixed data which the state estimation works estimate the parameters of the equivalent network, which has
on. The process of state estimation places huge confidence on only the general elements and no specific tap to estimate. This
the network database which is assumed to be fixed in nature. is specifically useful if the aim is to attain the correct state
Hence, any error in the concerned network would culminate in estimate. The equivalent parameters are updated whenever
erroneous interpretation which the state estimator would be estimated and are sufficient for accurate state estimation.
completely unaware of. If not accounted for, the errors, even Besides this, if the original parameter estimates are the need of
the hour, the equivalent parameters can always be converted
Amit Jain is Head, Power Systems Research Center, IIIT-Hyderabad, AP back with ease as shown in a further section.
500032 India (e-mail: [email protected]).
Sivaramakrishnan Raman is with Power Systems Research Centre, IIIT,
The equivalent network does not affect the identification
Hyderabad, India. (email: [email protected]). and detection of erroneous branches because the method
 2

employs measurement residuals for the purpose. A branch Side j:

index based on these residuals, described in a later section, is
I j = V j yo + (V j − aVi ) yij
obtained for each branch. The branches are ranked in order to
determine the suspicious one. The equivalent network’s = V j ( yo + yij ) − aVi yij
parameters are augmented to the state vector using the normal
equations method to obtain their estimate. The correct Adding and subtracting the term ayijVj would yield in (4).
estimated network parameters can also be extracted from the
estimates of the equivalent. This paper addresses any type of I j = V j ( yo + (1 − a ) yij ) + (V j − Vi ) ayij (4)
parameter error on a single branch.
The equivalent π network parameters are given by (5) with
II. EQUIVALENT MODEL APPROACH reference to (3) and (4) respectively. Fig. 2 depicts the
equivalent network.
A. Representation of π Network
A line or branch with a transformer can be represented as y eq = ayij
the π network. A transformer of ratio 1:a connected between
two buses i and j is shown in fig. 1. The transformer is yoieq = a 2 yo + a ( a − 1) yij (5)
connected to the bus i. The line admittance is given by yij and
yojeq = yo + (1 − a ) yij
the leakage admittance equal on both sides, is given by yo.
Voltage and current are denoted as V and I respectively.

Fig. 2. Equivalent Π Network Representation

Equation (5) can be split into the rectangular components for

Fig. 1. Transformer on line connecting ends i and j.
the requirement of use in state estimation. Equations (6), (7)
and (8) depict the same.
Transformer principles would confirm (1) and (2).
g eq = agij
(6)
Vt = aVi (1) beq = abij
I
It = i (2) g oieq = a 2 g o + a ( a − 1) gij
a (7)
Simple Kirchhoff laws would support the voltage current boieq = a 2bo + a ( a − 1) bij
relations for both the sides of the line.
Side i:
g ojeq = g o + (1 − a ) gij
(8)
Ii
= I i' + I oi bojeq = bo + (1 − a ) bij
a If there is no transformer present, a can be replaced the
= ( aVi − V j ) yij + aVi yo value of 1. If the transformer tap is of the form b:1, the same
procedure can be adopted with a = 1 / b. If the transformer is
I i = a ( aVi − V j ) yij + a 2Vi yo on the other end, the ends i and j are interchanged in the same
equations.
= a 2Vi ( yij + yo ) − ayijV j B. Significance
Adding and subtracting the term ayijVi would yield in (3). The equivalent terms given above are used to build the
network admittance matrix. The same terms are subjected to
I i = Vi ( a 2 yo + a ( a − 1) yij ) + (Vi − V j ) ayij (3) parameter estimation process described later. Once estimated,
the new values are stored in the database. This removes the
necessity of estimating the tap position as a separate entity.
The state estimation can be carried out successfully with the
 3

equivalent value itself. The model makes true sense when it where R is the measurement error covariance matrix, H is the
comes to transformer connected buses. measurement Jacobian and G is the state estimation gain
It is not always necessary to estimate all the six parameters matrix.
of the equivalent shown in (6), (7) and (8). A transformer is The measurement residuals detect any error only when
plainly used to shift the voltage level and is generally not a there is enough local redundancy. If the measurement is
part of a line having line impedance and line charging critical in nature, the error in it goes unidentified.
admittances. Such lines mostly have only the transformer
B. Identification by Branch Index
reactance in place. If this is the case, all the six terms
mentioned need not be estimated. If this is the case, only the The high normalized residuals need to be linked to the
three reactive components would have the transformer effect, correct branch to ensure correct estimation. The work
and would have to be estimated. On the other hand, if the line pertaining to this paper identifies branches by ranking them in
has no transformer, the π model would not have different order of their respective indices. For a given branch,
leakage admittances on either side thus reducing the number normalized residuals would arise from its branch flows or the
of parameters to be estimated, to four. power injections at the end nodes. The approach used here
The estimation of the equivalent parameters would yield in takes into account, all normalized residuals greater than 3.0,
the same state estimate as the estimation of the individual pertaining to a branch. The maximum among these is the
parameters. However, if the estimation of the individual index of that branch, as shown in (12).
parameter is also required, they can always be obtained from Indexi = max ( rjN )
the equivalent anytime, as shown in the further section on (12)
normal equations method of parameter estimation. i ∈ [1, m ] , j ∈ {rjN > 3.0}
where m is the number of branches and j is corresponding to a
III. PARAMETER ESTIMATION
measurement associated with the branch i. The branch with
A. Detection of Parameter Error the maximum index is chosen to have its parameters
estimated.
The method employs the normalized measurement
Critical Tuples
residuals [2] to detect any parameter error. A parameter error
In case, two branches end up with identical values for their
is analogous to the correlated errors in any of the
indices, the procedure has not been able to clearly identify any
measurements adjacent to the erroneous branch. These
one branch. The parameter error could be in either branch.
measurements are the power flow on the erroneous branch and
This generally happens in case of critical tuples which has
the power injections on the end nodes. Thus, the measurement
been discussed in [8]. The identical value may also be
residuals provide the window to lookout for parameter errors
attributed to the normalized residual of a power injection
as well. The normalized measurement residual, when
common to both the branches. In such cases, counter-checking
calculated to be above a threshold value, generally 3.0 [2],
the normalized residuals of the branch flows so as to identify
indicates an indirect error in the calculation of that
which branch has a higher value may help provided they too
measurement due to the actual error in one or more branch
are not similar.
parameters related to that measurement. This only detects the
error but does not identify the branch. The erroneous branch is C. Estimation by Normal Equations Method
identified by the branch index described in the immediately The normal equations method of augmenting the state
next section. vector clubs the conventional state variables with the
The residual for a measurement zi is given by (9). The suspicious parameters to be estimated together using the
normalized value is given by (10). Newton-Raphson technique, as shown in (13).
∧
ri = zi − hi ( x) (9)
∧
x = ⎡⎣ xV , xθ | x p ⎤⎦ (13)
where x is the estimated state of the prior state estimation, zi is where V, θ and p represent voltage magnitude, its angle and
the vector of given measurement set and hi is the vector of parameter respectively.
calculated measurements. The residuals can be determined One area of concern with this approach is regarding the flat
only after the state estimation process is done. start of the variables generally done before the start of the first
iteration. The flat start leads to Jacobian singularity on
ri account of last column of the Jacobian turning null very often.
ri N = (10) To counter this, the parameters are augmented to the state
Ωii vector after the first iteration.
Ω is the covariance matrix of the measurement residuals,
D. Equivalent and Individual Estimates
given in (11).
The approach in this paper directly gives the final estimates
Ω = R − HG −1 H T (11) eq eq eq
of the equivalent network, given by y , yoi , yoj . This
 4

would suffice for the purpose of state estimation to yield Table II gives the ranking of the branch indices for each of
correct states because at each iterative step, the network the three cases. The indices identify the branches distinctly
admittance matrix is updated with the equivalent element and correctly in each of the three cases.
estimates. However, the method is not incapable of updating
the individual parameters. The equivalent parameters are TABLE II
BRANCH INDEX RANKING
nothing but functions of these. Thus, if the equivalent network
estimates are given by k1, k2 and k3, the solution of (14) in 3
variables would yield the individual parameter estimates. Any Branch
Rank of Highest Corresponding
simple non-linear solution technique would yield the required Subjected
Normalized Residual Branch
results. To Error
k1 = ayij
12.5758 13-14
k 2 = a yo + a ( a − 1) yij
2
(14) 13-14 10.7922 9-14

k 3 = yo + (1 − a ) yij 10.196 12-13

This set of equations can be written twice to solve for g and b 6.63021 7-9
separately if required. 7-9 3.75991 4-7
Transformer Tap 3.75991 7-8
The variable estimated in (14) is a, which may not
necessarily be the tap ratio. If the original ratio was inverted to 6.46194 3-4
convert to 1:a form for developing this equivalent model, the 3-4 4.8282 2-3
true estimate would be 1 / a. 3.78833 2-4

IV. SIMULATION RESULTS

2) Case II: Multiple Errors on Single Branch
The technique discussed in the paper thus far has been This case deals with introduction of error in more than one
implemented on standard IEEE 14 and 30 bus systems data component except transformer tap, on a single branch. Table
[13], the results of which have been furnished in detail below.
III shows three different branches having errors on r, x and bs
This paper currently deals with single branch errors. Different
at the same time. But only one branch is considered to have
types of errors as depicted in the three cases given below have
the errors.
been handled on single branches. The state estimation results
after equivalent parameter estimation have been compared
with the state estimation results with the parameter errors.
TABLE III
A. IEEE 14 bus system MULTIPLE ERRORS ON SINGLE BRANCH

1) Case I: Single Error on Single Branch

This case deals with introduction of an error in any one of Branch
Component Original Error
the branch components apart from transformer tap, on a single Subjected
in Error Value Value
branch. Table I shows particulars of three different branches To Error
having a single error, but not simultaneously. They are three
separate cases and are handled individually. The symbols r, x r 0.05403 1.05403
and bs denote line resistance, line reactance and leakage 1-5 x 0.22304 1.22304
susceptance respectively. The true value and error value are bs 0.0492 0.2492
given too.
r 0.05811 1.05811
TABLE I 2-4 x 0.17632 1.17632
SINGLE ERROR ON SINGLE BRANCH
bs 0.0374 0.2374
r 0.05695 1.05695
Branch
Component Original Error 2-5 x 0.17388 1.17388
Subjected
in Error Value Value bs 0.034 0.234
To Error

13-14 r 0.06701 1.06701 Table IV gives the ranking of the branch indices for each
of the three cases. The indices identify the branches distinctly
7-9 x 0.11001 1.11001 and correctly in each of the three cases.
3-4 bs 0 0.7
 5

x 0.25202 0.55202
5-6
TABLE IV tap 0.932 0.862
BRANCH INDEX RANKING

Table VI gives the ranking of the branch indices for both

Branch
Rank of Highest Corresponding the cases. The indices identify the correct branch clearly on
Subjected
Normalized Residual Branch both occasions.
To Error
TABLE VI
20.3836 1-5 BRANCH INDEX RANKING
1-5 13.4019 2-5
Branch
13.4019 4-5 Rank of Highest Corresponding
Subjected
15.3021 2-4 Normalized Residual Branch
To Error
2-4 8.77077 2-3
8.77077 2-5 9.00789 4-9
4-9 6.05973 4-7
11.2689 2-5
2-5 4.56121 7-9
6.60812 1-5
6.60812 4-5 4.31149 5-6
5-6 2.34741 1-5
3) Case III: Transformer Tap Error on Single Branch 2.34741 2-5
This case deals with introduction of error in the transformer
tap and the reactance of a single branch. Table V shows two
The branches indicated by the branch index based on
different branches having the tap error. This case too pertains highest normalized residuals are estimated for their equivalent
to one branch at a time. parameters. Once they are estimated and updated, the state
estimation results can be computed. Due to space constraints,
TABLE V
TRANSFORMER TAP ERROR ON SINGLE BRANCH
the state estimation results for only the case of transformer tap
error in branch 4-9 from Table V have been shown. Table VII
Branch compares the state estimation results obtained when there is
Component Original Error no error with those obtained with error and corrected
Subjected
in Error Value Value equivalent parameters.
To Error
x 0.55618 0.25618
4-9
tap 0.969 0.869

TABLE VII
STATE ESTIMATION RESULTS

No Error With Error Estimated

Bus V θ V θ V θ
1 1.06 0 1.00439 0 1.06 0
2 1.045 -4.97895 0.988044 -5.54695 1.045 -4.97895
3 1.01 -12.7128 0.951511 -14.226 1.01 -12.7128
4 1.01862 -10.3201 0.958757 -11.5462 1.01862 -10.3201
5 1.02026 -8.77904 0.963251 -9.77973 1.02026 -8.77904
6 1.07 -14.2169 1.03053 -15.0642 1.07 -14.2169
7 1.06195 -13.3629 1.03727 -13.9027 1.06195 -13.3629
8 1.09 -13.3629 1.06805 -13.8438 1.09 -13.3629
9 1.05635 -14.9406 1.04488 -15.2161 1.05635 -14.9406
10 1.05133 -15.0983 1.03564 -15.4504 1.05133 -15.0983
11 1.05708 -14.7893 1.03049 -15.4033 1.05708 -14.7893
12 1.05522 -15.0714 1.01718 -16.0106 1.05522 -15.0714
13 1.05044 -15.1529 1.0137 -16.0458 1.05044 -15.1529
14 1.03579 -16.0325 1.01176 -16.5925 1.03579 -16.0325
 6

8.69994 29-30
Table VII shows clearly that the state estimation results
29-30 8.69994 27-30
after the estimation of equivalent parameters are the same
as without the error. In addition, when (14) is solved for 8.47777 27-29
these numerous cases given above, the individual
parameters match exactly with the original values 2) Case II: Multiple Errors on Single Branch
mentioned in the tables. This case deals with introduction of error in more than
B. IEEE 30 bus system one component except transformer tap, on a single branch.
Table X shows three different branches having errors on r,
1) Case I: Single Error on Single Branch
x and bs at the same time. But only one branch is
Table VIII shows particulars of three different branches
considered to have the errors.
having a single error, but not simultaneously. They are
three separate cases and are handled individually.
TABLE VIII TABLE X
SINGLE ERROR ON SINGLE BRANCH MULTIPLE ERRORS ON SINGLE BRANCH

Branch
Branch Component Original Error
Component Original Error Subjected
Subjected in Error Value Value
in Error Value Value To Error
To Error
r 0.0472 1.0472

2-4 r 0.0570 1.0570 2-5 x 0.1983 1.1983

4-6 x 0.0414 1.0414 bs 0.0418 0.2418

29-30 bs 0 0.5 r 0.0267 1.0267

6-7 x 0.0820 1.0820

Table IX gives the ranking of the branch indices for bs 0.0170 0.2170
each of the three cases. The indices do not identify the r 0.0169 1.0169
branches distinctly in the case of error in the leakage 6-28 x 0.0599 1.0599
susceptance of branch 29-30. bs 0.0130 0.2130
Identical Normalized Residuals:
The maximum normalized residuals of both branches
Table XI gives the ranking of the branch indices for
27-30 and 29-30 are identical. This is due to the maximum
each of the three cases. The indices identify the branches
residual arising from the common power injection at bus
distinctly and correctly in each of the three cases.
30. Bus 30 is connected to only two buses 27 and 29. So,
the power injection at bus 30 is a measurement adjacent to TABLE XI
both buses 27 and 29. However, on comparing the BRANCH INDEX RANKING
normalized residuals of the branch flows on the two
branches, the branch flow on 29-30 has a normalized Branch
residual greater then 8 whereas that on branch 27-30 has Rank of Highest Corresponding
Subjected
just above 0.5. It is clear from this that the erroneous Normalized Residual Branch
To Error
branch is 29-30.
21.6674 2-5
TABLE IX
BRANCH INDEX RANKING 2-5 16.5677 5-7
12.786 2-4
Branch
Rank of Highest Corresponding 10.6371 6-7
Subjected
Normalized Residual Branch 6-7 7.97217 5-7
To Error
5.53609 2-6
11.9403 2-4 5.42177 6-28
2-4 7.78801 3-4 6-28 3.44432 8-28
7.78801 4-6 3.44432 28-27
20.2966 4-6
4-6 11.7725 3-4 3) Case III: Transformer Tap Error on Single Branch
11.7725 4-12 Table XII shows two different branches having the tap
 7

error. This case too pertains to one branch at a time. XIV compares the state estimation results obtained when
there is no error with those obtained with error and
TABLE XII
corrected equivalent parameters.
TRANSFORMER TAP ERROR ON SINGLE BRANCH
TABLE XIII
BRANCH INDEX RANKING
Branch
Component Original Error
Subjected
in Error Value Value Branch Highest
To Error Corresponding
Subjected Normalized
Branch
x 0.2560 0.5560 To Error Residual
4-12
tap 0.932 0.832
5.0062 4-12
x 0.5560 0.2560
6-10 4-12 2.85563 2-4
tap 0.969 0.869
2.85563 3-4

Table XIII gives the ranking of the branch indices for 10.0644 6-10
both the cases. The indices identify the correct branch 6-10 5.64865 6-9
clearly on both occasions. 4.38077 9-10
The branches indicated by the branch index based on
highest normalized residuals are estimated for their
equivalent parameters. Due to space constraints, the state
estimation results for only the case of transformer tap error
in branch 4-12 from Table XII have been shown. Table
TABLE XIV
STATE ESTIMATION RESULTS

No Error With Error Estimated

Bus V θ V θ V θ
1 1.05997 0 1.04273 0 1.05996 0
2 1.04297 -5.34816 1.02515 -5.51571 1.04296 -5.3482
3 1.02068 -7.52967 1.00146 -7.69074 1.02067 -7.52976
4 1.01169 -9.28128 0.991674 -9.44705 1.01169 -9.28141
5 1.00997 -14.1623 0.991864 -14.645 1.00996 -14.1624
6 1.0102 -11.0616 0.991897 -11.4469 1.01019 -11.0617
7 1.00233 -12.8618 0.983997 -13.2959 1.00232 -12.8619
8 1.00997 -11.8105 0.991664 -12.2129 1.00996 -11.8106
9 1.05084 -14.1068 1.04116 -15.0713 1.05084 -14.1064
10 1.04505 -15.6973 1.03891 -16.8562 1.04505 -15.6966
11 1.08195 -14.1072 1.07328 -15.1103 1.08195 -14.1067
12 1.05708 -14.9378 1.06591 -17.2433 1.05708 -14.9362
13 1.07097 -14.9378 1.07906 -17.1475 1.07098 -14.9362
14 1.04224 -15.829 1.04972 -17.8828 1.04224 -15.8274
15 1.03765 -15.9201 1.04344 -17.8765 1.03766 -15.9186
16 1.04433 -15.5226 1.0463 -17.2031 1.04433 -15.5214
17 1.03983 -15.859 1.03535 -17.1524 1.03983 -15.8582
18 1.0281 -16.5369 1.02914 -18.1825 1.0281 -16.5358
19 1.02559 -16.7117 1.02443 -18.2237 1.02559 -16.7107
20 1.02967 -16.5155 1.02761 -17.9678 1.02967 -16.5146
21 1.03208 -16.161 1.02579 -17.3409 1.03208 -16.1603
22 1.03428 -16.0829 1.02895 -17.3209 1.03428 -16.0822
 8

23 1.02723 -16.3038 1.02941 -18.0118 1.02723 -16.3026

24 1.02188 -16.4669 1.01786 -17.8022 1.02188 -16.4661
25 1.01736 -16.0504 1.00849 -17.0302 1.01736 -16.05
26 0.999619 -16.4711 0.990864 -17.4414 0.999617 -16.4707
27 1.02322 -15.5303 1.01181 -16.3191 1.02322 -15.5301
28 1.00677 -11.6849 0.988792 -12.1027 1.00676 -11.6849
29 1.00338 -16.7601 0.99247 -17.6118 1.00338 -16.7599
30 0.991903 -17.6426 0.981041 -18.5183 0.991902 -17.6424
[9] M. R. M. Castillo, J. B. A. London, and N. G. Bretas, "Identification and
Table XIV shows clearly that the state estimation results estimation of power system branch parameter error," Power & Energy
Society General Meeting, IEEE, pp.1-8, 26-30 July 2009.
after the estimation of equivalent parameters are very similar [10] P. Teixeira, S. Brammer, W. Rutz, W. Merritt, and J. Salmonsen, “State
to those without the error. In addition, when (14) is solved for estimation of voltage and phase-shift transformer tap settings,” IEEE
these numerous cases given above, the individual parameters Transactions on Power Systems, vol. 7, no. 3, pp. 1386–1393, Aug.
1992.
match exactly with the original values mentioned in the tables.
[11] A. Debs, "Estimation of Steady-State Power System Model Parameters",
IEEE Transactions on Power Apparatus and Systems, vol. PAS-93, No.
V. CONCLUSION 5, pp. 1260-1268, 1974.
[12] I. Slutsker, and K. Clements, "Real Time Recursive Parameter
The paper motivates the idea of employing a π equivalent Estimation in Energy Management Systems," IEEE Transactions on
network for every branch in the system for state estimation of Power Systems, Vol. 11(3), pp. 1393-1399, August 1996.
a system affected by branch parameter and transformer tap [13] Power System Test Case Archive, Electrical Engineering, University of
Washington, http://www.ee.washington.edu/research/pstca.
errors. The equivalent network provides a common structure
for network lines and transformer branches as it includes the
VII. BIOGRAPHIES
transformer effect in the equivalent branches. The parameter
estimation algorithm estimates the correct equivalent
Amit Jain graduated from KNIT, India in Electrical
parameters rather than correcting the original network Engineering. He completed his masters and Ph.D.
parameters and taps separately. However, if the individual from Indian Institute of Technology, New Delhi,
parameter estimate values are also required, the estimates of India. He was working in Alstom on the power
SCADA systems. He was working in Korea in 2002
the equivalent can readily be used to estimate the parameters. as a Post-doctoral researcher in the Brain Korea 21
Illustrations of the method are provided on IEEE 14 and 30 project team of Chungbuk National University. He
bus systems for different scenarios including one case of was Post Doctoral Fellow of the Japan Society for
the Promotion of Science (JSPS) at Tohoku
identical values of normalized residuals. This paper caters to
University, Sendai, Japan. He also worked as a Post
single branch errors. The authors are presently working Doctoral Researcher at Tohoku University, Sendai, Japan. Currently he is
towards catering to errors in multiple and adjoining branches heading, Power Systems Research Center at IIIT, Hyderabad, India. His fields
and that work will be published in future. of research interest are power system real time monitoring and control,
artificial intelligence applications, load forecasting, power system planning
and economics, electricity markets, renewable energy, reliability analysis, GIS
VI. REFERENCES applications, parallel processing and nanotechnology.
[1] O. Alsac, N. Vempati, B. Stott, and A. Monticelli, "Generalized State
Estimation," IEEE Transactions on Power Systems, Vol. 13(3), pp. Sivaramakrishnan Raman is pursuing his Masters at
1069-1075, August 1998. Power Systems Research Center, International
[2] A. Abur and A. Gomez-Exposito, “Power System State Estimation: Institute of Information Technology, Hyderabad,
Theory and Implementation”, New York: Marcel Dekker Inc, 2004. India. He received his B. Tech degree from SASTRA
[3] P. Zarco, and A. Gomez, "Power System Parameter Estimation: A University, Thanjavur, India in 2008. His areas of
Survey," IEEE Transactions on Power Systems, vol. 15(1), pp. 216-222, interest include power system monitoring and control
February 2000. applications, protection, load flow, state estimation,
[4] T. Van Cutsem, and V. Quintana, "Network Parameter Estimation Using voltage stability and reactive power control.
Online Data with Application to Transformer Tap Position Estimation,"
IEE Proceedings, Vol. 135, Pt C, No. 1, pp. 31-40, January 1988.
[5] W. Liu, F. Wu, and S. Lun, "Estimation of Parameter Errors from
Measurement Residuals in State Estimation," IEEE Transactions on
Power Systems, Vol. 7(1), pp. 81-89, February 1992.
[6] M. B. Do Coutto Filho, J. C. S. de Souza, and E. B. M. Meza,
"Correcting electrical network parameters," Power & Energy Society
General Meeting, IEEE, pp.1-7, 26-30 July 2009.
[7] Jun Zhu, and A. Abur, "Identification of network parameter errors,"
IEEE Transactions on Power Systems, vol.21, no.2, pp. 586- 592, May
2006.
[8] Jun Zhu, and A. Abur, "Identification of network parameter errors using
phasor measurements," Power & Energy Society General Meeting,
IEEE, pp. 1-5, 26-30 July 2009.