Vous êtes sur la page 1sur 4

Presentation 8.

1

OMICRON electronics GmbH 2011 Instrument Transformer Measurement Forum
Principle of a Multipoint Compensation of CT Ratios and Phase Angles
for High Accuracy Electricity Meters
Erik Etien, University of Poitiers, France
Jean-Franois Tissier, ITRON, France

Introduction ction
Current transformers using magnetic cores are
widely used for metering purposes in electrical
power systems; they are connected to the high
power on primary side and to the metering
equipment on the secondary side. To guarantee
correct billing over all the dynamic of the primary
current, the complete chain must provide high
accuracy (up to 0.1%). Then It is essential to test
and calibrate the metering current transformer
(CT), to guarantee the accuracy of the entire
metering chain.
The proposal of the research is to compensate the
errors introduced by the CT not only at the main
frequency 50Hz or 60Hz and in one point of the
dynamic (single point compensation) but in a wider
domain of frequency and in a large dynamic of the
primary current using a multi point compensation
(Fig.1).
OMICRON CT Analyzer offers new perspectives in
the tests of CT used for metering according to
IEEE and IEC standards. With the measurements
delivers by the CT Analyzer, an electro technical
model of the CT sufficient for the frequency
bandwidth computed by the electricity meter can
be reconstituted.
Current transformer (CT) modeling
Transfer functions
The standard model of a current transformer is
classically presented on Fig. 2 where :
the current to be measured is noted i
p
(t).
r
1
and l
1
are respectively the primary
winding resistance and the leakage
inductance.
N
1
and N
2
are the number of turns (with
N
1
=1).
Lm is the magnetizing inductance.
R
f
is the parameter traducing iron losses,
hysteresis losses and supplementary
losses.
the measured current is noted i
s
(t).
R
CT
and l
2
are respectively the secondary
winding resistance and leakage
inductance.
Rs and Ls constitute Zs, the load
impedance (current channel impedance of
the meter)

Fig. 1 Principle of error compensation


Fig. 2 Current transformer model, f=50Hz, constant
induction.

The transfer function betwwen i
p
(p) et i
s
(p) (p :
Laplace variable) is:
, ) ' ' )( (
) (
) (
) (

+ + + = = p L R
p L R
K
p i
p i
p K
m f
N
s
p 1 1
1



(1)
with
s CT
R R R + = ' et
s
L l L + =
2
' .

In the ideal case ( =
m
L , =
f
R ), equation (2) is
reduced to
N
K p K = ) ( .
In other hand, the inverse transfer function is
defined by:

i
p
(t)
u
p
(t)
r
1
l R
ct
l
2
R
f
L
m

i
s
(t)
i
m
(t
)
i
p
(t)
N
2
N
1

u
s
(t)
R
s
L
s
Presentation 8.2

OMICRON electronics GmbH 2011 Instrument Transformer Measurement Forum



(2)
Measurement error expression
The error made between the measured primary
current and the theoretical primary current is
defined by :
) (
) ( ) (
) (
p i
p i p i
p
p
p pmes

= .

(3)
i.e
1 =

=
N N
N
K
p K
K
K p K
p
) ( ) (
) (

(4)
and
p L R
p L R p L R
p
m f
m f
.
) ' ' )( (
) (
+ +
= .

(5)
The ratio and angle phase errors are respectively
the real and the imaginary parts of (6):

100
2

+
+
+
=
f
s CT
m
s
K
R
R R
L
L l ) ( ) (
(%) .

(6)

+

+
=

m
s CT
f
s
L
R R
R
L l ) ( ) (
2
.

(7)

Error characterization of CT
Proposal establish the elationship between tests
results and model parameters ; the secondary
winding resistance R
CT
is measured both at
ambient and hot temperature.
The three other parameters are determined, for the
nominal load, from (7) and the phase deduced
from (3) which can be written as:


(8)
With:
' R
L
b
m
=
1
,
f
m m
R
L
R
L
R
L
a + + =
'
'
'
1
,
'
'
R R
L L
a
f
m
=
2
,
s CT
R R R + = ' and
s
L l L + =
2
' .

Then the phase (en radians) is defined by:

{ }
{ }

) (
) (
tan
p K
p K
1
1
1
.

(9)
where { } ) (p K
1
and { } ) (p K
1
are the imaginary
and the real parts of the transfer function (9), we
obtain :


=
1
2
2
1
a
a
) tan( .

(10)
i.e

[ ] + +

=
' ) ' (
' '
) tan(
R L R L L
L L R R
m f m
m f
2

(11)

.
Finally, the secondary time constant is defined by
CT s
R R +
+ +
=
2
l L L
T
s m
s
.

(12)
The system (7), (12) and (13) can be solved and
leads to parameters L
m
, l
2
and R
f
for the nominal
load.

Parameter estimation using results of
calibration tests
From data provided by CT Analyzer, fundamental
parameters can be derived. The load impedance
defined by Rs and Ls can be found directly from
tests at nominal load. The secondary resistance
(R
CT
) is directly measured. Ratio and angle phase
errors are measured for a secondary current from
1% to 200% of nominal. If the secondary leakage
inductance (l
2
) is measured and using (6) and (11),
we write:

0
2
=
+

+

f
s CT
m
s
K
R
R R
L
L l ) ( ) (
,

(13)
and
[ ] 0
2
1 1
= + + + ' ' ' ) ' ( ) tan( L L R R R L R L L
m f m f m
.
(14)
With
s CT
R R R + = ' ,
s
L l L + =
2
' and Hz 50 2
1
. = .











2
2 1
1 1
1
1
p a p a
p b
K
p K
N + +
=

. ) (
2
1
1
1
p
R R
L L
p
R
L
R
L
R
L
p
R
L
K p i
p i
p K
f
m
f
m m
m
N p
s
'
'
)
'
'
'
(
'
.
) (
) (
) (
+ + + +
= =


Presentation 8.3

OMICRON electronics GmbH 2011 Instrument Transformer Measurement Forum
This system can be solved and the parameters L
m

and R
f
are calculated for all current values. Figures
3 and 4 show Rf and Lm estimations calculated
from calibration tests. Other parameters are
derived from direct measurements (Rs=0.48,
L
s
=1.146mH, R
CT
=0.588 and l
2
=0.983mH).


Fig. 3 Estimation of losses resistance Rf


Fig. 4 Estimation of magnetizing inductance L
m

Errors compensation
Single point compensation
Ratio and angle phase errors can be compensated
for the fundamental frequency

=
2
1
1
F , for this
frequency, the transfer function ) (p K
1
is :
2
1 1
1
1
1
1
1
1
1
+ + +

'
'
)
'
'
'
(
'
.
) (
) (
) (
R R
L L
j
R
L
R
L
R
L
j
R
L
K j i
j i
j K
f
m
f
m m
m
N p
s .
We can define the gain and the phase provided by
the CT at the frequency F
1
:
) (
1
1 1
=

j K K
i
,

(15)
[ ] ) ( arg
1
1
=

j K K
i
.

(16)

The compensation is performed by introducing the
gain
) (
1
1
1

j K
K
i
,
And time delay
360
02 0
1
.
) (

=
i
K
T .
In presence of harmonics, the single point
compensation is not optimal. The filter ) (p K
1

provides different gain and phase for each
harmonic component. Indeed, considering linearity
hypothesis and constant values for parameters L
m

and R
f
, each component is applied to the single
filter ) (p K
1
. Applying the single point
compensation in presence of harmonic currents
lead to a variable ratio error and a phase error
larger than without compensation.
Multi point compensation
The principle of single point compensation can be
extended to each harmonic component using the
transfer function K(p):

Fig. 5 Principle of multi point compensation

The transfer function to be simulated can be
expressed as
p b
p a p a
K p K
N
1
2
2 1
1 + +
= ) ( .

(17)
With: )
'
'
'
(
f
m m
R
L
R
L
R
L
a + + =
1
,
'
'
R R
L L
a
f
m
=
2
and
' R
L
b
m
=
2
.
The transfer function (15) is not proper. In order to
simulate K(p), the order of the numerator must be
artificially increased with a filtering pole:
) (
) (
p p b
p a p a
K p K
F
N i nc

+
+ +
=
1
1
1
1
2
2 1


(18)





) (p K
1
i
p
(p) i
s
(p)
i
pmes
(p)
) (p K
Presentation 8.4

OMICRON electronics GmbH 2011 Instrument Transformer Measurement Forum
Simulation results
The multi point compensation is simulated using
Matlab/Simulink with continuous states and
variable step. The filtering pulsation is equal to

F=
10
7
rd/.

Comp.
F
1
=50Hz F
5
=250Hz F
7
=350Hz

single
(%)
K

0.000243 0.00244 0.00341

()
9.10
-5
0.215 0.31

multi
(%)
K

0.00145 0.0339 0.0201

()
0.00182 0.0092 0.0126

Comp.
F
11
=550Hz F
13
=650Hz F
17
=650
Hz

single
(%)
K

0.01252 0.0151 0.0211

()
0.487 0.575 0.755

multi
(%)
K

0.0588 0.0327 0.0537

()
0.0194 0.0236 0.0305
Tab.1 Simulation results

The table 1 shows the comparison between the
single and the multi point compensation in
presence of harmonic currents. These results
prove that the phase error is reduced until ratio
error is held small.
Conclusion
The multi point compensation permit to take into
account the harmonic component and leads to
better results that the single point method. In one
hand, an increased accuracy could be obtained
using all data available in calibration report
provided by the CT Analyzer. In other hand,
parameters as sampling frequency or filter design
are decisive in order to reach desired
performances.
Literature
[1] IEC 60044-1, Instrument transformers Part
1:Current transformers
[2] IEC 60044-6:1992, Instrument transformers.
Part 6: Requirements for protective current
transformers for transient performance
[3] Elleuch M., Poloujadoff M., Analytical Model
of Iron Losses in Power Transformers , IEEE
Transactions On Magnetics, Vol. 39 (2),
March 2003.
[4] Tellinen J., A simple Scalar Model for
Magnetic Hysteresis, IEEE Transactions On
Magnetics, Vol. 34 (4), July 1998.
[5] Emanuel A. E., Orr J. A., Current Harmonics
Measurement by Means of Current
Transformer, IEEE Transactions On Power
Delivery, Vol. 22 (3), July 2007.
[6] Locci N., Muscas C., A Digital Compensation
Method for Improving Current Transformer
Accuracy, IEEE Transactions On Power
Delivery, Vol. 15 (4), July 2000.
[7] Annakkage U. D., McLaren P. G., Dirks E.,
Jayasinghe R. P., ParkerA A. D., Current
Transformer Model Based on the Jiles
Atherton Theory of Ferromagnetic Hysteresis,
IEEE Transactions On Power Delivery, Vol.
15 (1), 2000.
[8] Sss F., Krger M., Kaufmann F., Method for
testing a transformer and corresponding test
device, United States Patent No US
6,987,390 B2, 17 January 2006.
About the Authors
Erik Etien (November 16 1970)
received the Ph.D. degree from
the University of Poitiers
(France) in Automatic in 1999.
His research interests the de-
velopment of soft-sensors in the
field of electrical engineering.



Jean Franois Tissier (France
October 17 1960), professional
path: 1984: Graduate Engineer
of ESE (Ecole Suprieure
d'Electricit) for industrial el-
ectronics; 1999:Schlumberger:
R&D Load Management Engin-
eering Manager; 2011 Itron: Uni-
versity liaison, member of SEE.

Vous aimerez peut-être aussi