0% found this document useful (0 votes)
146 views

Enhancing Differential Protection Stability During CT Saturation With Transient Bias

Uploaded by

gilbertomjc
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
146 views

Enhancing Differential Protection Stability During CT Saturation With Transient Bias

Uploaded by

gilbertomjc
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 4

ENHANCING DIFFERENTIAL PROTECTION STABILITY

DURING CT SATURATION WITH TRANSIENT BIAS


O. Bagleybter*, S. Subramanian*
*Alstom Grid UK Ltd, UK, [email protected]

Keywords: Current differential protection, transient biasing, The condition (1) essentially represents the equation 18 from
CT requirements, CT modelling. [2] when the inductive burden can be neglected.

Abstract In most cases, protective relays are connected to non-air-


gapped CTs of IEC P, TPS, TPX class or ANSI С, K class.
The paper describes a Transient Bias technique designed to According to [2], remanent flux as high as 80% of saturation
overcome the effects of Current Transformer (CT) saturation, flux can be observed for these CT classes. It should be also
which might severely affect the behaviour of a protection noted that once the remanent flux is established, it is
system. This Transient Biasing, as the name implies, is only dissipated very little under service conditions. To reduce the
active during transient conditions in the power system, and it remanence to less than 10% of saturation flux density a
decays quickly once the transients disappear. A voltage of about 60% of the knee-point voltage must be
complementary algorithm for fast detection of external faults applied [2], which is extremely unlikely under service
is proposed, which prevents the Transient Bias feature from conditions.
increasing the differential protection operating time in case of
internal faults. The operation of current differential protection If we assume X/R equal to 20 (quite a common value for
with Transient Bias is checked using Simulink® models of transmission systems), then Equation (1) transforms into the
the relay and current transformer. In addition, the paper following equation:
reviews the results of RTDS® tests carried out with actual I ⋅ R ⋅ 21
current differential relays and provides detailed information Vx > s ∑ = 105 ⋅ I s ⋅ R∑ .
1 − 0.8
on CT requirements testing method adopted by Alstom Grid.
It can be seen that taking into account the DC component and
the remanence leads to rather high CT requirements.
1 Introduction Compared with the simple calculation for a symmetrical AC
current and zero remanence, the equation (1) defines an
Even while such an advanced technology as “Non
overdimensioning factor of 105. Moreover, in modern EHV
Conventional Instrument Transformers” is proliferating fast,
systems the X/R ratio as high as 65 can be observed, which
the majority of protection devices are still connected to
conventional electromagnetic current transformers (CT).
results in the overdimensioning factor of 330.
Apparently, this situation is not going to change for at least
Therefore, manufacturers of numerical protection relays have
another 5 to 10 more years. The reliability of any protection,
a difficult task in defining CT requirements for their devices.
and especially differential, is largely dependant on the correct
CT dimensioning. Theoretically, it is possible to dimension a
First of all these requirements must guarantee secure and
CT to completely prevent its saturation [2]. To achieve this it
dependable relay operation in all cases. On the other hand the
requirements should be reasonably low, otherwise the choice
is necessary to comply with the condition:
of CT could become impossible or lead to unnecessary extra
costs. There are different techniques of reducing CT
I s ⋅ R∑ ⋅ (1 + X R )
Vx > (1) requirements, such as saturation detectors and waveform
1 − REM p.u. recovery algorithms [4]. The situation with CT dimensioning
where is aggravated by the fact that there are no documents defining
Vx is the saturation voltage, a common approach to CT dimensioning tests for
manufacturers of Current Differential protection. An attempt
I s is the maximum expected fault current (the primary
to develop such a standard was made in report [1], though this
divided by the turns ratio), document still has a “draft” status.
R∑ is the total secondary burden, with the internal CT
resistance and the neutral lead burden.
X R is the X/R ratio of the primary system,
REM p.u . is the maximum possible remanence of the CT.
I DIFF
2 Current Differential Protection with
Transient Bias
K2
The initial idea of transient biasing was for any positive R > K2
change (delta) of the bias current I BIAS to increase the [I BIAS (n), I DIFF (n)]
tripping threshold I DIFF .THR. , and this additional biasing had
to be accumulating. If, however, the bias current was not
R < K2
increasing, the transient bias value ITR. BIAS had to decay K1
I S1 [I BIAS (n), I DIFF (n)]
exponentially.
[I BIAS (n − 1), I DIFF (n − 1)] IS 2
I BIAS
Unfortunately, testing this algorithm with models
immediately showed that it had a drawback, since the tripping
time increased for internal faults. The reason was that an Figure 1: The bias characteristic and the external fault
internal fault also led to the positive change of the bias detection principle.
current, and the protection was effectively desensitized by the
additional bias. To obtain the transient bias I TR.BIAS , deltas of differential and
bias currents are calculated, and also the ratio of these deltas
To avoid slowing the protection down an external fault R:
detector was implemented. This detector is based on the ratio ΔI DIFF (n) = I DIFF ( n) − I DIFF (n − 1) , (3)
of the differential to the bias current deltas [5].
ΔI BIAS (n) = I BIAS (n) − I BIAS (n − 1) , (4)
With this detector the positive delta of the bias current is only R = ΔI DIFF (n) ΔI BIAS (n) . (5)
used to increase the additional biasing if this delta is caused
by an external fault or a sudden surge of the load current. The transient bias is then calculated with the following
algorithm:
The transient bias stops increasing during CT saturation
because differential current occurs, but some value of If R < K i and ΔI BIAS (n) > 0 (6)
I TR.BIAS must be already built up which provides sufficient
I TR.BIAS (n) = D ⋅ I TR.BIAS (n − 1) + S ⋅ ΔI BIAS (n) , (7)
stability until the next undistorted segment of the current
waveform.
otherwise
I TR.BIAS (n) = D ⋅ I TR.BIAS (n − 1) , (8)
2.1 Transient biasing algorithm
For a sample n in the current differential algorithm the where
following equation is applied given the additional transient K i is the slope percentage ( K1 or K 2 depending on where
bias ITR. BIAS : the point [I BIAS (n), I DIFF (n)] falls on the bias characteristic),
S is the scaling coefficient,
I DIFF .THR. (n) = SLOPE (I BIAS (n) ) + I TR.BIAS (n) , (2) D is the decay coefficient, D < 1.
where
I DIFF .THR. (n) is the threshold of the differential protection, The differential algorithm in the line differential relays
SLOPE (I BIAS (n) ) is the bias characteristic with two slopes produced by ALSTOM Grid has a sampling frequency of 8
samples per cycle or 400 Hz (providing that the system
used in line differential relays manufactured by ALSTOM
frequency is 50 Hz). Therefore, after one cycle (20 ms) the
Grid (Figure 1).
transient bias will decay to D8 of its initial value (if the
condition (6) was not fulfilled during this time). In this paper
the coefficient D = 0.8 is adopted, then after 20 ms
D 8 = 0.1678 or about 17% of initial value. Such a fast decay
helps to avoid delaying the trip in the case of a cross-country
fault.

2.2 Testing of the transient biasing algorithm with


SIMULINK models
While developing and testing the Transient Bias feature, a
SIMULINK® model of CT was used extensively. This model
is based on the Jiles-Atherton theory of hysteresis and
constitutes a replica of a well-known CT implementation in The test stops once one of the relays trips, the RTDS system
the RTDS® and PSCAD® systems [3]. However, the records a limiting burden Rlim equal to the burden one step
SIMULINK® model is more flexible in use. before the trip occurred.

The CT model employed in the paper has the following For internal faults the criteria to stop the test is non-operation
parameters: the CT ratio KCT = 10/1, the rated secondary of the protection or unacceptable tripping time.
current 1А, the knee-point voltage VK = 160V.
It is essential to test the protection with different CT
Figure 2 illustrates the CT saturation, the spurious operation remanence values. To achieve that, the initial remanence is
of the differential function without the transient bias and the set to zero and every fault is applied six times: three times
stabilizing effect of the transient bias. with a positive DC component (the remanence grows from
zero to its positive limit) and three times with a negative DC
60
40 component (the remanence drops from the positive limit to its
20
0 negative limit).
-20
a) The CT saturation
20 I DIFF .THR. Trip!!! After completing all the tests the RTDS system generates an
10
Trip!!! array of limiting CT burdens Rlim for external and internal
I DIFF
0 faults with different combinations of the following
b) The differential current and the tripping threshold without parameters:
the transient bias
30
20 I TR .BIAS 1. Knee-point voltage of the CT Vk : 40-320 V
10
0
2. Fault current If : 2-50 А secondary
c) The transient bias current 3. X/R ratio of the network: 5-120
40 I DIFF .THR. 4. Fault type: A-N, B-C, A-B-C
20
0
I DIFF 5. Point on wave (POW): 0°-150°
0.48 0.49 0.5 0.51 0.52 0.53 0.54 0.55 0.56 0.57 0.58 0.59 0.6
d) The differential current and the tripping threshold with the For the second slope setting K2 = 150% the number of tests
transient bias
totalled 3967 for external faults and 1915 – for internal,
where a single test comprises of the six-shot fault sequence
Figure 2: The stabilizing effect of transient biasing. described above.
Once the tests in the Simulink® environment had
2.4 Processing the results of the RTDS testing
demonstrated the merits of the transient bias algorithm, this
feature was implemented in the software of actual line The data produced by the RTDS system is processed in the
differential relays. MATLAB environment.

2.3 Testing relays with RTDS system and determining It should be noted that the limiting burden for internal faults
new CT requirements is much higher than for external ones owing to the external
fault detector and the fact that the transient biasing does not
The final tests of the differential relays with the transient slow the relay down during internal faults.
biasing feature, and the determining of the new CT
requirements, were carried out with the RTDS (Real Time A coefficient K is calculated from the limiting burden Rlim .
Digital Simulation) system. The set-up involved the This coefficient can be used as a dimensioning factor for an
following: arbitrary CT.
1. A protected line with settable X/R ratio and fault current
VK
(for internal and external faults). K= , (9)
2. Two models of three-phase CT with settable knee-point I n ⋅ (2 ⋅ Rlim + 0.5)
voltage Vk and burden. where
3. Several line differential relays connected to the CT models VK is the knee-point voltage of the CT used in the RTDS
through digital-to-analogue converters and amplifiers. test,
I n is the rated secondary current of the CT used in the RTDS
The stability check for external faults is done by applying
multiple faults with a pre-defined fault current and X/R ratio test (1А),
while incrementing the CT burden until one of the relays (2 ⋅ Rlim + 0.5) - the total CT burden including both forward
trips. The burden RL (one-way) is increased step by step from and return leads and the internal CT resistance (fixed at 0.5
0.1 Ohm to 16 Ohms (38 steps in total). Ohm).

In real applications the maximum fault current If and X/R ratio


of the network are defined by calculation, while the fault type
and POW are arbitrary factors for any fault. Therefore, the tests and benefits of using the Transient Bias feature are
highest K value for different fault types and POW (the worst shown as well.
case) must be chosen for each combination of the fault current
and X/R ratio. The result of this procedure was a matrix of 56 References
values of K for every tested [If , X/R] combination.
[1] “Coordination of Relays and Conventional Current
After applying a two-dimensional fitting procedure (from the Transformers”, CIGRE Report, CIGRE_B5.02 Draft
MATLAB Optimization Toolbox) the following equation was 10a, (August 2005).
derived: [2] “IEEE Guide for the Application of Current
Transformers Used for Protective Relaying Purposes”,
K = (1.42 ⋅ I f + 53.7) ⋅ (6.06 ⋅ 10 −3 ⋅ X R + 0.515) . (10) IEEE Std C37.110-2007.
[3] Annakkage U. D., McLaren P. G., Dirks E., Jayasinghe
R. P., Parker A. D. “A current transformer model based
Finally, the required knee-point voltage Vk can be calculated on the Jiles-Atherton theory of ferromagnetic
as follows: hysteresis”, IEEE Transaction on Power Delivery,
volume 15, No.1, pp. 57-61 (2000).
Vk = K ⋅ I nom ⋅ RΣ , (11) [4] Kang Y.C., Yun J.S., Lee B.E., Kang S.H., Jang S.I.,
where Kim Y.G. “Busbar differential protection in conjunction
I nom is the rated secondary CT current, with a current transformer compensating algorithm”,
R∑ is the total secondary burden, with the internal CT IET Gener. Transm. Distrib., volume 2, No.1 pp. 100-
109, (2008).
resistance and the neutral lead burden. [5] Pradeep K.G., Tarlochan S.S. “Current Differential
Protection Relays”, US Patent Application Publication
2.5 Comparison with existing CT requirements No.: US 2009/0009181 A1, (2008).
The existing CT requirements for ALSTOM line differential
relays are defined by the following equations:

(
If ( I f ⋅ X R) ≤ 1000 : K = max 65, (40 + 0.07 ⋅ ( I f ⋅ X R )) ,)
If 1000 < ( I f ⋅ X R) ≤ 1600 : K = 107 .

Table 1 shows the comparison between existing and new CT


requirements for several combinations of the fault current and
X/R ratio.

If X/R K K Effect
w/out trans. bias with trans. bias %
5 5 65 33.2 49
10 10 65 39.1 40
20 20 68 52.2 23
30 30 103 67.1 35
40 40 107 83.7 22
40 65 - 100.4 -

Table 1: Comparison between existing and new CT


requirements for line differential relays.

3 Conclusion
The transient biasing algorithm is proposed in the paper that
significantly increases the stability of Current Differential
protection during external faults. The algorithm is first tested
with CT and relay models in the MATLAB/SIMULINK®
environment and then with the RTDS® system.
The paper also describes in detail a testing technique used by
ALSTOM Grid to define CT requirements of Current
Differential protection. The results of the CT requirements

You might also like