Beruflich Dokumente
Kultur Dokumente
JULY/AUGUST 1985
Understanding Asymmetry
CRAIG N. HARTMAN, MEMBER, IEEE
INTRODUCTION
WHENEVER a short-circuit occurs on a conductor there
exists a transient response time during which the circuit Time (Cycles)
attempts to reach a steady-state fault condition. In ac circuits (a)
this transient response may produce currents in excess of those Magnitude (P.U.)
2.0
encountered either before the fault occurred or after the -
DC Componn Asymmetria
steady-state condition is achieved. These high currents must be
taken into consideration when specifying power system
equipment.
It is instructive to note the effect which high efficiency
equipment is having on this type of analysis. Generally
transformers constitute the bulk of the systems impedance, and
thus have a predominate effect on the system X/R ratio. A
cursory look at this situation would indicate that increasing the 00 0 Q1 0.2 0.3 0 4 0.5 0.6 0.7 0 8 0.9 1 0
Time (Cycles)
efficiency of a transformer from 98-percent to 99-percent (b)
would have the effect of doubling the X/R ratio. This effect is Fig. 1. Fully offset wave.
seen in virtually all high efficiency equipment. The X/R ratio
of modem distribution systems commonly exceeds the estab- 3) the fault impedance is purely inductive;
lished X/R ratio of the protective equipment, thus requiring 4) no current was flowing prior to fault inception.
derating. Referring to Fig. 1(a) we note that at time zero both current
The author feels that it is important not only to take these and voltage are zero. Recall that the rate of change of current
factors into account when sizing the equipment, but also to in an inductor is proportional to the voltage across that device.
have some visual conception of what actually takes place At time 0.05 cycles the voltage is beginning to rise but is still
during those first few cycles following inception of a fault. It is rather low. The current in the inductor is also beginning to
felt that a good visualization of the situation would clarify a rise, but the slope of current change is small in response to the
number of misconceptions which exist relative to this subject. low voltage value. As the voltage increases, the rate of change
Accordingly this article shall begin by "walking through" a (slope) of the current wave increases in response. At time 0.25
hypothetical fault. cycles the voltage reaches a maximnum, and the slope of
FAULT VISUALIZATION current is at its steepest value. NoW the voltage begins to drop,
but since it is still positive the current continues to increase. At
A worst-case condition is assumed in which time 0.45 cycles the current is still increasing, but due to the
1) the fault occurs at an instant of time when the source low voltage level the current increases at a very slow rate. At
voltage is equal to zero; time 0.5 cycles the current finally reaches its maximum value.
2) the source voltage can supply an infinite amount of Not until the voltage goes negative can the current begin to
current with no voltage dip; decrease. Using the same arguments one can follow the course
Paper IPSD 84-38, approved by the Power Semiconductor Committee of of the current wave back to zero magnitude at time 1.0
cycles.
the IEEE Industry Applications Society for presentation at the 1984 Industrial Using the same reasoning above, but with the fault
and Commercial Power Systems Conference, Milwaukee, WI, May 9-12, occurring at a voltage maximum, one could trace a current
1984. Manuscript released for publication November 2, 1984. (no
The author is with the Westinghouse Electric Corporation, 3900 S. wave which is symmetrical with regard to current zero
Wadsworth Blvd., Lakewood, CO 80235. offset). By looking at the two waves it bbecomes obvious that
Authorized licensed use limited to: CHILECTRA. Downloaded on May 03,2010 at 22:48:37 UTC from IEEE Xplore. Restrictions apply.
HARTMAN: UNDERSTANDING ASYMMETRY 843
the fully offset current wave is simply obtained by adding a dc zero degrees so that in the purely resistive circuit the voltage
current equal to the peak magnitude of the symmetrical wave and current will be completely in-phase. Also note that at time
to the wave itself (Fig. 1(b)). The actual value of the dc zero the dc component is exactly equal in magnitude to the
component would depend on the time at which the short-circuit value of the ac component but opposite in sign. This condition
occurs in the voltage wave and is quantitatively equal and must exist due to the fact that current cannot change
opposite to the value of the steady-state symmetrical current instantaneously in an inductive circuit. The argument of the
wave at time zero. When the dc current assumes a value equal exponential indicates that the dc component will decay at a rate
to the peak value of the symmetrical current, we say that the dependent on the X/R ratio.
wave is fully offset (maximum asymmetry). It is possible
under certain circumstances to get more than 100-percent MAXIMUM ASYMMETRY VERSUS MAXIMUM PEAK
offset, but these special circumstances are beyond the scope of CURRENT
this paper. It is Interesting note that the maximum asymmetrical peak
to
currents do not generally occur during conditions of maxi-
QUANTIFYING THE CURRENT VALUES mum asymmetry. It has been shown [I] that the maximum
In real life, of course, perfect inductors do not exist any peak asymmetrical current will be produced for any X/R ratio
more than do perfect voltage sources. In order to obtain more when the fault occurs at voltage zero (O = 0). As shown in
realistic actual values the hypothetical fault circuit discussed (I), however, maximum asymmetry exists when the fault
above will be modified to include a resistance. This circuit is occurs at a symmetrical current component maximum (0 =
now as shown in Fig. 2. arctan (X/R) - 90). (Maximum asymmetry is defined as
If the switch is closed at time zero to simulate initiation of being that state which produces the maximum dc component
the fault and if current values are per-unitized according to the value.) Thus the fault angle for maximum peak current is
peak symmetrical current (IO) then the current plot is a always zero whereas, the fault angle for maximum asymmetry
function of three independent variables (0, X/R, and time). ranges from 0-90 degrees.
Equation (1) results from using Kirchhoffs voltage law to sum Table I shows the maximum actual peak current for several
the voltages around the circuit of Fig. 2 and then solving the X/R ratios, along with the dc component, ac component, and
resulting nonhomogeneous first order differential equation for the time in cycles at which the current peaks. Table II
current.
TABLE I
I= IO{sin [arctan (X/R) -
MAXIMUM FAULT CURRENTS
exp [- 27rt/(X/R)1 (dc component)
X/R I Ii I. I Time
+ sin [2rxt arctan (X/R)I} (ac component) (1)
-
Authorized licensed use limited to: CHILECTRA. Downloaded on May 03,2010 at 22:48:37 UTC from IEEE Xplore. Restrictions apply.
844 IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS, VOL. IA-21. NO. 4. JULY/AUGUST 1985
TABLE II
DIFFERENCES IN CALCULATED FAULT LEVELS ACCORDING TO METHOD OF CALCULAFION
.I
Magnitude (P.U.)
compares the actual maximum peak current to the peak current 20
under conditions of maximum asymmetry. 1.8
1.6 -
DC Components
g _Actual Current
Since finding maximum current magnitudes using (1) is 14 -
1.2-
rather unwieldy, the traditional method for calculating cur- 1.0
rents has been to assume a fault of maximum asymmetry and 0.8p A tu l
XI
then calculate the peak current at 0.5 cycles. Substituting these -02
0.2
assumptions into (1) yields the following: 04
- 0 6
-,i 11olI I I I 7 I AI I I I X I I
- 0.8
Authorized licensed use limited to: CHILECTRA. Downloaded on May 03,2010 at 22:48:37 UTC from IEEE Xplore. Restrictions apply.
HARTMAN: UNDERSTANDING ASYMMETRY 845
2.6
compromise, generally containing both conservative and 2 = Peak value (in P.U. of sym. Peak)
=
Authorized licensed use limited to: CHILECTRA. Downloaded on May 03,2010 at 22:48:37 UTC from IEEE Xplore. Restrictions apply.
846 IEEE TRANSACTIONS ON INDUJSTRY APPLICATIONS, VOL. IA-21 NO. 4, JULY/AUGUST 1985
TABLE I1I While (6) gives the estimated capability factor, it should be
CIRCUIT BREAKER OPERATION TIMES understood that for each circuit breaker a certain capability
factor is mandated by standards and that circuit breakers are
Rated Contact tested to those stated values. These values were noted for each
Interrupting Opening Parting
Time Time Time Capability breaker in Table III.
(cycles) (cycles) (cycles) Factor It would appear from the above that a circuit breaker with a
large capability factor could be applied on systems with low
2 1.0 1.5 1.3 X/R ratios at a symmetrical current higher than its rating.
3 1.5 2.0 1.2
5 2.5 3.0 1.1 Current interruption, however, is a function of the transient
8 3.5 4.0 1.0 recovery voltage (TRV) as well as the thermal energy in the
arc. TRV is a function of di/dt which relates to the
symmetrical component of the current, so the symmetrical
In ac high-voltage circuit breakers, the predetermined X/R value of the fault current is just as important as the
ratio value has been established at 15. Thus, all ac high- asymmetrical value. Circuit breakers should therefore never
voltage breakers must be able to withstand the 1.52 per-unit be applied where the symmetrical fault current exceeds their
rms current and 1.81 per-unit peak current (2.55 times the symmetrical rating, no matter what the X/R ratio.
symmetrical rms value) as shown in Fig. 4 at 0.5 cycles. As For systems with larger X/R ratios, however, the capability
stated previously they are actually rated at 1.6 per-unit rms or factor can be put to good use. Equation (7) may be used to
2.7 per-unit peak.) determine a multiplying factor for X/R ratios higher than 15
The interrupting rating is somewhat more difficult to assess by which the symmetrical fault current must be multiplied.
since different circuit breakers will interrupt the fault at (The reciprocal of (7) could also be used as a derating factor
different times. As a result these circuit breakers are classified for the breaker.)
according to their rated interrupting times. Table III shows V I + 2 exp [ - 4xt/(X/R)]
circuit breaker operating times according to ANSI/IEEE multiplying factor (7)
C37.04-1979 [7]. S
The rated interrupting time is the maximum permissible where
interval between the energizing of the trip circuit at rated
control voltage and the interruption of the main circuit on all t contact parting time in cycles
poles. A breaker should be rated to interrupt the amount of X/R actual circuit X/R
current present at the contact parting time. Contact parting S capability factor.
time is determined by adding 1/2 cycle relay time to the A plot of (7) will correspond to the graphs shown in ANSI/
breaker opening time. The curves of Fig. 4 may be drawn for IEEE C37.010-1979 [10, fig. 10, p. 36]. The equation is not
any X/R ratio using (3), (4), and (5). meant to imply greater accuracy than the curve it represents,
but only greater convenience should the user wish to use some
dc =
exp [ 27rt/(X/R)] t in cycles sort of computerized selection procedure for large distribution
peak current dc + 1
= (4) systems.
So far we have discussed derating only with regard to the
rms current- l2dc2+l . (5) rms values. Should we not have equal concern for the peak
A five-cycle breaker with a standard three-cycle contact current values? Table IV shows the peak values of current
parting time will be used as an example. As seen in Fig. 4 the along with the multiplying factors for rms current from (7).
rms current which the breaker must interrupt is 1.08 times the
Note that under certain conditions the peak ratio multiplying
symmetrical value. This value occurs at the contact parting factor exceeds that of the rms by about 2 percent. Since the
time of 3.0 cycles. The ability of a circuit breaker to interrupt actual momentary tested value of 1.6 discussed earlier exceeds
a higher rms value of asymmetrical current than its symmetri-
the calculated value of 1.52 by 5 percent it would appear that
cal value is called "capability factor." The breaker just the multiplying factors provided by (7) would provide the
discussed would need a capability factor of about 1.08. The necessary protection for peak currents as well as rms.
actual capability factor of this breaker is 1.1. Therefore a five- According to ANSI/IEEE C37.010-1979 [10], if the fault is
cycle circuit breaker with a contact parting time of three cycles fed predominantly from generators through a) not more than
one transformation or b) a per-unit reactance external to the
can interrupt 10 percent more asymmetrical rms current that
its rated symmetrical value. The approximate capability factor generator which is less than 1.5 times the generator per-unit
for any circuit breaker may be obtained from (6) subtransient reactance on a common system megavolt-ampere
base, then one should use Figs. 8 and 9 from [10] for derating.
S = 1+2 exp[-4irt/(X/R)] (6) These figures include the effect of ac decrement and have been
where determined by empirical methods.
The IEEE working group has determined that the standards
S capability factor are slightly "pessimistic" as far as handling the dc component
t contact parting time in cycles is concerned, but "optimistic" as far as handling the ac
X/R 15. component is concerned. Combined, the results are quite
Authorized licensed use limited to: CHILECTRA. Downloaded on May 03,2010 at 22:48:37 UTC from IEEE Xplore. Restrictions apply.
HARTMAN: UNDERSTANDING ASYMMETRY 847
TABLE IV TABLE V
COMPARISON OF DERATING HIGH-VOLTAGE BREAKERS BY PEAK POWER FACTOR OF TEST CIRCUIT FOR MOLDED CASE BREAKERS
VERSUS DERATING BY RMS
Rated Interrupting Current
Peak -rms (by maximum asymmetry)- Amperes Power Factor
X/R ratio Two cycle Three cycle Five cycle Eight cycle
10000 and less 0.45-0.50
15 1 0.964 0.977 0.980 1.034 10001 to 2000 0.25-0.30
16 1.006 0.978 0.992 0.992 1.042 Above 20000 0.15-0.20
17 1.011 0.991 1.006 1.003 1.051
18 1.016 1.004 1.019 1.015 1.060
19 1.020 1.015 1.032 1.027 1.069 TABLE VI
20 1.024 1.026 1.044 1.038 1.078 CAPABILITIES OF MOLDED CASE CIRCUIT BREAKERS
21 1.028 1.036 1.056 1.049 1.087
22 1.031 1.046 1.067 1.060 1.097 Rated
23 1.034 1.055 1.077 1.071 1.107 Interrupting
24 1.037 1.064 1.087 1.082 1.116 Current iaym lsym
25 1.039 1.072 1.097 1.092 1.126 (amperes) X/R (rms) (peak)
26 1.041 1.079 1.106 1.102 1.135
27 1.044 1.087 1.114 1.112 1.145
28 1.046 1.093 1.123 1.121 1.154 10000 and less 1.7-2.0 1.03-1.04 1.16-1.21
29 1.048 1.100 1.131 1.130 1.163 10001 and 20000 3.2-3.9 1.13-1.18 1.37-1.44
30 1.049 1.106 1.138 1.139 1.172 Above 20000 4.9-6.6 1.25-1.33 1.53-1.62
Peak ratio: peak current in per-unit of peak current with X/R = 15. Iaym (rms) is in per-unit of symmetrical rms.
rms (by maximum asymmetry): derating factor by (7) [6-10] for each I8,y. (peak) is in per-unit of symmetrical peak.
breaker interrupting time.
Using (3), (4), and (5) as before we can calculate the
close. It should be stated, also, that long experience with asymmetrical currents shown in Table VI. For values of Iasym
applying circuit breakers by these methods has seemed to in per-unit of symmetrical rms current, simply multiply by the
verify their adequacy. square root of 2. These values are usually taken to be 1.7, 2.0,
and 2.2, respectively, for the three breaker ratings. For system
LOW-VOLTAGE POWER CIRCUIT BREAKERS X/R ratios higher than those in Table VI, the multiplying
Low-voltage power circuit breakers are assumed to have a factor to use for modifying the calculated fault values is
arrived at using (9)
contact parting time of 1/2 cycle unless otherwise stated. This
means that their momentary and interrupting ratings should be 1 + exp [ - 7r/(X/R)]
identical. An X/R ratio of 6.6 is also used rather than the 15 multiplying factor =
used for high-voltage breakers. Finally considering the charac- Iasym (peak)
teristics of low-voltage breakers (both power and molded case) (see Table VI). (9)
it has been decided that peak currents rather than rms currents
will determine the interrupting capability. From (3) and (4), OTHER METHODS OF BREAKER RATINGS
using a time of 0.5 cycles and an X/R ratio of 6.6, it can be
seen that these breakers must interrupt a peak current of 1.62 Occasionally circuit breakers (especially older models) may
times that of the symmetrical peak (2.3 times that of the be rated according to asymmetrical single-phase amperes or
symmetrical rms). For X/R ratios higher than 6.6 one may asymmetrical three-phase amperes. In the asymmetrical sin-
use (8) to modify the fault current prior to selecting a breaker gle-phase case one may use (3) and (5) to find the nns current
and match this to the interrupting rating. In the asymmetrical
1 + exp [-rl(XIR)] three-phase case, (10) should be used and matched to the
multiplying factor (8) interrupting rating. A convenient table to use for this purpose
1.62
is included in NEMA standard publication No. AB1-1975 [4,
Equation 8 would correspond to table 3, ANSI/IEEE C37. 13- pt. 2, p. 121.
1981. Again, (8) should be not used for X/R ratios lower than
6.6. Iave (asym)- {{Il + 2 exp [- 4it/(X/R)
MOLDED CASE CIRCUIT BREAKERS
Due to the lack of standards, the following procedures are + 2 V11+ 1/2 exp [-47rt/(X/R)] } (10)
suggested for applying molded case circuit breakers. t in cyles.
Unless otherwise stated, molded case circuit breakers are
assumed to have contact parting times of 1/2 cycle. The rated APPLICATIONS TO OTHER DEVICES
X/R ratio, however, depends on the size of the breaker. Table Fuses will not be covered in this paper, but the point should
V, taken from NEMA publication No. ABI-1975 141, shows be made that (3), (4), and (5) of this paper are much preferred
the power factor of the test circuit under which each of these to the "up-over-down" method used on most fuse curves, the
breakers must be tested. "up-over-down" method assumes an X/R for you.
Authorized licensed use limited to: CHILECTRA. Downloaded on May 03,2010 at 22:48:37 UTC from IEEE Xplore. Restrictions apply.
848 IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS, VOL. IA-21, NO. 4, JULY/AUGUST 1985
Authorized licensed use limited to: CHILECTRA. Downloaded on May 03,2010 at 22:48:37 UTC from IEEE Xplore. Restrictions apply.