ransrormer TL

A
s
Magnetizing .L
.
nrus
E I
urrent rthe, coil and the core. The equation for
thecircuitisthen
-WE sin wt =iR+ Ndo
N 10-8 (1)
T. R. SPECHT dt
MEMBER AIEE
The quantity 4) is the flux linked by the
winding excited. Integrating this equa-
THIS paper will develop formulas and tion. The eddy losses in the core are neg- tion from -wr/w to t:
curves for calculating approximately lected. Essentially they represent a re- -VE t
the magnetizing inrush curve of a trans- sistance load in parallel with the trans- (1+ cos wt) iRdt+
former. Formulas have been given by former, and if the system impedance is w
J -r/
others for calculating the peak of the low, they should not materially affect the N(c-fr)10- (2)
first cycle of inrush current,' but no con- magnitude or rate of decay of the inrush Let 0,m be the maximum normal flux due
venient method has been given for de- transient. Experience factors will be to the applied voltage.
termining the shape of the inrush curve. used, which will help to take into account
This information is of interest for a num-
ber of reasons. Where a transformer is
the effect of the iron.
Referring to Figure 2, the voltage is
= -_E1o3
coN
protected by fuses, the fuse must have assumed applied to the transformer at Then from equation 2 and 3,
characteristics such that it will carry the -180 degrees at the voltage zero point.
magnetizing inrush current. Overload Also the residual flux is assumed to be in 1+cos cot)-Rdt
and differential relays must have suffi- the same direction as the flux change due N J,r/w
cient time delay so as not to operate to the applied alternating voltage so that (4)
falsely on magnetiz4ng inrush. With the they will add. This combination will give
availability of a method for determining the maximum inrush current. When the The fu adc
the shape of the inrush curve, it should be flux is below the saturation value, the C,t + s:
possible to precalculate to some degree
the time delay it is necessary to include in
current flowing is negligibly small. See
Figure 3, a typical oscillogram of a mag-
108 f
& 4rNRdtj
dsr/co (5)
any protective device and then determine netizing inrush. The system resistance
how much protection the devices will give and the transformer winding resistance From equation 5 it is seen that the inte-
under steady loads. are combined into one, R, for the purposes grated iR drop over a cycle has the effect
of calculation. Also, the system react- of reducing the residual flux. As long as
Theory ance and the transformer reactance are there is a d-c component in i, this term
combined into one, X. For purposes of will continue to reduce the residual flux
The circuit that will be analyzed is analysis, the transformer magnetic cir- each cycle. The method used in this
shown in Figure 1. A source of alternat- cuit will be divided into two parts, the paper is as follows
ing voltage, E, is applied to a series circuit core and the effective space between the 1. Use a simplified equation 4 for the first
consisting of the system reactance Xs, coil and the core. For purposes of visu- cycle assuming the integrated iR term is
system resistance R,, transformer winding alization, the system reactance is equiv- negligibly small. For most transformers
resistance Rt, and transformer magnetiz- alent to an increase in the space between
0
~~~~~~~~~~~~~~~~~is
true. supplied from normal sources of power this
ing reactance X8. The transformer is as-
sumedtonbe singphae transformer.isas-
sumed to be aa single-phase transformer.mietecrntfrheistyl. 2. Using this simplified equation 4, deter-
mine the current for the first cycle.
Reference 1 describes how to use this data
withca3pscestranformer bano or Xs HT rasfrmrorbak 3. From equation 5, determine the residual
Rs wh a 3-hae
fflux at the start of the second cycle. The
transformers. The winding of the trans- 'evaluation of the integrated iR term for a
former is assumed to have infinite react- cycle will subsequently be explained.
ance up to the saturation point and to E IXTr 4. Use this new value of residual flux in
have a finite reactance X, above satura- , simplified equation 4, and determine the
. . ~~~~~current for the second cycle.
Paper 51-53, recommended by the AIEE Trans- -- J E _ _-_-
_
formers Committee and approved by the AIEE 5. Continue this process to find the corn-
Technical Program Committee for presentation at Fgr .Crutfrtasomr h otg lt nuhcre
N. Y~., January 22-26, 1951. Manuscript sub- applied is E. Xs and Rs are the supply react- Neglecting the last term of equation 4,
imitted Octobering
5,19550; made available for print- aneand resistance. ti h eitneo h hseuto eoe
December
5, 1950.
~tancfrer .idn Rn1
T. R. SPESCHT is with Westinghouse Electric Corpo- trnfmewidgadX1sthstutd
istheeitne oaturthed tiqainbcms
ration, Sharon, Pa. reactance of the transformer 4+='4r+4m(l+cos cot) (6)

195,1, VOLUME 70 Specht-Transformer Magnetizing Inrush Current 323

Authorized licensed use limited to: Universidad Industrial de Santander. Downloaded on September 19,2023 at 17:54:32 UTC from IEEE Xplore. Restrictions apply.
 4T_sLx
_ ,N NlRdt saturation angle for the first cycle is from saturation angle 0 for the n'th cycle is:
, > 0r +0m+0ocoswt equation 13, where OR is the initial re- n-I
Om
20n it | 't' 0 sidual flux: e sm-m-R+R
=s A(0) (22)
-I~
{ O0 5
=cos 01\(k
=- sk-m-<R (15)
( m X
E / 1'1 (APPROX.)n i
Or /I(ACTUAL)R
L_____ 0tAPPROX) Or, cos 0 cos 01+ zA(0) (23)

CCOS 01-BsB Bm BR (16) 1

Bm The maximum current for a particular

Figure 2. Instantaneous voltage, current, and since these fluxes are all in the iron and cycle would be from equation 14,
flux in a transformer for one cycle of magnetiz- are all referred to the same core area.
ing inrush The current flowing during the first cycle iM = \/2E
x
(1 - cos 0) (24)
is,
The
Taneaverage current for the first ii
and will apply for any particular cycle. -

U=s (cos cot - cos 01) (17) cycles will next be derived. For any par-
ticular cycle, for that cycle, from equa-
space between the core and coil. Above
saturation the permeability of the core The flux at the end of the first cycle is tion 14 and 19,
will be assumed to be constant, hence the from equation 5, ffO /w V2E
coil will have an effective inductance L
above saturation. The saturation flux
108
k=O-R N X X/
V\.ER JiJ X
(25)
in the core is k8. Above saturation the
following equation applies,
V\/2E A(0)
( (cos cot-cos 01)dt (18) X c
(26)
d(k di OJ /
N-
dt 10-8=L--ddt (7)
Let
For the first n cycles, the average current
would be,
Integrating this from the instant at which J (=0), , ct-cos 0)dt
(cos
n
fidt
the core saturates: J2(sin 0-0 cos 0) (19) iq
-n V2E 1 (27)
27rn
Figure 8 is a curve of A(0).e()Then,
J d(k L di dt X 2(rn
s dt dt co
N( -Os)10-8=Li (9) R The -nms current for the first n cycles
N10- ( X will next be derived. For any particular
i=L - s) (10) cycle, for that cycle, from equation 14,
The flux at the end of the n'th cycle,
Substituting into equation 10 the value which is the residual flux, is then by con- fi2dt= (COS o- cos o) dt
of 0 from equation 6, tinuing the process, J J _LX
N10- n, (28)
L ((rr+(Rm-ks(+(mcOs Ct) (11) =r R-kmjE (0) (21) (VE)2Ro (29)
Using equation 3, w
Substituting this in equation 13, the
X \ Om
+ cos cot / (12)
R(0) =20+0 cos 20-- sin 20
3 (30)
2
Let 0 equal the angle at which i= 0, that
is, the angle at which saturation occurs: 1CK 9 iscaecurve aforthcrent
Figren cycles R( uor the
Os -Ora -Or _c='- __ first the -ms current would be,
cos (13) 6s r
(m ,/2 E/2E ZR(0)
Substituting this in equation 12, Z 1 27rn i- L2ms
jn

i= V 2E (cos cot- cos 0) (14)

1/)
Next will be derived an expression for w6 i __ -S Practical Formulas
the saturation angle for any cycle. The , 3 r- -
_=____+ 5
ac-
The foulasdenr veda bove should be
t t_ _ _ _ _ l -1t81modified by experience factors. These
MA VAV
A~~~~~~
VVVV~~ VVV~A~
V V AA
VV VV c Jo O o o o g8 g also w ill be affected by the formulua chosen
CYCLE NUMBER tO calculate X8. It wTas deCided tO cal-
JU! _
Figure 4. The saturation angIe 0 as a function ~~~~~~~~~~~~~~~~~~~culate
give theXtsame
in such a manner that it will
maximum current for the
Figure 3. Oscillogram of magnetizing inrush of the number of cycles for various values of first cycle as in reference 1.
for a 10-kva transformer. The peak of the first k3 R/X. The curves are drawn for 0=_180
cycle of current, lower trace, is 663 amperes, degrees except for the curve for k3 R/X=0.2 X 20fN2A8 10-8 h(
and the rms value of the voltage is 450 volts where 0,= 110 degrees h

324 Specht-Transformer Magnetizing Inrush Current AIEE TRANSACTIONS

Authorized licensed use limited to: Universidad Industrial de Santander. Downloaded on September 19,2023 at 17:54:32 UTC from IEEE Xplore. Restrictions apply.
 Three experience factors were decided moved over one cycle number. This 200 _ _ _
on, the first is for the saturation angle for suggests the possibility of using one curve I
I _K3 _
the first pulse: for all values of 01 for a given value of 80
k3R/X. Such a set of curves are shown in 60
0, =kcsI
Bo
Bms-BR (33) Figure 4 for various values of k3R/X. To 3
Bm ~~~~~~use
this curve, calculate the value of 01 20-
The second experience factor is for the according to equation 33.
maximum current for a cycle, On the appropriate curve on Figure 4 '
find the cycle number where the angle is 6
S4
imk2 v
X (1- cos 0) (34) equal to 0a. This number need not be an
integer. Call this cycle number p
The third experience factor affects the 02 Will then be read from the curve at
decay of the inrush. p+l 03 at p+2, and so forth. The por- .1 .2 3 .4 .6 .8 2 3 4 6 810 20 30
n-1 tion of the curve between the first and CYCLE NUMBER
R x < second cycle number curves quite rapidly.
cos e = cos 0, + k3 L-'(0) (35 ) This portion of the curve was determined Figure 7. 2R(o) as a function of the number
X indirectly. An intermediate value of 01 of cycles for various values of k3 RIX
An analysis was made of seven units
that lies between the first and second
cylnubrwsaumdad02a-
from 10 kva to 20,000 kva that had had Ycle by e as tin u5eT,hn t2 cyle tegrated iR term in equation 4 cannot be
their magnetizing inrush measured and number fy 0q is n 35. Then the cycle neglected. The value of 0.2 for k3R/X is
the average values of these constants are 0umark on the curve
at c actually about as large as can be used.
as follows: 0, marked on the curve at ths cycle see the Appendix for further discussion of
number less one. The curve for k3R/X = this point.
k, =0.90 0.2 starts at 01=110 degrees. This is Figure 5 is a curve of 1- cos 0. The
k2=1.15 necessary, for angles greater than 110 value of iS
by this,~~~~~~~~~~~~~~~
p determnined from Figure 4 as
k3 =2.26 L 45aD X .
degrees cannot be considered described previously. The value of 1-
The value of k3 was chosen so that the method for k3RX =0.2, for then the in- cos 0 for the first cycle is read at cycle
peak of the sixth cycle of current was number p, the value for the second cycle
matched. 2 at cycle number p+1, and so forth. The
A value of 20.2 kilogausses has been t maximum current for each cycle is then
used for B,, the saturation flux density, for calculated by equation 34.
both hot and cold rolled steels. The .6 Figure 6 is a curve of 5A(e) using the
values for the residual flux density used '4 -N values of 0 from Figure 4. To use this
are as follows: V curve to find the value of 2A(0) for n
BR =60 per cent B,, for hot rolled steel. cycles proceed as follows. Find p as
BR=90 per cent Bm for cold rolled steel. -N IdescribedfromFigure4. Readthevalue
BR =50 per cent Bm for type C cores. .06 of -(0) at n+p-1 and at p-i. The
The coefficient k2 also will apply to .046
.03 iV TIIXI }X1X N lhl A1 difference of these two quantities is the
desired summation. The average cur-
equation 27 and 31 for the average and .02 rent can then be calculated from equa-
rmscurrentsasfollows: -ll E-111 1111111 l$lYX1g\ tion 36. On Figure 6, the portions of the
n CYCLE NUMBER to 0D 'o curve for cycle numbers less than one
k2V-2E A(0) (36) Figure 5. The value oF I - cos 0 as a function
were deduced by an indirect process.

X L 27rn J
L1 - '/2
of the number of cycles for various values of
k3 R/X. The maximum current for a particular
Of
of 0a
2val(o)
at be equal to
p- d t be equal to
v(al)
11/, where 0l corresponds to p. Therefore the
e

r cycle is proportional to 1- cos o

irms-n= l
kIV2EIZER(6)
2
1
j
(37) value of A(0) at p-i isequal to the value
of 2tA(0) at p, minus A(0,). For this, of
2001 1111 11111 11 11111 11 1i course, p lies between 1 and 2.
It is possible to draw curves of these - Figure 7 is a curve similar to Figure 6
various functions aud reduce the labor of 60 except for ZR(0). It is used in the same
calculations to a minimum. From equa- 40 manner as Figure 6. With this, the rms
tion 35 it is seen that 0 is a function of 0I, 20 current for the first n cycles can be cal-
k3R/X, and the number of the cycle, n. °l |L culated by equation 37.
If, for example, the series of points for 1 EXAMPLE
0~= 180 degrees and k3R/X=0.1 are13
plotted, itwillbefound02=112degrees, |||||||||||| Anexamplewillbheworkedoutforthe
03=92 degrees, and soforth. If the series N ;&l1111\111111 1 111 1 transformer which has the magnetizing
of points corresponding to 01=112 de- .6 lll | il||| i |1i | nrush curve shown in Figure 3. This
grees and k3R/X= 0.1 are plotted, the 3|t111111!111111 } 111 l transformer is rated 10 kva, 60 cycles.
new 02=92 degrees, the same as 03 above. ili11111111 W1 The value of 0s calculated by equation 33
It is rather evident from an inspection of l_ > tt - @ 0t 0000o°o was 95 degrees compared to 85 degrees
equation 35 that all the points for the CYCLE NUMBER '9 measured. The value of k3RjX was
second series of points must coincide with Figure 6. 2A(0) as a function of the number 0.154. Referring to Figure 4, for 01-=95
the first series if the second series are all of cycles for various values of k3 R/X degrees and k3R/X=0.154, p= 1.8. To

1951, VOLUME 70 Specht Transformer Magnetizing Inrush Current 325

Authorized licensed use limited to: Universidad Industrial de Santander. Downloaded on September 19,2023 at 17:54:32 UTC from IEEE Xplore. Restrictions apply.
to = _ _ - tc = maximum current for a particular
8. = = = = = = == = = S = = _ _ = =====
6 irms -,n = rms value of the current for the first
0- - - - - - -0- - - - - - -n cycles, amperes
~~ 3 --- k =cycle number where the saturation angle
3
=- _ _ =--=O,
= - -== = = = = f = =
- = - 8 is '/4 radian
2
_ =_== _S_ X L k_ =_correctionabout
factor for the saturation
10o -.A3 -7 - ; -.5, S,l -.I _ 2 \angle, 0.90
l=-~~NE4 COS
OS 0 (NEGATIVE.)

_ = = = . = _ =
s
=
T

= =
SNEG
ETI
E --.correction factor for the maximum cur-
rent, about 1.15
- =

k3=correction factor for the decay of the

.8 =
.6 A ~~~~~~~~~~~~~~~~~~~~~~~inrush,
about 2.26
.4.. .
0 X C-CX X C C C
L =X/w, henrys
n=cycle number
3... N =turns in the winding excited
2 \ __ p=number of the cycle where 0 is equal to
0, on Figure 4
ax _}\\ t t _.. _ _ R =sum of R, +RI, ohms
III_ _ t ]t\| |t _ Rs =resistance of the supply circuit, ohms
406 \1 .08 = _ =-) Rt =resistance of the winding excited, ohms
R(O) =R function of 0. See equation 30
t =time, in seconds
.04 _ III II .04 -._ X=sumofX,+Xt,ohms
.03- .03 _L X,=reactance of the supply circuit, ohms
.02 _ __ \
..t | | | .02
t i_ | _
- t
r _ Xt=saturated reactance of the winding
excited, ohms. See equation 32
.. - l- - l A(6)=the |l|\|||ll l| \A function
|| l of 0. See equation 19
.I Z .4
3.5i9
. 0 .1 .2 .3 .4 .5 .6 ol
.8 .9 10 _ l l0
=the angle before the voltage zero point
where the core saturates, radians
G
Ck =the value of 0 for the k'th cycle about
3/4 radian, radians
Figure 8. A t0) as a function of cos 0. It is 0, =the value of 0 for the first cycle, radians
Figure 9. R (0) as a function of cos 0
more convenient to have this plotted as a func- 4, = instantaneous flux linking the winding
tion of cos 0 rather than 0 for use with equation excited, lines
23 m=maximum flux in core due to the ap-
The difference is 3.4. Then the rms cur- plied alternating voltage, E, steady
rent for the first 15 cycles is: state, lines
,= residual flux in core for a particular
determine the maximum current for the 1.15 V2(450) r 3.4 1'/2 cycle, lines
fifteenth* cycles, read the ~~~~~~~~ ~~irms - 1 = +=residual flux in core at moment voltage
I,
first, sixth,1-and
value of
'~~~~~~~~~~~~~~~~1
cos 0 from Figure 5 at the
X2i' 0.956 L15 X
= 146 amperes
24, residua
is applied to winding, lines
is
s = saturation flux of core, lines
fu wiinncramoen
following cycle numbers: w = 27rf, radians per second

1+1.8-1= 1.8: 1- cos0=1.1 Conclusions

6+1.8-1= 6.8: 1-cosO=0.29
15+1.8-1=15.8: 1- cos 0=0.10 This method of calculation will make it
possible, by use of the curves, to readily Appendix
The maximum current for the first calculate the maximum value of the mag-
cycle is from equation 34: netizing inrush current for any cycle of Maximum Values of k3 RIX
1.15V2- (450) X.X1 current, the average value of the current
'large. If R ki
0-956 =841 amperes for the first n cycles, and the rms value of This value of R/X must nottobeX,tooit
0.956 is too large compared
the current for the first n cycles. The ac- will start to control the current. The
For the sixth and fifteenth cycles the cur- curacy is apparently as good as usual formula was derived on the assumption that
rents are 222 and 77 amperes. The meas- calculations for magnetizing inrush,! 40 R is negligibly small in comparison to X.
ured values are 663, 265, and 106 am- per cent or so. Substituting into equation 10 the exact
value of 4 from equation 4,
peres.
To detennine the average current for N10-8
the first 15 cycles, read from Figure 6 the Sym o s mL r O4,m-sv+ m cos cut-
value ofZA(O) at the following points. A, =area inside the mean turn of winding 10 iRdt) (38)
15+1.8-1 =15.8; 2;(0) =12
151.8-1=15.8; 21(0)=5.3 excited, square inches
Bm=maximum flux density in core due to
N J3
1.8-1=1.8; Z(O) =5.3 the applied alternating voltage, E, If the integrated iR term for a full cycle is
steady state, in kilogauss held to less than 1/2 of the maximum of the
The difference is 6.7. Then the average B. =residual flux density in core at moment rest of the terms, it should not be too serious
current for the first 15 cycles is: voltage is applied to winding, kilo- to neglect the last term.
gauss
av =1.15V<(450) 6.7 5s= saturation
B4 flux density of core, kilo- 8r 1
0.956 15 >K2ir amperes E =supply alternating voltage, rms volts 545gus12jRt
N.J 4r2, ,)(9
To determine the rms current for the f =frequency of supply voltage, cycles per
first 15 cycles, read from Figure 7 the h-height of winding excited, inches
values of ZR(0) at the following points: = instanltaneous current in winding excited, 10' R X0E A(OK5 (4,r+24,m -4,s) (40)
amperes N X X 2
15+1I.8-1 =15.8; ZR(0) =11.2 yiavg -n=average current for the first n
1.8-1 = 0.8; 2ZR(0) =7.8 cycles, amperes And with equations 3andl13, equation 40

326 Specht-Transformer MCagnetizing Inrush Gurrent AIEE TRANSACTIONS

Authorized licensed use limited to: Universidad Industrial de Santander. Downloaded on September 19,2023 at 17:54:32 UTC from IEEE Xplore. Restrictions apply.
can be rewritten, Let k be the cycle number where 0 is 3/4 ik3R x
A(8) (48)
k

R cos 6 r+24r-s radian or less. Then equation 35 can be cos 0R Cos 81+
- 'A(0) < 2 2
2 (41) written as follows:

Introducing k3, + \ Also by appropriate substitutions anid itnte-

cos 0= cos 0l+k3x A(O)+ grations,
k3 RIX:I-2A(o)
kX
cos 6
(42)
1R n-I
n
n2 k32R
A d

For 0=180 degrees it is 0.16, for 120 degrees k3 f A(O)dn (45) A()=/
3 A(O)dn=J 3L3X X
it is 0.20, and for 90 degrees it is 0.25. For X It k
smaller angles this increases indefinitely.
Hence, if it is all right for the first cycle, it is Substitution of the integral for the summa-1 - X k32R
all right for the rest of the inrush. tion is a good approximation for A(O) does (tJ 2k3R L 3X
not change appreciably in one cycle for 0 1 -21n
Relations for 0 Less Than 3/4 less than 3/4 radian.
~~~~~~~~~substituting Differentiating this and
the value of A(0) from equation
-
0kJ
k
k 2k1R
(0k2 02) (49)
RadianRadian 43,
By a similar process,
It is possible to make certain simplifica- dB2R
tions when 0 is less than 3/4 radian that will -sin 0- =k3 - 3 (46)
be quite useful and reduce considerably the dn 3X X
) x (o -0)(50)
time required to draw the curves such as
Figures 4, 5, 6, and 7.
al etn4Ok3R
less
For angles less than 3/4 radian, sin 0 is ap-
F

k
(50)
proximately equal to 0 for the accuracy
A(6) =2(sin 0-0 cos 0)_ desired. Making that substitution and
203 (1-10 2) 3 (4 integrating the resulting equation from Ok ReFerences
1 1. TRANSFORMER ENGINEERING (book), L. F.
3
R(O) 20+6 cos 20 - 2 sin 20ctk k32R =_ 1 (47)
Blume. John Wiley and Sons, Inc., New York,
N. Y., 1938, pages 23-36.
3X(n-k)+-k 2. ELECTRICAL TRANSMISSION AND DISTRIBUTION
4B5 t 402 40 REFERENCE BOOK, (book). Westinghouse Electric
154) \ 21 15JWhereand Manufacturing Company (East Pittsburgb,

Discussion element, very little being dissipated from

the element during the short interval needed. if 10 times normal full-load current had
been flowing for 0.1 second.
G. Camilli (General Electric Company, The following equation may be written This leads to a rather simple criterion for
Pittsfield, Mass.): The author of this paper applying to the heating during the inrush selecting fuses that will withstand magnetiz-
has developed an interesting method for the period: ing inrush currents safely. Allowing what
calculation of the magnetizing inrush cur- 12t K appears to be a reasonable margin in excess of
the total I2t in the transformer inrush, it
=
rent curve of a transformer. From the
practical standpoint, however, it appears where I = current in the fuse in amperes has been suggested that fuses be selected on
that the formulas have very limited appli- t = time in seconds . the basis that at the 0.1-second point on the
cations. It is stated that the information K = constant depending on the size characteristic curve they be capable of car-
will be useful in precalculating the values of and material of the fusing ryig at least 12 times the transformer full-
any protective device, and for the selection element load current.
of fuses. As far as the protective devices Mr. Specht's formulas apply only to a
are concerned, the knowledge of the effective The Pt of any fuse may be obtained from single-phase transformer with one wound
value and average value of the inrush cur- the manufacturer's published characteris- (that is, active)rleg.
rent for the first few cycles may be of some tics, usually given as a time-current curve In energizing a 3-phase transformer the
value only when it is desirable to desensitize plotted on logarithmic co-ordinates so that inrush current generally will be quite dif-
the differential relays or to make them en- this portion is represented as a straight line. ferent in all three phases. Furthermore,
tirely ineffective for a few seconds or until It is only necessary then to know the I2t analysis may be complicated by the fact
the inrush current has had an opportunity corresponding to the transformer magnetiz- that all poles of the switch may not close
to decay. Unfortunately, this practice ing current, in order to select a fuse that will simultaneously. Statistical studies, in these
tends, in some measure, to defeat the pur- withstand it safely. cases, similar to those made in reference 1 of
pose of the differential relays, since insula- The 12t representing the inrush for any this discussion are more useful than analyti-
tion failures in transformers have been specific transformer is, of course, a function cal studies based on uncertain conditions.
known to occur at the instant of energizing. of both the magnitude of the first current
The calculation of the wave shape of the loop and the rate of decay, neither of which REFERENCE
inrush current has no practical consequence is usually known without a rather exhaustive
when harmonic restrained differential relays study of the transformer and the system to 1. TRANSFORMER MAGNETIZING INRUSH CURRENTS
.ri-icirl. In +'pQt-
are used il
thzese relays
oln-u the harmonic re-
tlli-Ihnrm"10 r- xx1 .-
which.fit is.Q connected.
i of tPqtv'h
SerieS nf
AA qnrip,,- tests1 has AND INFLU1NCE ON
S. B.SYSTE;M
Blume, G. Camilli, OPERATION,
Farnham, L. F.
H. A. Peterson.
straining forces are provided by harmonic indicated that the total IPt of the inrush AIEE Transactionzs, volume 63, 1944, pages 366-75.
components of current, and, since typical in- may be approximated with sufficient ac-
rush current wave forms are rich in harmonic curacy for the purpose of fuse application.
content, relay operation is prevented so long Although the maximum crest inrush current T. R. Specht: It is interesting that it has
as the differential current is of this type. of the several sizes of transformers tested been concluded by Mr. Camilli that fuses
In applying fuses in series with trans- ranged from 11 to 20 times their normal full- can be applied by the rule that they be able
formers, care should be taken to select fuses load crest current, and the interval re- to carry 12 times the transformer rated
which are large enough so that they will not quired for these transient currents to decay current for 0.1 second. I calculated the
be damaged by the transformer inrush cur- to normal values extended from 33 to 120 equivalent magnetizing inrush current for
rent. Essentially all the heat developed in a cycles, it was found that for each trans- three transformers referred to 0.1-second
fuse during the transient inrush period goes former tested the total 12t in its magnetizing basis. The values for 10-, 25-, and 6,000-kva
into raising the temperature of the fusible inrush was numerically almost the same as transformers are 11, 7.5, and 9.2 times nor-

1951, VOLUME 70 Specht-Transformer Magnetizing Inrush Current 327

Authorized licensed use limited to: Universidad Industrial de Santander. Downloaded on September 19,2023 at 17:54:32 UTC from IEEE Xplore. Restrictions apply.
imal cLurrent With the number of variables to single-phase tranisformers with more than formers ani approximation to the current
that cotuld affect the imagnetizing inrush one wound leg The saturated reactance of can be determined from the single-phase
current, this is in good agreement each coil could be calculated and the sum magnetizing inrush current by the methods
Tlle formnulas in the paper are applicable used for the formulas. For 3-phase trauis- given in reference 1 in the paper,

A n Investigation or Audio oise weighted scale normally uised for noise

levels up to 60 db; the B scale for noise
levels up to 90 db; the se C scale'for
in Substation-Type
in station- Transformers
u pe rans ormers rprwrrcreoihAcl
~~~~~levels
above 90 db. All readings in this
ic
report were recorded on the A scale, since
it is almost impossible to obtain average
J. H. VIVIAN R. R. PECK noise level readings in a given study by
FELLOW ATEE MEMBER AIEE jumping from one scale to another, and
since the majority of readings taken "ere
below 70 dl).
U NTIL RECENT years, the audio 1. Are the present NEMA standards ade- Noise Level of New Transformers. In
noise emnanating from substation quate? order to verify or disprove the contention
transformers occasioned few complaints, 2. If the present standards are not ade- that the transformers now being received
l)ut as sulbstation installations in con- quate, what is acceptable? are close to NEMA Standards, a number
gested areas increased, utilities became 3. What is the practical solution for cor- of new transformers were checked in the
aware of the necessitv for limiting the in- recting transformers with high noise levels Test Laboratory. Table I lists some of
terisity of transformer noise. Recogni- which have already been installed or have the typical transformers.
tion of this developing noise problem a few The noise levels of new transformers are
vears before World War II led to the The Southern California Edison Coin- generally close to the NEMA upper limit.
collaboration of utilities and transformer pany felt that the problem should be in- It is interesting to note that although the
manufacturers for adopting some sort of vestigated by studying the conditions on NEMA Standards allow higher noise
noise standards. This action resulted in its own system. This study would then levels for larger transformers, the cus-
the establishment of the National Elec- give the company firm grounds for any tomer who hears the transformer noise
trical Manufacturers Association (NE- change in standards it might request, and makes no such distinction. Some noise
MA.X) Standards* which consist primarily would also be of value to other utilities complaints have been received when a
of tabulated ratings of the average noise facing the same problem. small bank has been replaced with a bank
levels of trailsformers then being pro- of larger rating, even when the noise level
duced, plus a margin of safet. Souther California Ed.ison of the new bank conformed to the NE-
Attention is called to two developments Compan Investigation MA Standard.
relating to the application and manufac- ompanyDetermination of Ambient Noise Levels.
ture of transformers that have taken
place since the adoption of the present Noise Measurements. For the pur- One of the prime factors in deciding
standards: pose of this study, the unit decibel is used able r no ise ientise
throughout. In general, we may define able or not depends on the ambient
o
noise
1. Economiiic studies have favored the therdecibelas
h
the smallest changeinsound
eie stesals hnei ond level and how much above or below this
utilizationi of the unit-type substatioii lvlatasomrniecnb dni
This has resulted in a marked increase in the level which can be detected by the level a transformer noise can be identi-
use of such substations located in well-
developed residential and comnmercial areas.
human ear. The zero decibel (db) level
is at the threshold of hearing and is very
fied. Atypicalresidential,commercial,and
industrial area was selected for a 24-hour
Often the transformers are immediately seldom reached except under laboratory test. Readings were made with a General
adjacent to homes and apartment buildings. conditions. too
2. Manufacturers, in producing these unit- Measurements were made with the 90
type substations, incorporated the latest General Radio Sound Level Meter and X 80
type cold rolled steel which could be worked . . ' c, 801
up to 100 kilo lines per square inch. This whenever possible, were made in accord- 7
was done to make the use of such stations ance with the NEMA Standard pro-
econoically feasible, but in so doing, cedure. Three scales are available from 60 X.
nlecessarily increasedl the noise level of the this meter, namely, the A scale which is a ,,5
tranlsformlers.o 0p/\

Th
-lcn fniirtasomr
close to residences hlas, of course, resullted
Paper 51-55, recommended by the AIEE Trans-
formers Committee and approved by the Tech- 20
inl comlplainlts. Utilities nzow face the nicai Programi Comgmittee fo presetntatiwonkat
following three questions: N. Y., January 22-26, 1951. Manoscript submitted
Cl261
________ ________ ________ ________

* NEMA Standards for Transformers.

_______

Publica-
October i8, i950;
December 5, 1950.
made available for printing
AM.
PM
12.
tions Number 48-i32, Section TR3-130, National J. H. VIvIAN is with the Southern California Edison Figure 1. Noise ievels encountered in a
Electrical Manulfactulrers Association (New York), Company, Los Angeles, Calif., and R. R. PECR is eidnilae
tyia
Septembller 1948. with the Line Material Company, Milwaukee, Wis. tpclrsdnilae

8328 Vrivian, Pcck-Audio Noise in Substation-Type Transformers AIEE TRANSACTIONS

Authorized licensed use limited to: Universidad Industrial de Santander. Downloaded on September 19,2023 at 17:54:32 UTC from IEEE Xplore. Restrictions apply.