oisture Lqulibirium Between Gas formers and in the 120,000-volt gas-pres-

sure cable.'-3 This increasing of the

moisture content of the gas can be of
i
pace an rous ateria s in ~~~~practical
significance only if the dew
point of the gas is reached at local cool
spots over considerable periods of time,
Enclosed Electric Equipment thus causing progressive transfer of
moisture from the warmer fibrous ma-
terials, and if the resulting condensate
JOHN D. PIPER drops upon a vulnerable part of the equip-
NONMEMBER AIEE ment. Any condensation that takes
place as a result of an over-all cooling of
the equipment t or of a cool spot of short
THE USE of an inert gas_ in enclosed at electric stresses that are higher than
-~~~~~~~~~~~~~~~~~~~~a durationl nocsignificanc seot
durati iSisn of noof significance bbecause the
*
electric equipment has increased con- otherwise would be possible.
siderably in recent years. In transform- The purpose of this paper is to call at-
water content of the gas is relatively low.
During severely cold weather, it is
ers an inert gas is used in the space pro- tention to a phenomenon that occurs in Dring seve coldeather,oituis
vided above the transformer fluid to al- enclosed electric equipment in which in all aleof the transformers that haveoccurs
thatsomerndnat been
low for its volumetric expansion with both an inert gas and fibrous material investigated. Moreover, it is probable
temperature rise. There the function of are present. The systems involved tend that, in certain transformers, condensa-
the gas is to provide an inert medium by to establish an equilibrium between the tion occurs during the cool part of the
which the pressure at the surface of the moisture content of the fibrous materials day nearly every day of the year. Ap-
transformer fluid is kept slightly above and that in the gas space. Considera- parently the condensate usually drops
atmospheric. Thus the fluid is not ex- tion, therefore, should be given to the in locations where it does no harm. Con-
posed to the moisture and oxygen of the moisture content of the fibrous materials densate has been found, however, on the
air, as it is with the so-called "open- present as well as that of the gas itself. terminal board inside transformers.
breather" or "oil-conservator" types A quantitative method is suggested for Whether or not service failures actually
unless the latter communicate with the predetermining the equilibrium condi- have been caused by condensate on vul-
outside atmosphere through chemical tion for both unimpregnated and impreg- nerable parts is unknown, but there seems
agents. In gas pressure cables, however, nated fibers. It is believed that use of little question that such condensation
the inert gas is used at pressures up to the method may assist in the solution of constitutes a potential source of trouble.
approximately 15 atmospheres in order manufacturing and operating problems. In the case of the 120000-volt gas-
that the cables may operate successfully pressure cable, in which the paper-insu-

Paper 46-160, recommended by the AIEE

commit- Field Experience and Data lated unsheathed cables rest on the bot-
telcntiin-ip,neryl
tomE ofa
tom of a steel containing-pipe, nearly all
tee on power transmission and distribution for pres-
entation at the AIEE summer convention, De- Th
troit, Mich., June 24-28, 1946. Manuscript sub- e inreasing of the moisture content locations are vulnerable to attack by liq-
mitted March 29, 1946; made available for print- of dry nitrogen by the transfer of moisture uid water should condensation occur at
ing May 28, 1946. from fibrous materials has been observed local cool spots. Such cool spots might
JOHN D. PIPER iS in the research department of The
Detroit Edison Company, Detroit, Mich. in The Detroit Edison Company in trans- occur at a terminal, a length that crosses

to be free from deterioration caused by requisite high voltage introduces a safety able ductive coating on glass is the most desir-
abrasion. hazard, although placing the NESA electrical method for providing heat
aging, sunlight, overheating, and coat-
the ing
Dulring operation the inner surface of During next tothe
operation theinner Wil.l tend.
vinylsurfe mitigate
of thto ig vat
the windshield. Indeed, unless operat-
nt to te ing experience proves otherwise, this method

windshield is not excessively hot, which the danger. Windows or shields made of appears to be suited ideally for the purpose
improves pilot comfort by minimizing the plastics such as Plexiglas cannot be de- of windshield anti-icing. For most air-
radiant heat. The installation should iced or defogged by this method since the planes, it ultimately should supplant other
prove reliable and safe. Structurally, a NESA coating cannot be applied to their presently known systems.
clean design should result, the only com- surface, at least not by present techniques.
plication being the means for attaching References
the power wires and control elements. Conclusions
There are disadvantages to the scheme, 1. AIRPLANEB AIRWORTHINESS, TRANSPORT CATEB
however. Most important is the weight,' prime anti-icing, GORIES.
additional Washington,
an windshields Manual Aeronautics Board,
Civil 04.38013.
D. C. 04,Section
reasonfrom
1. Aside for heating aircraft
which may be high because of the neces- is evident For a given weight and con- 2. IMPACT RESISTANT WINDSHIELD CONSTRUC-
sity for using high voltage alternating struction, vastly improved impact resist- TION, George L. Pigman. AeronauXtical Engineering
current delivered by inverters or engine ance can be obtained b y using a l aminatedReiw.oue4 aur 95 ae -5
driven alternators . The control system panel with the vinyl plastic maintained at 3. AN AIRCRAT DOUBLE WsNoSsssRLn-ITS D}-
adds significant weight, although a unit about 11 egrees Fahrenheit. Society of Automotive Engineers, volume 51,
to effect the same order of regulation in 2. The method of heating a windshield October 1943, pages 350-5.
preention
Other ice s chefiguration of the cockpit and of the wind- I CB ON TEE1 AIRPLANE WINDSIIIBLD, L. A. Rodert.
comparable weight. Inability to operate shield system. Availability of heated air National Advisory Committee for Aeronautics,
the ,VESA film directly from the ship's and (or) electric powrer are equally impor- Technical Notes, number 754, March 1940.
28-volt power supply and provide the tant factors. 5. Goldschmidt,
H. John A.
Partridge. (bookr),
E. P.TRANSFE:R
INDUSTRIAL HEAT Schackr,
Wiley and
necessary power is a disadvantage. The 3. Use of the transparent electrically con- Sons, New Yorkc, N. Y., 1933, pages 173-207.

DECEMBER 1946, VOLUME 65 Piper-Moisture Equilibriuxm TRANSACTIONS 791

 A0ooc 4
90o
* >o.\\ & z 800 C
\\\\ /z0
IIOC 70 C DATA FOR SPRUCE WOOD
60 C INTERPOLATED TO 25 C
\ \ \\X\\ 100 / ,/; f / < FROM tNSERT BELOW VIA
KRAFT PAPER AT 90 C 50 C FIGURE
0cia. (VALUES EXTRAPOLATED 40C
FROM DATA IN FIGURE II 40 C
o ~~~~ ~~~~~~~~~~~~~~~~~~~~300 42
20 3C29023
\1_0~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~2
0 0.115~~~~~~~~~~~~~~~~~~~~~~~~~0l

\ \O\5\ \ \ ~ // 0.087TO

O\0F005b
~
\
\ \
060NE FIGURE
0.030
2
DATA FOR SPRUCE
'a
\H U 0MA0.RIALS. WOOD FROM FIGURE7
\ \OTO t5J ~ \ " O CAND6OF R12 4 0 W0517QX
*EMPENTURE, OF IN FIBROUS T2N6OLEF
20
(7).
BC /// M

0.019
\0 S P U E W O D ()\\8 / OT
0.012
lil 00l l l
0.1 WATER
NUMRALALONE
SHO \QI1RI \ \ \/
oCOTTON (4)
COTTON 15)
KRAFT PAPER (6) xOTTO
"SPRUCE WOOD (9) COTTON
INTERPOLATED FROM INSERT ~ ~ ~ ~ ~ ~ ~~00INTERPOLATED OR EXTRAPOLATED vALUES
OF FIGURE 2. 0.01 ~ ~ ~ ~ ~ ~~~~~~~~~~~~~~-
~ ~ ~ ~~
SPRUCE
KRAFT PAPER b erD
CONCENTRATIONS DMDED
NUJMERALS SHOW EQUILIBRUIM
0.001595 BY I..

CONCENTRATIONS OF WATER 0.00106

IN THE FIBROUS MATERIALS.' 0.000795
TEMPERATURE,o
150
0OW12O
130 120 110 100 90 10
80 70 60 5040
OCOOSI, 0.000sembly cDo
~~~~~WEIGHT
0.01
OF WATER IN FIBROUS MATERIAL
0p
foun0 3 OO 7 02 0
WEIGHT OF DRY FIBROUS MATERIAL
240 250 260 270 280 290 300 310 320 330 340 350 360 370

Figure 1. Vapor pressure of water alone and muslin tape. This was followed by two Figure 2. Equilibrium between water con-
in Ribrous matefialse data from literature layers of paraffin-impregnated canvas* centrations inabrous materials and vapor pres-
xum
t~~~cn to provide mechanical protection
lolllto tecblduringa sure at selected temperatures
isalto.Atrasiae kdwr
a canal, or a location in proximity to a
when itwaswas
wrapped over the canvas and the as- six samples in which the content varied
watermarin.go
watr

poreason of theintwasla
this
min.Forthi resonwhe sembly coated with petrolatum for lubri- from 1.0 to 2.0 per cent.
fiound durin
the progstressnt-n intaya cant during installation, the cable was 2. Evacuation of the completed line, which
ntiognthatsth moistur contientrofdthed enclosed in a tempora-ry lead sheath that is a normal procedure in the installation of
nediatren gas,ewhichwaslntouclin pedm-
im-
was stripped within a few feet of the en- this type of cable, was extended for eight
mediately after the cable-pulling opera- trance tube to the pipe line as the cable days while the cable was heated by con-
tion, was beginning to increase, steps were was being installed. ductor heating to give a tape temperature
taken co-operatively by the manufac- e g of between 40 and 42 degrees centigrade.8
turer and user to determine the cause and inroen theintodrythenitrogen,
introduced
which was
steel pipe line with
bon-dioxide-alcohol trap and measured.
The water removed was collected in a car-

cfforrective comleasurelysucessary. Tesdem the cable and held at slightly over at- From the average water concentrations
efforts were completely successful as dem- mospheric pressure by temporary plugs, and the weight of water removed, the aver-
onstrated both by the data herein con- . ' age moisture content of the protective tapes
tained and by the performance of this the source presence
the of the moisture of
moisture ,w as the line was placed in operation was cal-
was sought. culated to be 1.73 per cent. Presumably,
circuit1941.
since w
iThThedat.a
atasenergzedingDecem-
The pipe was known to be dry. No water most of the water that was removed came
cable are presented in this papewas
cbe
concermng
this found in the petrolatum grease but from the wetter tapes.
paper only from 2.4 to 5.0 per cent of water** was Inodrtdermewhhrtes
because they indicate that a predictable . .
. In order to determme whether these
equilibrium exists between the moisture found int the impregnated muslin and efforts to prevent the possibility of con-
content of the fibrous material ' even canvas tapes,* based upon the weight of densation were successful, periodic deter-
dry fabric. The average for 14 samples minations . .' were made of the moisture
when impregnated, and that of the dry
gas was 3.3 per cent. Two steps were faric.ntheavrgf
r1 sames taken concentrationst or the dew
points of the
space. to minimize the moisture content of the
As has been described elsewhere1'2 muslinandcanvastapes gas space and of the temperature at
the 120,000-volt gas-pressure cable is of various points along the line. During
preimpregnated-tape construction; that 1. The manufacturer modified his process the first year of operation the dew points
is, the cable was taped in an air-condi- inf stuch a manner that the tmoisture cfontent rose until late summer when at the more
tioned room with tapes that prevriously shipments of cable was reduced to an aver- mitlctos hc otze al
had been impregnated and from which age of 1.07 per cent as judged by tests on not given the special factory treatment,
the excess impregnant had been re- -they were above the winter operating
moved.
moved.Over the~~~
ver th tae isulto
tapea nsulaton two tw * This type of and
manufacturer construction no longer
hence it does is usedthatbydew
not follow the t Originally the moisture concentrations were de-
layers of metallised-paper shielding tapes point problems are inherent in other gas-pressure- terminied gravimetricaiiy at atmospheric pressure
wer aple an vr hs th_ brnz cable installationB.
** As determined by Dean-Starkc apparatus and
and the resuts calculated to line pressure. Later
dew points were determined at line pressure usinlg a
tape intercalated with an impregnated boiling tolnene.' dew-point apparatus.

792 TRANSACTIONS Piper-MIoisture Equilibrium ELECTRICAL ENGINEERING

temperature of the pipe line. In winter, for cotton containing over one per cent of narrow, it is not clear whether the data
however, the dew points fell to values water. The data for each concentration are best represented by straight or curved
safely below the winter operating tem- are represented as a continuous curve in- lines. The slopes of the lines are approxi-
perature. stead of as the two intersecting straight mately the same, however, as those of the
Inasmuch as the rise and fall, with lines used by the original authors. As adjacent curves for cotton.
temperature change, of the vapor pres- thus represented, the curves for the higher In using the four sets of data to prepare
sures that corresponded to the several dew concentrations are nearly parallel to, and a general equilibrium chart, no effort was
points or moisture concentrations fol- have nearly the same curvature as, the made to distinguish between absorption
lowed the vapor pressure rule for liquids, curve for water. With diminishing con- and desorption data. The difference,
within experimental error, it became evi- centration of water, the curves become where shown by the original authors, is
dent that an equilibrium existed between more divergent and their slopes approach small as compared with the effect of
the moisture content of the gas space and that of the curve for ice. This curve for temperature change, which is the primary
that of the tapes. This behavior, coupled ice is not shown in Figure 1. The solid concem of this paper, and probably is also
with the change in dew points of trans- curves at the lower right hand corner of small as compared with the effect of the
formers with temperature change, led to the figure represent the data of Neale and previous history of the samples.9
a search of the literature to determine Stringfellow6 for cotton of low moisture The data for the equilibrium chart for
whether quantitative data existed con- content. These curves are straight, par- cotton were obtained essentially by inter-
cerning this equilibrium for the fibrous allel lines having the slopes of the vapor polation, for selected temperatures, of the
materials generally used in the electrical pressure curve for ice. The curves that vapor pressures that correspond to chosen
industry. Literature on the subject re- appear to extend as broken lines from the concentrations of moisture. The princi-
vealed that although there are many ar- curves representing Neale and String- pal object of the interpolation, of course,
ticles describing the water concentrations fellow's data represent the data of Houtz was to bridge the gap between the data of
of various fibrous materials under various and McLean7 for kraft paper of low Urquhart and Williams and those of Neale
relative humidities, only a few show the moisture content at elevated tempera- and Stringfellow in the concentration
effect of temperature change. Fortu- tures. The data for the samples having range between 0.0016 and 0.012 gram of
nately two such sets of data were available the higher moisture concentrations are water per gram of dry cotton. Interpola-
for cotton and one each for kraft paper best represented by curved lines whereas tion was carried out by means of Figure
and spruce wood. From these data an the data for the samples of lower concen- 2, where the vapor pressures and concen-
equilibrium chart has been prepared that trations fall along straight lines as shown. trations of water in cotton are both
is useful for predicting and interpreting The latter are not parallel with each other. plotted on logarithmic scales for selected
the behavior of the moisture within elec- For the data of Urquhart and Williams, temperatures to give the long S-shaped
tric equipment containing fibrous ma- the slopes of the curves become steeper as curves that represent both the data of
terials and a gas space. the moisture concentrations of the sam- Urquhart and Williams and those of
ples represented diminish. The fourth Neale and Stringfellow. In order to ob-
Preparation of the set of data is that of Pidgeon and Maass' tain the 90-degrees-centigrade curve of
Equilibrium Chart for spruce wood. These data appear in Figure 2, the data of Neale and String-
the same region as those of Urquhart and fellow were extrapolated from 80 to 90
Figure 1 shows the data as obtained Williams. Because the temperature range degrees centigrade by extending the
from the literature. In this figure the for the investigation on spruce wood was straight lines as shown in Figure 1.
vapor pressures, on a logarithmic scale, Because only one set of data was avail-
are plotted against the reciprocals of the able for kraft papert and that at high
absolute temperatures on a linear scale. Figure 3. Comparison between the concen. temnperatures only, interpolation similar
The top curve is for water alone. The to that described for cotton could not be
tration of water in cotton and that in kraft
long solid curves below the top curve rep- carried out. Instead, the curves in Figure
paper or spruce wood required to give the
resent the data of Urquhart and Williams5 1 for the data of Houtz and McLean were
same vapor pressure at the same temperature
extrapolated downward from 100 to 90
e ..... .__. . .... . ._ ___ -. | Wdegrees
w J centigrade, and the resulting
xCOTTON VS KRAFT PAPER AT 90 C
/ COTTON VS KRAFT PAPER AT 90 Cvalues of vapor pressure were plotted
COTTON VS SPRUCE WOOD AT 25 C against their respective concentrations
in Figure 2. The straight line that re-
sulted when the extrapolated values for
kraft paper at 90 degrees centigrade were
g 00 plotted on the double-log scale of Figure 2
w / lies slightly below the curve for cotton at
i / , 80 degrees centigrade as shown. Next, it
r
/ was reasoned that inasmuch as kraft pa-
o2 per often is made from spruce pulp, pos-
S // ~~~~~~~~~~~~~~~~~~~sibly
the equilibrium in kraft paper and
Sxo t ~~~~~~~~~~~~~~~~~~spruce
pulp might be related. For this
0 i / ~~~~~~~~~~~~~~~~~reason
the data of Pidgeon and Maass,8
/ l | | ~~~~~~~~~~~~~~shown
in the insert of Figure 2, were in-
/^~~~~~~~~~~~~~~~~~~~~~~~
Vincent and Simons15 evidently investigated the
/ ~~~~~~~~~~~~~~~~~~~~~~~~field
over a wide range of temperatures and moisture
/x ~ ~ concenltrations but unfortunately did not include in

CONCENTRATION OF WATER H KRAFT PAPER OR SPRUCE WOD equilibrium char.

DECEMBER 1946, VOLUME 65 Piper-Mfoisture Equzilibrium TRANSACTIONS 793

 WEIGHT WATER PER UNIT WEIGHT OF DRY FIBER
WOOD a KRAFT PAPER

N~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~-

IN,~~ ~ ~ ~ ~~'
IN~~~~~

cr-
I~~~~~~~~~~~~~~
Ul)
w N
a.

0
a-

m ln

~ 0.01
- ~ igr 4o Equilibrium
t charts tw
h c i
terpolated by the use of Figure 1 to 25 sorbed moisture, it might appear that pleted line, as calculated from the analyti-
degrees centigrade. Finally, the concen- impregnated fibers behave entirely dif- cal results and from the quantity of wa-
trations of water in cotton required to ferently from unimpregnated ones. The ter removed, was 1.73 per cent. The
give the same vapor pressure at 90 data for the paraffin-impregnated cotton points indicated by dots in Figure 5
degrees centigrade as are given by se- protective tapes of the 120,000-volt cable show the concentrations of water in cot-
lected concentrations of water in kraft previously described provide a compari- ton at each of the 30 joints at which the
paper and those concentrations required son between the behavior of commercially samples were taken. These were deter-
to give the same vapor pressure at 25 de- made American impregnated-cotton tapes mined with the equilibrium chart from
grees centigrade as are given by selected and that of the two sets of laboratory- the average temperature of the line and
concentrations of water in spruce wood prepared unimpregnated-cotton fibers the vapor pressures that correspond to the
were plotted on a double-log scale as made in England. dew points that were determined in Sep-
shown in Figure 3. The straight line The comparison between the impreg- tember 1944. The values, which because
through these points indicates that the nated and the unimpregnated cotton they were determined by the chart are
relationship between the water concen- samples is given in Figure 5. The solid for unimpregnated cotton, ranged be-
trations that are required to produce the lines in this figure, and the numerals ar- tween 0.9 and 2.65 per cent and averaged
same vapor pressures in cotton and in ranged in a column upon them, are repro- 1.68 per cent. The agreement between
either kraft paper or spruce wood is sim- duced for cotton from Figure 4. The 1.73 and 1.68 per cent is much closer than
ply the ratio 1 to 1.7. crosses represent the average conditions the analytical methods warrant. The
By dividing the concentrations of wa- at three locations at which the higher agreement indicates that it is safe to use
ter in kraft paper, as given in Houtz and moisture concentrations were found in the the equilibrium chart for impregnated
McLean's data, by 1.7, an estimate was gas space as the average temperature of cotton at least in the solution of practical
made of the vapor pressures that would the pipe line rose and fell with seasonal problems in which an error of perhaps
result from low concentrations of water in and load change. The low-temperature plus or minus two in the second signifi-
cotton at 100 and 110 degrees centigrade. point was determined while the line was cant figure can be tolerated.
These data were combined with the ex- not in operation. The circles represent No data are available to show whether
perimental data of Urquhart and Wil- the average conditions at two locations impregnation with insulating oil changes
liams in that temperature range to pro- at which the lower moisture concentra- the equilibrium temperature for kraft
duce the top two curves of Figure 2. tions were found. It is evident that, re- paper. The data for the paraffin-im-
The equilibrium chart was prepared by gardless of whether the fibers are impreg- pregnated cotton support the belief that
plotting the vapor pressures for cotton, as nated or not, the change of the vapor pres- although the impregnant greatly slows
given from the curves of Figure 2 for se- sure of water in the gas space is essen- down the diffusion process by which
lected concentrations, on a logarithmic tially the same function of temperature equilibrium is attained, the equilibrium
scale against the reciprocals of the abso- change. It is also evident that although concentration of water in a fibrous ma-
lute temperatures, similarly to Figure 1. equilibrium had not been reached when terial is not affected greatly by an organic
This chart is shown in Figure 4. All of some of the data on the cable line were ob- impregnant. It seems probable that the
the curves are extrapolated from 20 to 10 tained, especially those taken in Septem- equilibrium conditions shown for unim-
degrees centigrade. Those for concen- ber 1942 the moisture concentrations in pregnated kraft paper in Figure 4 will
trations above 0.03 gram of water per the cotton, as indicated by the chart for apply approximately for oil-impregnated
gram of dry cotton also are extrapolated different dates and temperatures, are kraft paper as well.
for temperatures above 80 degrees cen- nearly the same. Even including the
tigrade. For this extrapolation, the gen- September 1942 data, the indicated con- Use of the Equilibrium Chart
tle curvature of the vapor pressure curve centrations ranged only between 2.2 and
for water, shown as the broken line con- 2.6 per cent for the tapes in the more moist The following examples are intended
stituting the top curve, was followed. regions and between 1.2 and 1.5 per cent to illustrate a few of the uses that may be
The equilibrium concentrations of wa- for those in the drier locations. Data made of the equilibrium chart in the elec-
ter in cotton that were selected for the taken during the initial operation of the trical industry. The first three of these
preparation of the equilibrium chart are cable in the winter of 1942, however, fall examples concern cable of the type dis-
shown by the numerals of each curve in far below the equilibrium values shown. cussed.
the left-hand column. The numerals in It does not necessarily follow, that be-
the right-hand column represent the cause the vapor pressure changes in the PLE 1
equilibrium concentrations of water in same manner with the temperature Did the precautionary treatments of
kraft paper or spruce wood, as obtained change regardless of whether the cotton protective tapes (previously described)
by multiplying the respective values for is impregnated or unimpregnated, that effect any substantial reduction in the
cotton by 1.7. the equilibrium concentration of water dew point of the gas?
in the cotton is the same in impregnated This question is answered by the informa-
Effect of an Impregnant cotton as in unimpregnated. Naturally tion in Figure 6, which also shows the
it was impractical to obtain samples of method of using the equilibrium chart.
Although the practical value of an thie tapes in the completed
* .
120,000-volt
'
Aspreviously
~~~ments stated, before
the maximum the drying
moisture
treat-
concentra-
equilibrium relationship, such as that cable line after the evtacuation treatment tion was 5.0 per cent, and the average 3.3
described in this paper, has been recog- in order to determine whether the impreg- per cent. After the factory treatment on
nized previously for unimpregnated fi- nant alters the equilibrium. Neverthe- the part of the cables and evacuation of the
brous materials,'0 there seems to be no less, a means for obtaining a qualitative entire line, the maximum concentration as
record of its application to impregnated answer exists. Ais previously stated, the dermndfo thdwpitsas26
per cent and the a-verage concentration from
fibers. Indeed, to many who have tried average concentration of water in the both the dew point measurements and the
to dry impregnated fibers that havre ab- impregnated cotton tapes of the com- calculated residual water concentration was

DECEMBER 1946, VOLUME 65 Piper-Moisture Equilibrium TRANSACTIONS 795

 20.0 NN F\\\ b \ | Figure 5. Similarity of which is of wood origin, the probability

10.0 N
\\Nv@\
N
N \
\\\<\s l in equilibrium be-
havior between im-
of condensation is greater than for gas-
pressure cable. The reason is that the
ECS N-X)&N>
\ D \ : pregnatedand unim- top of the transformer casing may be
6. \ pn cotton only a fewdegrees above outdoor tempera-
4.0
3.0;\
N#$) 'tX
~
+. >
k'o;\ t4
>1\
N|
ture while the fibrous material is at the
oil temperature, which may reach a daily
2.1\ \ < > < lx ISN maximum of 50 to 60degreescentigrade.
EE 2.0 . \ Xl The equilibrium for transformers is more
E l\\ complicated than for the paraffin-impreg-
tlcc1:1.0 \\ X 0
LINES AARE REPRODUCED
nated tapes because the amount of water
~cn~ ~ ~ ~~~0 =) 0.8 \\FR.OM THE EQUILIBRIUM
CHART OF FIGURE 4.
in the large mass of oil is not negligible
w 0.6 PRESSURES AND TEMP- as compared with that in themuch smaller
X cr_\\\ \ \ ~~~~~~~~ERATURES IN THE DE- S mass of fibrous materials. It is probable
a. ETROIT EDISON COMPANY
00: 0DOTS SHOW\ \ PRSSURE CABLE LINE that a definite relationship also can be es-
03\
0. THE EQUILIBRIUM
CONDITIONS AT EAC
ON DATES GIVEN
x AVERAGE OF THREEmosuecnntfoiadthtfte
tablished for the equilibrium between the
02 OF JOINTS
THE THIRTY 1944
IN SEPTEMBER LOCATIONS WHERE
MOISTURE CONCENTRATION
moisture content of oil and that of the
HAVERAGE OF TWO
LOCATIONS OF LOWER
gas space. Until such information is
available, however, it is not possible to
0.1 MOISTURE CONCENTRA-
TION estimate the water content of the oil from
\ \\\ . \ \ \ dew point data. In the following the oil
50 40 30 20 10 is considered only as a barrier that mark-
TEMPERATURE C
edly reduces the rate at which equilib-
rium becomes established between the
1.7 per cent. The dew point of the gas EXAMPLE 3 moisture contents of the gas and fiber.
space corresponding to a point that repre-
sents a given moisture concentration and What concentration of water would EXAMPLE 4
selected temperature of the fibrous material preimpregnated kraft paper tapes ac-
is obtained by drawing a horizontal line quire if they were exposed in the air- The gas in a new transformer had a dew
from that point to the curve for water as conditioned roo described in example point of ten degrees centigrade when op-
shown by the arrows in Figure 6, and ob- condiinedgroom desribein eombe poingtof te degrees
taining the dew point temperature as given 2 long enough for equilibrium to be eratig at an otltemperature of 50 degrees
by the point of intersection. Thus the chart reached? centigrade while the outdoor temperature
shows that the dew point in the vicinity of was well above the dew point of the gas.
tapes having the maximum moisture con- The point representing 507 milimeters of What was the probable moisture content
tent of 5.0 per cent would be 26 degrees mercury and 30 degrees centigrade on the of the kraft paper insulation within it?
centigrade when the summer temperature chart indicates a moisture concentration of
of the cable line averages 38 degrees centi- 4.4 per cent in kraft paper A dew point of ten degrees centigrade cor-
grade, and nine degrees centigrade when the 4 responds to a vapor pressure of 9.2 milli-
winter temperature of the line averages meters of mercury, which at 50 degrees
20 degrees centigrade. Thus it is shown The foregoing example does not imply centigrade is shown by the chart to be in
that the drying treatments reduced both that the moisture content of a preimpreg- equilibrium with a moisture content of ap-
the maximum and the average dew points natedpaeinuainihihjsbeas prxmtl27prcntnteislto.
by approximately 15 degrees centigrade over te paper insulation is high just because pr y . p
the entire operating temperature range of the equilibrium moisture content is high EXAMPLE 5
the line, as shown. for the humidities that feasibly may be
maintained ofinnormalthe taping room. Under The gas in a reconverted oil-conserva-
EXAMPLE 2 conditions operation the time tor type of transformer had a dew point
What concentration of water would a of exposure is too short to allow signifi- Of 20 degrees centigrade when operating
cotton protective tape acquire if the ca- cant absorption. For example, the aver- at an oil temperature of 42 degrees centi-
ble on which the tape is wrapped stands, age moisture content of the insulation of grade while the outdoor temperature was
until equilibrium takes place, in a room the cable previously mentioned was not well above the dew point of the gas. How
air conditioned to a relative humidity of 4.4 per cent but only 0.055 per cent, much water would have to be removed
20 per cent at 30 degrees centigrade, and which is essentially the same as that for from the fibrous insulation of the trans-
the cable thereafter is sheathed? other well-made cables impregnated in a former in order for the gas to have a dew
A 20-per-cent relative humidity at 30 de- conventional manner. An equilibrium point of minus ten degrees centigrade
grees centigrade is equivalent to a moisture value that low would require, according while the oil temperature was 50 degrees
content of 6.0 grams per cubic meter which to the chart, the relative humidity at 30 centigrade?
corresponds to a vapor pressupr of 5.7
milietr of mercury A poit epesnt degrees centigrade to be of the order of A dew point of 20 degrees centigrade, or a
ing 5.7 millimeters and 30 degrees centi- 0.0)17 per cent. Obviously, the low rate vapor pressure of 17.5 millimeters of mer-
grade on the chart indicates a moisture con- of moisture absorption is of much more cury, indicates an equilibrium moisture con-
tent of 2.6 per cent based upon the weight importance to the manIufacturer of this centration of 5.4 per cent at 42 degrees
of dry unimpregnated cotton. type of cable than is the equilibrium centigrad e, according to the chart. A dew
concentration. 1.95 millimeters of mercury indicates a
e Because the concentration of water in grams per The two examples that follow illus- concentration of 1.05 at 50 degrees centi-
cubic meter and the vapor pressure are nearly the
same numerically up to 50 degrees centigrade, onlly trate the use of the equilibrium chart for grade. Thus it would be necessary to re-
pressure shown on the chart by the relIative hulmidi- prbems involving eodnsto in tranls- 4.35 pe ent based upo bthe weight of the
ties to obtain the equilibrium vapor pressures. In .formers. For given equilibrium mois- paper fibers. For a transformer containing
this case, instead of 5.7 millimeters, a value of 20
per cent of 32 or 6.4 millimeters would he obtained. ture content in the fibrous material, most 3,000 pounds of such insulation, it therefore

796 TRANSACTIONS sPiper-Moisture Equilibrium ELECTRICAL ENGINEERING

would be necessary to remove approxi- CONCENTRATION Of DEW POINT AT AVG
mately 130 pounds or 15.6 gallons of water 2W WATER IN TAPES LINE TEMPERATURES
which does not include the amount that
would have to be removed from the oil to \\.ORIGINAL AVG 3.37.
8ORIGINAL MAX 5.07 2 C
18 C
9 C
0 C
reach the same equilibrium conditions. c,c FINAL MAX 2.65 % II C -6 C
d.FINAL AVG 1.7 . 0 C -16 C
Vapor Barriers E 6 \ "
Consideration of the information given en _ \\__

clusion that, if no vapor barrier existed 2

between the cotton and the kraft paper o
2

d
tapes of the cable previously mentioned, g
equilibrium ultimately should become l _____'_____
established between the moisture con- bo 40 30 20 10 0 -10 -20
tent of the cotton and kraft tapes with TEMPERATURE C
resultant reduction in the dew point of
the gas space. The two kinds of tapes Figure 6. Use of the equilibrium chart in tures, in predicting the dew points of the
are present in the fiber-weight ratios of 22 estimating the reduction in dew point that was gas in enclosed electric apparatus, and
to 485. Considering the region in which effected by reducing the water concentration in employing dew point measurements as
the final concentration of moisture in the in impregnated cotton tapes a nondestructive analytical tool for deter-
cotton tapes was highest (2.65 per cent), mining the moisture content of fibrous
if equilibrium were established with the materials in electric equipment.
kraft types, which had an average con- ter vapor. It seems entirely possible that Metallized-paper electrical shielding
centration of 0.055 per cent, the final if an elastomer can be developed to have tapes placed over fibrous materials are
moisture concentration of the cotton suitable life in contact with oil on one side shown to be remarkably effective in pre-
tapes would be 0.1 per cent and of and ground water and air on the other, venting diffusion of water into these ma-
the kraft tapes 0.17 per cent. The equi- it should be a suitable sheathing for ca- tenrals. The significance of this pheeom-
librium vapor pressure at a temperature of bles provided the outside of the insulation enon in connection with the use of an elas-
38 degrees centigrade then would be, ac- of each conductor is shielded by metal- tomer as a sheathing material, particu-
cording to the chart, approximately 0.05 lized paper or thin metal tapes. The larly for solid-type cables, is pointed out.
millimeter of mercury, and the corre- latter without some sort of sheath are not
sponding dew point approximately minus adequate because they will not keep out Ref
45 degrees centigrade. Although estab- liquid water nor provide mechanical pro- eferences
lishment of the equilibrium between the tection. Use of a corrosion and abrasion- 1. 120-Ky HIGH-PRBSSURE GAS-FILLED CABLE
moisture contents of the cotton and kraft resistant elastic material that can be ap- L.Atkinson.
T. Faucett, L. I.TRANSACTIONS,
AIEE Komives, H. W.volumeCollins,
61, R.1942,
W.
tapes would be expected to be slow, there plied competitively with lead seems pages 6°5865.
was no evidence that any exchange had worthy of investigation, particularly for 2. DETROIT PIONEERS WITH 120-Kv PIPE-TYPE
taken place in over 21/2years of operation. solid type cables. The elaborate "sand- Hull. GAS-FILLED CABLB-I,
Electrical World, G. B. McCabe,
volume 117, 1942,F. pageM.
wich"scheme of Rihl and Heering.
This is remarkable when it is realized that WiCh" scheme Of Ril1. and Heering" to to
2214.
the conditions are the same as though the use a plastic sheath seems unnecessary. 3. DETROIT PIONEERS WITH 120-Ky PIP-TYPE
GAS-FILLED CABLE-I,
Electrical World, volume G.118,B. 1942, pageF.278.
McCabe, M. Hull.
cable had been exposed continuously for cable had been exposed continuously for

over two years to a relative humidity of Conclusions 4. SCOTT'S STANDARD METHODS OF CHEMICAL
approximately 30 per cent. Furthermore, ANALYSIS, N. H. Furman, editor. Fifth edition,
volume 2, D. Van Nostrand Company, Inc., New
exposure for several months of unsheathed An equilibrium chart has been pre- York, N. Y., 1939, page 1342.
samples of the cable to Detroit's humid pared from data from the literature to 5. THE MOISTURE RELATIONS OP COTTON, THE
summer atmosphere out of direct contact relate the equilibrium concentration Of EFFECT OP TEMPERATURB ON TEB ABSORPTION OF
WATER BY SODA-BOILED COTTON, A. R. Urquhart,
with rain resulted in no apparent in- water in cotton or kraft paper with the A. M. Williams. Journal, Textile Institute (Man-
crease in moisture content of the kraft vapor pressure of water, over a tempera- chester, England), volume 15, 1924, page T559.
insulation. Of three samples of cable so ture range of 10 to 110 degrees centi- S. 6. M.THENeale, PRIMARY W. SORPTION
OF WATER BY COTTON,
A. Stringfellow. Transactions,
exposed before the line was completed, grade. The concentrations of moisture Faraday Society (London, England), volume 37
the moisture concentrations were 0.01, covered by the chart range from 0.01 to 1941,ADSORPTION
page 525.
OF WATER BY PAPERS AT ELE-
7.
0).01, and 0.05 per cent, respectively, all 10 per cent foradcotton and1.7
e e1.7times
t
re c
those
t ea
VATED TEMPERATURES, C.C. Houtz, D. A. McLean.
s

of which values happened to be less than values for kraft paper. Journal of Physical Chemistry, Baltimore, Md.
volume 43, 1939, pages 309-21.
the average for unexposed samples. The Although the equilibrium chart was 8. THE
prepared from data for unimpregnated Pidgeon, ADSORPTION
OF WATER BY WOOD, L. M.
author's interpretation of the reason the 0. Maass. Journal, American Chemical
insulation does not become moist is that fibers, data taken on The Detroit Edison Society, Washington, D. C., volume 52, 1930, page
the metallized-paper shielding tapes bar Company's 120,000-volt gas-pressure ca- 9.HAOFDSRTNOFWERBPPR,
the diffusion of water vapor into the ble line show that the chart is also valid C. C. Houtz, D. A. McLean. Journal of Physical
insulation. This fact is believed to be of for impregnated materials, at least for Chemistry, Baltimore, Md., volume 45, 1941, pages
further practical significance, for no cotton impregnated with paraffin. 10. A SENSITIVE METROD FO THE DE£TERMINA-
longer can it be said that plastic cable Examples are given .of.the use of the TION
.. . . ......SULATING OP MOISTURE, APPLIED TO ELE:CTRICAL IN-
MATERIALS, R. S. Vincent, A. Simons.
sheaths cannot be used for oil-impreg -chart in predicting the equilibrium ml- Poedxs hsclScey(odn nln)
nated cables on the basis that all such ture content that fibrous materials will volume 52, 1940, pages 489-500.
sheaths, although th}ey provide 'a barrier acquire when exposed to atmospheres Of ~~~~~~~~~~~~~~~~~~~~~~~~~trotechnische
11. CABLE SHEATH, Rihil, H.Germany),
Zeitschriftw.(Berlin, Heering. volume
Elek-
for liquid water, are too permeable to wa- various relative humidities and tempera- 62, 1941, page 1917.

DECEMBER 1946, VOLUME 65 Piper-Moisture Equilibrium TRANSACTIONS 797