Sie sind auf Seite 1von 4

Conference Record of the 1998 EEE International Symposium on Electrical Insulation, Arlington, Virginia, USA, June 7-10, 1998

La s in Paper Tapes for

A Voltage Generators
F. T. Emery E. Barmtia
Westinghouse Power Generation Cogebi
4400 Alafaya Trail Huysmansiaan, 65
Orlando, FL 32826-2399 65B-1651 Lot Belgium
J.D.B. Smith
Westinghouse Science and Technology Center
1310 Beulah Road
Pittsburgh, PA 15235

Abstract: Muscovite mica paper with narrow particle size such high voltage and high temperature requirements.
distribution has been developed to improve properties of tapes to Manufacturing costs are a factor because large coil
be used in the VPI process using epoxy resin. The small particle manufacturing is a series of many critical operations
size, and narrow particle size distribution, results in uniform throughout the manufacturing process. Therefore to realize
penetration of the bonding resin and accelerator into the mica
paper in the process of tape manufacture. This particle size and some cost advantages, the area of the best cost savings is in the
distribution are also responsible for uniform penetration of the reduction of material costs through the use of shorter
impregnating resin. The use of appropriate epoxy binder resins manufacturing cycles, reduction in coil materials, ability to
ensure compatibility with the impregnating resin. The accurate operate at higher temperatures and voltage stresses, and
control of the combination of mica paper appropriate resin and through the manufacture of coils with consistent
accelerator are important contributors to the overall mechanical characteristics.
and electrical properties of the high voltage coils insulated with
these new tapes. In order to optimize the epoxy groundwall insulation used in
the design and fabrication of high voltage stator coils, mica
This paper presents the results of the dielectric performance
when using these new tapes for vacuum-pressure-impregnated tape assessment criteria were defined and are detailed in Table
(VPI) groundwall insulation. The objectives are to develop a glass 1. Working with Cogebi, a major supplier of mica paper and
backed mica tape with higher dielectric properties than mica paper tape, a number of custom tape designs in
commercially available tapes. Custom tapes were developed; conjunction with standard production type mica tapes were
working with Cogebi. These tapes were used to make up insulated evaluated. The evaluation process included making and
test bars which in turn were subject to electrical, chemical, and processing test bars and then subjecting the test bars to thermal
mechanical testing. This paper reviews the results of the tape cycling and electrical evaluation. Limited mechanical and
improvement effort which leads to an improved groundwall chemical tests were also done as part of the evaluation process.

INTRODUCTION Table 1. Groundwall Mica Paper Tape Assessment

Groundwall insulation systems for VPI'ed high voltage stator Criteria
coils usually use an epoxy or polyester resin, glass backed
mica paper (GBMP) tape system. The coils are insulated with Used for coil ratings from 13.8kVrms to 26.0 kVrms
mica paper tape and then vacuum-pressure-impregnated with Operating Voltage Stress Capabilities--20% Above
the resin. Stator coil design, fabrication, and testing has Present Design
generally followed standard practices for many years. Coil Breakdown Voltage( 1 min)--2500 vpm
reliability and long life are major requirements that must be Voltage Endurance--IEEE 1043-1996 @120"C.
met when considering any improvements in the insulation Power Factor Tip-Up--5 1% @, 25OC. (.8U-.2U). Before
system. Coil operation at higher temperatures and higher and after Thermal Cycling at 155OC.
voltage stresses can be shown to have a significant effect on Hot Tan Delta--<lO% @ 0.2 V@15S°C.
machine size and cost. Operating Temperature Class-- 155OC.
Cured Insulation Tensile Strength--16OOOpsi @lOO°C.
Groundwall tapes with improved dielectric properties allow a
reduction in coil groundwall thickness which improves the Compatible with VPI Resin and Processing
thermal performance which in turn allows stator winding U = Coil Design Voltage
design of higher power density. Material costs are a factor
because the material used in the fabrication of coils must meet

0-7803-4927-X/98/$10.000 1998 IEEE

COMPOSITION OF MICA TAPE Even though the mechanical properties of the SM160 and
Improved dielectric performance of high voltage generators QM160 papers are superior to the other two, and even though
has been achieved by developing tapes with improved there is evidence the higher aspect ratio, is beneficial for a
penetrability that allows the uniform resin impregnation of higher modulus of elasticity, their large particle size and in-
thicker insulation used on high voltage coils. A side feature is homogeneous distribution on the paper make them less suitable
the possible use of shorter manufacturing cycles especially for for dry porous tape, where conformability is of great
the impregnation step. Improved impregnation also improves importance. The TM160 paper exhibits physical properties
the dielectric, thermal, and mechanical characteristics of the which are substantially different from non calcined
groundwall insulation. Muscovite. The structure of the Muscovite scraps is altered
after calcination, This renders the porosity and penetration
Five different muscovite .mica papers were considered for this time values much higher than non-calcined papers.
evaluation. The physical properties of the five mica papers are
given in Table 2. The -physical
- structure of the candidate mica The particular mica paper selected is based on the required
:hen in Figure I. insulation requirements- (1). The MM 160 mica aper was
selected for the evaluation reported on here.

Conveyor side Upper side Conveyor side Upper side

QM 160 QM 160 MM160 M M 160

Figure 1. Electromicrographes for two of the Mica Papers.

From left to right, the magnification is XSO, X100, X50, The smaller particle size and narrow particle size distribution
XSO. result in low stiffness values, low porosity time, and short
penetration times which allows a more rapid penetration of the
The SM160 is a paper with a large particle size distribution impregnating epoxy resin. The MM160 mica paper was used
and has large platelets. The QM160 paper is also a large size in the groundwall tape study presented in this paper. The
platelet mica paper, but the particle size distribution is epoxy bonding resins were carefully selected to obtain
narrower than in the SM160 paper. The AM160 paper has optimum compatibility with the impregnating resin. Five
medium particle size platelets and a wide particle size different tapes were manufactured.
distribution. The MM160 paper has small particles with
narrow particle size distribution. The TM160 paper is made of For the mica papers with accelerators, Zinc or Chromium
calcined Muscovite scraps. catalysts were added in an organic solution. Following the
fabrication of the mica paper, the mica paper was impregnated
with the accelerator. The mica paper was then dried at room
temperature and then sprayed with a bonding resin used to
secure the mica paper and glass backing. A sprinkling system
was used to bond the mica paper to the glass backing with the
objective to minimize the amount of epoxy binder. The tapes
using the MM160 mica paper; physical properties are given in
Table 3.

The Cogebi trial tapes were applied to aluminum test bars at
Table 3. MM160 Mica PaDer TaDe Characteristics the same insulation thickness and all samples were processed
and tested in the same exact manner. Following the processing
of the test bars, each test bar is power factor tested at room
temperature. The power factor tip-up (.8U-.2U) was calculated
for each bar and the average of 6 samples for each tape type is
shown plotted in Figure 2. Tape W is included as a control
tape and is a commercially available GBMP tape. All the taped
test bar samples exhibited low power factor tip-up values for
the initial values.
lnltlal Power Factor Tip-Up

A 0
e -0.05
The Cogebi tapes were used to insulate ?4inch by 2 inch by 48 -
Q) -0.1
inch long aluminum test bars insulated to 13.8 kV using a n
double wall thickness (DWT) of 236 mils. Groundwall tape is
applied using an automatic taping machine. Typically 10 half
lap layers are applied to each test bar. All test bars are vacuum
pressure impregnated using an epoxy resin. The test bars were AR AW CR B16 RI6 VP
fblly cured using the same cure cycle as would be used for full
Tape identification
size production coils. Ground electrodes and voltage grading
electrodes are applied to each test bar. Normally, a ground Figure 2. Comparison of the initial power factor tip-up of
electrode length of 12 inches is used with 5 inch long voltage test bars taped and processed with Cogebi tapes and one
grading electrodes on each end of the test bar. The test bars are commercially available tape (VP).
then subject to both electrical and mechanical testing and
evaluation. In order to determine if the insulation system de-laminates
after exposure to elevated temperatures, a four cycle thermal
Following the tests performed on the test bars, the data is cycling test was conducted. Each test bar was heated at 155OC;
compared to data obtained from similar testing performed on unrestrained, for 48 hours and allowed to return to room
other mica tapes. The best performing tapes are selected for temperature. After the test bars cooled to room temperature,
evaluation on full size stator coils. power factor measurements were performed. The plotted
power factor tip-up after the 4* thermal cycle is given in
TEST RESULTS FOR MICA TAPES Figure 3. Low tip-up values indicate minimum groundwall
Electrical tests and limited mechanical tests were conducted on separation has occurred based on the results of the thermal
the processed test bars. The electrical test results were cycle.
evaluated and had to meet the requirements as specified in
(JU-.PU)Power Factor Tip-Up
Table 1. If the tape met this set of electrical test criteria, then
the limited mechanical tests were performed. The experimental
test results are given for the Cogebi tapes used in the


The Cogebi tapes evaluated in this program are given in Table
Tape identification
Table 4. Cogebi Tapes
0 (AR) 609ASR16-24 with Zinc Naphtenat Figure 3. Power Factor Tip-Up after qfh Thermal Cycle.
0 (AW)Normal Cogebi Commercial Tape Plotted Value is average for 6 Test Bars in each group.
0 (CR) 609CAFU6-24 with Chromium Accelerator
0 (B 16)609B16-24 no Accelerator/Same bonding resin as At the conclusion of each of the 4 thermal cycles, the 155"C.,
AW tape power factor was measured at a voltage level of 2.8kVrms.
0 @16)609R16-24 no Accelerator The values are shown plotted in Figure 4. The hot power factor
is an indication of the dielectric loss of the insulation system
when operated at elevated temperature and is desired to be
below 10% at the test temperature.

Tensile Strengthof Test Bar lnsulatlon
155 deg. C Hot Power Factor for 4 Thermal Plotted Value Is Average of 6 Test Bars
Cycles, Measured at 2.8 kVrms
16ooo . ~ - ~ ~ ~ - ”
14013 1
L 4

za 3
c 2

AR AW CR B16 R16 VP v7L AR AW CR 616 RI6 VP

Tape Identification Tape Identification

Figure 4. High Temperature Power Factor measured at Figure 6. Tensile Strength of Test Bar Insulation.
155°C and 2.8 kVrms. The plotted values are the average
of 6 test bar values. The 4 thermal cycle Tan Delta values Percent Uncured Resin
are plotted in the order the thermal cycles were done. The PlottedValue Is Average of 6 Test Bars
numbered bar is the final Tan Delta value after 4 cycles. 0.49
$, 0.45
-E 0.4
$ 0.35

Test bar voltage breakdown tests were conducted under oil, in F5 0.3
order to obtain a true breakdown value and to reduce the c
0 0.2

possibility of voltage flashover. The plotted value, in Figure 5 , 2C

is the actual hold voltage. In units of volts per mil (VPM) the 8 0.1

test bar will hold prior to a breakdown voltage. The hold value 2 0.05

is less than or equal to the actual failure voltage of the test bar. 0
AR AW CR 816 R16 VP

Tape identification

Voltage Hold Value(sv8 of 6 bars) Figure 7. Percent Extract of Uncured Resin.

It has been demonstrated that improved dielectric
characteristics of glass backed mica paper tapes, when used in
a VPI process, is obtained through modification of the tapes
construction and chemistry. The tapes that exhibited the
highest dielectric strength (Figure 5 ) are the AR and the CR
tapes which use a Zinc Naphenat and a Chromiun accelerator
respectively. The accelerators are added to the MM160 mica
paper in these samples. Improved curing and higher strength
groundwalls are also obtained from the tapes that use the
accelerators (Figures 68~7).However, it has been found that
AR AW CR EIS RI6 VP VZ when using the accelerators, a higher elevated temperature
Tape Identification power factor loss is experienced (Figure 4). The higher power
Figure 5. Voltage Hold Values in VPM for the Cogebi factor at 155‘32, for sample CR, is at half the value where a
Tapes design concern would occur.

Dogbone shaped samples were cut from each test bar. REFERENCES
Groundwall tensile strength (see Figure 6) and percent of (1) Kenzo kadotani, “Effect of Mica Paper on Properties of
uncured resin (see Figure 7) were measured. The tensile Coil Insulation”, Electrical Engineering in Japan, Vol. 102, No
strength is a measure of how well bonded the insulation system 6, 1982.
is and the percent extractables is a measure of the remaining
uncured resin in the insulation system after coil cure.