CA0000009

Improving the Reliability of Stator Insulation System in Rotating Machines

Bal K. Gupta*, Howard G. Sedding, and Ian M. Culbert
Ontario Hydro, Toronto, Canada

•Address: 800 Kipling Avenue, Rm KR-151 1. INTRODUCTION

Toronto, Ontario, Canada, M8Z 5S4.
Rotating machines are essential to the operation of
Abstract: Reliable performance of rotating a nuclear generating station. Premature failure of
machines, especially generators and primary heat a generator or an essential motor can result in large
transport pump motors, is critical to the efficient financial losses to a utility. Also some of the
operation of nuclear stations. A significant machines, eg, standby generators and emergency
number of premature machine failures have been power generators are mandatory for safe operation
attributed to the stator insulation problems. of nuclear stations. The present competitive
Ontario Hydro has attempted to assure the long utility environment makes it even more imperative
term reliability of the insulation system in critical to have the equipment operating reliably without
rotating machines through proper specifications any premature failures and the consequent forced
and quality assurance tests for new machines and outages which may severely reduce revenue.
periodic on-line and off-line diagnostic tests on Almost a third of rotating machine failures are
machines in service. The experience gained over caused by a breakdown of stator insulation []].
the last twenty years is presented in this paper. Hence taking steps to improve the reliability of
the stator insulation system in all important
Functional specifications have been developed for machines in a nuclear station will pay good
the insulation system in critical rotating machines economic dividends. Ontario Hydro has had its
based on engineering considerations and our past share of premature failures of important machines
experience. These specifications include in its nuclear stations and has taken steps to
insulation stress, insulation resistance and improve the reliability of stator insulation systems
polarization index, partial discharge levels, in important rotating machines, ie, turbine
dissipation factor and tip up, AC and DC hipot generators (TG), standby generators, and large
tests. Voltage endurance tests are specified for motors rated 4.1 kV and above.
groundwall insulation system of full size
production coils and bars. For machines with In Ontario Hydro, the process of assuring a long
multi-turn coils, turn insulation strength for fast term reliable service life of the stator insulation
fronted surges is specified and verified through system begins with the proper functional
tests on all coils in the factory and on samples of specification and continues through acceptance
finished coils in the laboratory. tests during and/or after the manufacturing and
commissioning stages. Periodic diagnostic tests
Periodic on-line and off-line diagnostic tests are are also performed through the service life of the
performed to assess the condition of the stator machine as discussed in the following sections. In
insulation system in machines in service. Partial this paper we summarize our experience in this
discharges are measured on-line using several area.
techniques to detect any excessive degradation of
the insulation system in critical machines. Novel
sensors have been developed and installed in 2. VENDOR QUALIFICATION
several machines to facilitate measurements of
partial discharges on operating machines. Several Whether purchasing the stator winding for new
off-line tests are performed either to confirm the generators and motors or for the rewind of existing
problems indicated by the on-line tests or to assess ones, the insulation designs from prospective
the insulation system in machines which can not vendors should be evaluated. This evaluation
be easily tested on-line. Experience with these should include a review of the materials used,
tests, including their capabilities and limitations, winding design (including voltage stresses), and
are presented. manufacturing processes including sample coil
testing. The vendor's capability and experience to
comply with the acceptance criteria, to perform
required tests (eg, voltage endurance and thermal

52
cycling tests for groundwall, surge tests for to 0.2 us range [4]. Although not accepted yet
multiturn coils etc, as described below), and to as a standard, a recently described endurance
assure good quality control during the test for turn insulation [5] should be specified
manufacturing process should be reviewed. Only for multiturn coils of critical machines.
manufacturers whose windings are likely to meet
the specifications consistently should be asked to In some cases, Ontario Hydro has specified
bid on the job. the maximum acceptable void sizes in the
insulation system. The voids can be measured
in the coils dissected after the type tests
3. FUNCTIONAL SPECIFICATION above.
AND ACCEPTANCE TESTS
d. Insulation resistance (IR) and polarization
The functional specification for the machine index (PI) for the coils and/or complete
should be based on the application and the winding: Acceptable values and the voltage to
expected performance of the machine. It should be used for test should be specified. IEEE
contain all the tests and the associated acceptance Standard 43 [6] provides the required
criteria required at various critical stages of the guidance. However, the IR values in this
manufacturing process and/or after the final standard should be treated as bare minimum
assembly of the stator winding. The specifications values; modern insulation systems have much
should include (along other pertinent higher values. This is being addressed in the
specifications) the following important features of next revision of IEEE Standard 43.
the stator winding insulation system design and
quality assurance during manufacture: e. Dissipation factor at low voltage (10 % of the
rated line to ground voltage) and tip up
a. Class of insulation and maximum temperature (difference in dissipation factor at 10 % and
rise: Ontario Hydro often specifies class F full rated line to ground voltage) for the
insulation systems for class B temperature rise winding [7]. For stator windings with modern
to ensure an adequate thermal life of critical insulation systems, the acceptable values are 1
machines. to 2 % for the low voltage dissipation factor
and less than 1 % for the tip up.
b. Requirement (or not) for semi-conducting
coating on the groundwall in the slot section f. Hipot tests for the winding: The voltage
& stress grading treatment at the slot to levels and the type of voltages used to hipot
overhang transition area. Both treatments are the complete winding should be specified [8].
required for machines rated higher than 4.1 One minute ac test at (2E+1) kV is normally
kV. specified for the new windings, where E is the
rated line-to-line voltage of the machine in
c. Various type tests to be performed on coils kV. Whether equivalent DC hipot tests [9] or
randomly selected from the production run: equivalent lightning surges can be used as an
By nature these tests are destructive and can alternative should also be specified.
not be performed on the coils or the winding
to be used. These tests evaluate the adequacy g. Partial discharges: Partial discharge (pd)
of the winding design. If more than one unit activity is an important parameter for judging
of the same winding are ordered, these tests the quality of an insulation system. It detects
can be performed once to prove the design. the imperfections like voids, contamination
etc. However, measuring pd levels
Voltage endurance tests, as described in IEEE quantitatively in a stator winding is not a
Standard 1043 [2] are conducted on at least simple task. The measured results for a
two coils. The coils are expected to survive winding at a specified voltage depend on the
250 hours at 110 C and 2.5 pu voltage (eg, 35 size and type of defects, the location of pd
kV for 13.8 kV coils). For machines to be sites, characteristics of the measuring system,
used in cycling duty, thermal cycling tests [3] and the connections of the winding for the
are specified. For machines with multi-tum measurement. There is no standard at present
coils, the turn insulation strength should be for partial discharges in rotating machines.
determined for complete coils, and it should Working groups are active in IEEE, CIGRE,
exceed 3.5 pu for surges with rise times in 0.1 and IEC in this area. Based on extensive

53
 experience with measurements done on large monitoring of the winding for evidence of
number of new and old machines, Ontario continued "good health", or degradation while in
Hydro specifies pd levels for new machines at service. As the windings may settle during the
the rated line to ground voltage as measured first few months of operation, these tests should
with its own wide band pd measurement preferably be performed about a year after
system. PD measurements are made on commissioning and should include the following.
individual phases of a winding, with the other
two phases connected to ground. For good a. IR and PI tests
machines with modern insulation systems, the b. Off-line PD and capacitance/dissipation factor
pd levels are 100 mV or less. Another utility, tests
Electricite de France, also specifies pd c. On-line PD tests (for machines fitted with PD
acceptance limits. sensors)

New techniques for on-line pd measurements

have been recently developed [10]. However, 5. IN-SERVICE DIAGNOSTIC TESTS
for such measurements the pd sensors have to
be built into the winding. Therefore Once the stator winding is in service, tests should
specifications should include installation of be performed to trend the insulation condition and
any sensors (with necessary details) required to detect any significant degradation. The goal is
for later on-line measurements. to detect a degraded condition before it leads to an
unexpected premature failure and to either correct
h. Quality assurance tests during manufacture: the situation or to remove the machine from
There are some tests which can and should be service before a costly forced outage. Some
performed at various stages of manufacturing researchers have reported methods to estimate the
process. For example hipot tests, turn remaining life of machines [11-13]. However,
insulation tests, partial discharge their results have not been confirmed in other
measurements, hipot tests may be specified studies. A large EPRI study [14-16] concluded that
for each coil to weed out coils with defects (a) estimation of the remaining life was impossible
before insertion in the stator core. This is at present; (b) no single parameter gave consistent
easily specified for coils with vacuum results on the condition of stator insulation
pressure impregnated (VPI) or resin rich systems, and (c) an assessment of the insulation
insulation systems; in this process the coils are condition can be made by measuring multiple
effectively finished before insertion in the parameters and trending the results for the same
stator core. However, for the global VPI machine over time or comparing the results for
insulation system, the coils are not similar machines.
impregnated before installation in the slots.
The hipot test levels (including surge levels Ontario Hydro uses both off-line and on-line tests
for the turn insulation in multiturn coils) for to assess the condition of stator insulation systems.
this type of green coils should be lower, On large turbine generators, partial discharges are
typically 2/3 of those for the finished coils or measured on line periodically. If a large
the complete winding. deleterious change is noticed, the machine is
subjected to off-line tests and visual inspection.
i. Acceptance tests for the complete winding: On other critical machines, like standby
Insulation resistance, polarization index, hipot generators, off-line measurements are made every
tests, dissipation factor, and partial discharges five years or so. If a problem is observed, the
are specified for the whole winding and, if measurements are repeated at shorter intervals to
possible, for individual phases. Also tests assess the rate of degradation.
should be carried out to check the tightness of
wedges in all TGs and SGs before the rotor is Off-line measurements include IR and PI,
installed. capacitance and dissipation factor, and partial
discharges. The measurements are made on
individual phase windings as it provides an
4. POST COMMISSIONING TESTS opportunity to compare three phases. Further,
energizing one phase, with the other two phases
Once the new or rewound stator is installed at site, grounded, produces an electric stress in the
some baseline test data should be obtained to allow endwindings. This stress permits probing for

54
problems in these areas due to contamination or significantly with time. Again, a significant
incorrect design of interphase clearances. The increase in pd over time is a better indicator of
hipot tests are potentially destructive and not degradation than the absolute value of pd
applied every time but only in case of suspected magnitudes. On-line pd measurements are used to
weaknesses to assure survivability in operation. examine if any deleterious change has taken place.
Performance of a hipot test requires that the utility In cases where such measurements indicate a
is prepared for the consequences of a failure. In problem, further investigations using off-line tests
our tests on many machines, no single test and inspections are warranted
provided consistent assessment of the insulation
condition. However, examination of all the Hipot tests are performed on machines in service
parameters, comparison of their values for the to assure their serviceability. They should not be
same machines at different times, and a used as routine tests as they are potentially
comparison between similar machines or different destructive in nature. The voltage used is (1.5E+1)
phases of the same machine were used to assess ac or equivalent DC level. The relationship
the insulation condition. The results for a number between the ac and dc test levels is not completely
of standby generators and emergency power understood [20] and Ontario Hydro prefers the ac
generators were reported in earlier papers [17-18]. test. The ac and dc hipot tests stress the
groundwall insulation only and not the turn
The IR values are a measure of moisture, insulation in multi-turn machines. For multiturn
contamination, or cracks in the stator insulation machines in service, Ontario Hydro does not use
and tell little about the insulation aging or or recommend turn insulation tests using fast rise
degradation. PI values less than 2 are often time surges on complete windings as the detection
indicative of aged insulation. However, for the of a turn insulation fault is quite difficult.
modern insulation systems with IR values in
several GC2 range, the values of PI are not very
relevant. Also for such high IR values, the 6. VISUAL INSPECTION AND
measurement of PI is difficult. Most often the IR MAINTENANCE
values are used only to decide whether high
voltage can be applied or not and not to assess the If a problem is indicated by diagnostic tests, a
insulation condition. visual inspection of the winding is necessary.
Normally one looks for damage by overheating or
The measured values for modern insulation discharges, looseness of the coils in the slots,
systems (epoxy resin systems) in good condition general contamination or cracks in the winding,
are 1 to 2 % for the dissipation factor and less signs of damage to semiconductive coating in the
than 1 % for the tip up. However, the measured slot section and the stress grading coating, and any
values are affected by the stress grading coatings softness or puffiness in the insulation. Wedge tap
applied in the machine. Hence the comparison tests are performed on large machines to examine
with earlier measurements or baseline the tightness of wedges. In recent years Ontario
measurements is more useful than the absolute Hydro has been evaluating the effectiveness of
values of the parameters. Also for some of the robotic equipment to perform tests and inspections
older machines with softer insulation systems, the in turbine generators without removing the rotor.
values can be substantially higher. A negative tip These devices can move along the air gap to
up can be an indicator of serious degradation of perform wedge tap tests, core insulation tests, and
the insulation system [19]. visual inspection with a video camera. Some of
the problems can not be corrected by maintenance
Partial discharge measurements are made both off- and a rewind of the stator is then required. In
line and on-line. On good stators, using the some cases maintenance actions, like cleaning
Ontario Hydro wide-band detector, measured pd and/or drying the winding, tightening the wedges,
magnitudes can be expected to be 100 mV or less. and touching up the stress grading coating, may
Magnitudes above 400 mV indicate serious make the machine serviceable with little cost.
problem. In off-line measurements, discharge
inception voltages (DIV) and discharge extinction
voltages (DEV) are also measured. A low DEV 7. CONCLUSIONS
value indicates poor condition of the insulation
system. Some machines may have high pd levels The reliability of stator insulation systems in large
from the beginning which may not increase rotating machines can be assured and improved by

55
proper specifications for the winding, quality 11. H. Yoshida and K. Umemoto, Insulation
assurance during the manufacturing process, Diagnosis for Rotating Machine Insulation,
acceptance tests in the factory, proper application IEEE Transactions on Electrical Insulation,
of machine, periodic on-line or off-line tests, and El-21,No.6, December 1986, p 1021-1025.
necessary maintenance. 12. K. Kadotani, T.Hakamada and S. Yamatake,
A Proposal for Insulation Diagnosis of 3 kV
Motor Stator Windings, IEEE Transactions on
Electrical Insulation, El-18, No.l, February
REFERENCES: 1983, p 59-63.
13. M. Krecke and R. Goffaux, Attempt at
1. Improved Motors for Utility Applications, Estimating the residual life of the HV
EPRI RP1763-1, Final Report, Report EL- Insulation of AC Rotating Machines, Cigre
4286, Volume 1 and 2, October 1982. 1988, paper 11-12.
2. IEEE Standard 1043-1996, Recommended 14. G.C. Stone, H.G. Sedding, B.A. Lloyd, and
Practice for Voltage Endurance Testing of B.K. Gupta, IEEE Trans, on Energy
Form Wound Bars and Coils. Conversion, EC-3, No.4, December 1988, pp
3. IEEE Standard 1310 -1996, Recommended 833-841.
Practice for Thermal Cycle Testing of Form 15. I.M. Culbert, H. Dhirani, and G.C. Stone,
Wound Stator Bars and Coils for Large Handbook to Assess the Insulation Condition
Generators. of Large Rotating Machines, EPRI EL-5036,
4. IEEE Standard 522-1992, Guide for Testing Volume 16, June 1989.
Turn-to-Tum Insulation on Form-Wound 16. Motor and generator Insulation Life
Stator Coils for AC Rotating Electric Estimation, EPRI Project RP2577-1, Final
Machines. report, Publication EPRI TR-100185,
5. A Proposed Type Test for Interturn Insulation Volumes 1-3, 1992.
in Multi-Turn Coils, B.K. Gupta and W.T. 17. B.K. Gupta and I.M. Culbert, Assessment of
Fink, Paper Presented at 1996 IEEE Insulation Condition in Rotating Machine
International Symposium on Electrical Stators, IEEE Trans, on Energy Conversion,
Insulation, Montreal, June 16-19, 1996 and EC-7, No.3, September 1992, pp 500-508.
published in the Conference Record, IEEE 18. B.K. Gupta, Experience with Off-Line
Publication 96CH3597-2, pp 235-238. Diagnostic Tests on Rotating Machines,
6. IEEE Standard 43-1974, IEEE Recommended CIGRE-EPRI Colloquium on Maintenance
Practice for Testing Insulation Resistance of and Refurbishment of Utility Turbogenerators,
Rotating Machinery. Hydrogenerators, and Large Motors, Florence,
7. IEEE Standard 286-1975, IEEE Italy, April 14-16, 1997.
Recommended Practice for Measurement of 19. R.W. Sillars, Electrical Insulating Materials
Power Factor Tip-Up of Rotating Machinery and Their Applications, Peter Peregrinus (on
Stator Coil Insulation. behalf of IEE), England, p 60. Also see C.G.
8. IEEE Standard 4-1995, IEEE Standard Garton, J.IEE, Vol 88, Pt. 3, 1941, pp 23-40.
Techiniques for High Voltage Testing. 20. B.K. Gupta, Use of AC and DC Hipot Tests to
9. IEEE Standard 95-1977, IEEE Recommended Assess Condition of Stator Insulation,
Practice for Insulation Testing of Large AC Presented at the Electrical/Electronics
Rotating Machinery with High Direct Voltage. Insulation Conference, Chicago, Illinois,
10. H.G. Sedding, S.R. Campbell, G.S. Stone, and September 18-21, 1995, and published in the
G.F. Klempner, A New Sensor for Detecting Proceedings, IEEE Publication 95CH35487,
Partial Discharges in Operating Turbine pp 605-608.
Generators, IEEE Winter Power Meeting,
New York, NY, February 1991, Paper 91 WM
065-3 EC.

56