Stator Rewind Solves Corona Problem

Download as pdf or txt
Download as pdf or txt
You are on page 1of 5

GENERATOR OVERHAUL

Stator rewind
solves corona
problem

om Hundley, plant manager,


Elk Hills 35R Cogen, located
near Bakersfield, Calif, said
he knew the day was coming
when a new generator, or a rewind,
would be required on his facilitys
two LM2500-powered gas-turbine
packages. He listened over the years
to experts who participated in the
Western Turbine User Group Incs
conferences (visit www.wtui.com)
and they all essentially had the same
message: Be prepared for major work
or replacement after your air-cooled
generator passes its tenth birthday.

The two units Hundley is responsible for were installed in 1994 and,
as summer 2006 approached, the
annual borescope inspection on Unit
2 revealed some deposits of white
powder on stator coils characteristic
of those associated with corona (Sidebar 1). There was sufficient evidence
to warrant pulling the rotor for a
closer look.
First a little background: The Elk
Hills cogen facility is operated by
Production Services Network, a relatively new firm based in Aberdeen,
Scotland, that provides O&M servic-

es to the oil and gas industry. Hundley, who is employed by Production


Services, has worked in the Elk Hills
O&G field for nearly 30 years and
has managed the plant since 2001.
He was not affiliated with the plant
when it was built in 1994.
National Mechanical Services
(NMS), National City, Calif, which
specializes in generator work, was
hired to pull the rotor and assess
the condition of the stator windings.
An independent consultant also was
engaged. NMS President Bill da
Silva, well known in the land-based

1. What is corona?

ill Moore, PE, National Electric


Coils director of development,
is well known by owner/operators
for his presentations at user-group
meetings while the companys manager of engineering. Heres how
Moore describes corona, plus his
thoughts on prevention:
Partial discharge, sometimes
called corona, is quite common in
air-cooled generators. It is caused by
a partial voltage breakdown within
the generator coil insulation, in gaps
between the coil and the stator core,
and/or in the end turns when the coils
are in close proximity to each other.
Because it is not a complete
breakdown of the insulation system,

it does not cause a full electrical


ground. Over time, however, these
discharges can eat at the insulation,
causing its deterioration until a full
ground does occur and the unit trips
offline.
Evidence of partial-discharge (PD)
activity often is visible to the eye,
appearing as a white powder on the
surface of the stator winding (left).
Severe PD damage on the outside
surface of a stator coil shown below
was attributed to a lack of semi-conductive coating on the surface of the
coil in the slot portion.
Prevention. Special equipment is
needed to detect PD activity. Stator
slot couplers can be inserted under

COMBINED CYCLE JOURNAL, First Quarter 2007

the stator wedge to monitor magnitude and frequency of the discharges.


It is important to trend PD activity
over time because different machines
have different baseline values.
Doubling of PD levels over a
period of six months is cause for
concern. The machine should be
opened up and inspected visually.
Special corona suppression treatments can be applied to coil surfaces
to minimize some types of PD activity.
Photo, right, shows special cell- and
corner-section corona suppression
treatment for a stator coil. Other
types of shielding arrangements can
be used as well, especially for phaseto-phase discharges.

67

GENERATOR OVERHAUL

13

14

15

19

20

21

Profiling a generator
overhaul/rewind
project in the field
and in the shop

ational Mechanical Services


personnel eased the Elk Hills
rotor out of the generator stator (1), bracing and blocking it in
critical locations to prevent damage.
Next, crane straps were installed (2)
and the rotor positioned so the lifting rig could be attached (3). The
rotor then was moved to the flatbed,
packed in protective covers, and
secured for the trip cross-country
to National Electric Coils shops in
Columbus, Ohio.
A stator rewind of the unit had
been decided on months earlier when
an annual borescope inspection
68

25

revealed corona, later confirmed by


a rotor-out inspection that showed
discoloration of the ground-wall insulation (4).
First step after the rotor was
removed for the November outage:
Label and document bus connections
and other critical information before
removing the old windings (5). On
this project, bus rings were reused,

but surge rings and blocking were


replaced.
When the rotor arrived in Columbus, its physical condition was
assessed to see if any damage had
occurred in transit. None in evidence
and photographs were taken for a
permanent record. Next, the rotor was
installed in a lathe and run-out checks
made at more than 30 locations (6).
Rings off. Retaining rings were
prepped for removal (7) using a highfrequency heater. Ring measurements
of OD, ID, length, etc, were recorded
and a dye-penetrant examination conducted to detect cracks or flaws in the
forging (8).
Damper or amortisseur winding
is visible under rotor end wedges (9).
Note the brass squares used as buildup in front of the wedge. They had
been soldered onto a copper ring (at

COMBINED CYCLE JOURNAL, First Quarter 2007

GENERATOR OVERHAUL

10

11

12

16

17

18

22

far left in 9) that encircles the rotor end


winding; damper windings normally
are silver-plated aluminum. This design
required especially careful disassembly
of the rotor to prevent damage.
The design of the copper damper
winding used on the Elk Hills rotor is
considered unusual when compared
to generators of a similar age made
by other OEMs. The copper was laid
in layers and joints were soldered.
Note the buckling of the second layer
of copper under the first layers failed
joint (10).
A temporary set of slip rings was
used to induce a current into the rotor
to conduct flux-probe and other tests
(11). In 12, the rotor is fully assembled
with its exciter and ready for highspeed balancing.
Stator rewind. NEC field personnel constructed an enclosure on the

24

23

front end of the generator housing to


provide shelter for technicians and
storage for tools and materials (13).
Removal of the old windings is shown
in 14. Cramped workspace and the
hard epoxy-impregnated slot fillers
made coil removal difficult and timeconsuming (15).
Some of the coils are shown ready
for the scrap heap in 16. Dark sections visible on a few coils illustrate
the effects of corona activity.
Every inch of the core was thoroughly cleaned, then checked and
rechecked to be sure all resin and
debris had been removed (17). The
original surge rings of insulated steel

COMBINED CYCLE JOURNAL, First Quarter 2007

were replaced with fiberglass rings,


which are stronger (18).
Last of the bottom bars is installed
in 19. Wedges are driven into place
to maintain the proper bar-to-slot
contact and tightness (20). At the connection end of the machine, original
bus rings were tested; no repairs were
required. Permanent ties and blocking were installed when photo 21 was
taken; note that leads are prepared for
the brazing of clips.
After turn-to-turn testing for shorts,
clips were brazed to connect the top
and bottom coilstwo at the turbine
end and one on the connection end of
the machine (22). Brazing complete,
clips await installation of protective
caps (23). El CID test confirmed that
core iron had not been damaged during coil installation (24). Protective caps
were installed and unit painted (25).
69

GENERATOR OVERHAUL
aero community, recalled that the
damage suggested a rewindor
a new generatoras soon as the
plants operating schedule could
accommodate the work.
Hundley said preliminary numbers suggested that a replacement
Brush generator would not have
cost much more than a rewind; however, the long lead time associated
with a new unit was unacceptable.
It would have put the plant at risk
given the condition of the existing
generator.
With the rotor out the following
steps were taken to assure, to the
extent possible, reliable operation
until the stator could be rewound:
1. NMS cleaned up affected areas
of the stator and made temporary
repairs using an insulating epoxy.
2. The consultant recommended
the installation of partial-discharge
analysis (PDA) instrumentation to
trend PD activity. The instrumentation was installed by Hampton
Tedder Technical Services, Montclair, Calif. The thought here was

2. Inside an aircooled generator

that if the condition of the stator


did not deteriorate, the unit would
continue to operate until the November planned outage. If PD activity
increased to the point that failure
was considered imminent, the plant
would take a forced outage to rewind.
To learn more about partial discharge and predictive diagnostics,
visit www.partial-discharge.com.
3. National Electric Coil, which
committed to rewinding the stator
and conducting a full shop inspection of the rotor within the plants
demanding outage window, was
brought in to survey the generator
and take the samples necessary to
plan the project and make replacement coils (Sidebar 2).
Assessment and repairs complete,
NMS reinstalled the rotor and the
plant successfully restarted the unit.
Hundley had only good things to say
about the abilities of NMS and its
Project Manager Rudy Gaitan: The
rotor was pulled in about a day and
a half and replaced within three;
no problems encountered. NMS da

20

Silva attributes that performance to


an experienced in-house staff.
NECs Gary Barton, project manageroperations, said he had two
immediate tasks: (1) Get the data
gathered during the Elk Hills site
visit to engineering so work could
begin on the design of new coils,
and (2) develop a plan and tooling to
remove the resin-saturated Dacron
felt filler used to secure the original
coils in the slots in the iron core.
Barton said the 28-day window
for the entire project was made
more challenging by the resin-saturated felt, which is very difficult to
remove. A hiccup at this stage could
adversely affect schedule. Normally
Barton would add a few days to the
outage just in case, but this was
not possible for the Elk Hills project. Also, NEC had to be prepared
to deal with a forced outage, should
corona activity escalate to a critical
level. This demanded an accelerated schedule for coil design and
manufacture.
In addition to a tight sched-

21
1

21

21
20

ey to the long-term health of any


14
piece of equipment is following
15
1
2
9
to the letter the inspection and maintenance instructions provided by the
15
8
manufacturer. Regarding the main
19
11
18
stator windings (1 in the diagram),
8
12
9
which convert rotor power into volts
4
17
2
6
and amps, your primary goal is to
7
5
keep them clean, says Derek King,
11 3
1
22
10
4
who manages Brush Turbogenerators Inc, Houston.
16
23
14
Visually inspect for dust, oil, for13
eign-object damage (FOD), etc, at
24
least annually. Continuous monitor22
ing of winding temperature allows
1. Main stator windings
2.
Stator
core
tracking of trends that would indicate
18
3. Exciter permanent
the presence of crud and the need
16
magnet generator
for more frequent attention. If you
4. Exciter field
24
find crud, use a vacuum, brushes,
5. Exciter armature
13.
End
frame
6.
Exciter
fan
19. Drive end/coupling
and/or lint-free cloth to clean the
14. Winding supports
7. Diode wheel
20. Air inlet
windings. For stubborn deposits,
15. Fan shroud
8. Rotor and body wedges
21. Air outlet
wipe sparingly with orange oil or
16. Mounting feet/
9. Retaining end caps
22. Access doors
hold-down bolts
23. Access to heater and
equivalent. Take care to avoid forc10. Exciter end of machine
17. Fans
back of stator
ing contaminants into the insulations 11. Oil seals
24. Sole plates
12. Main bearing (exciter end) 18. Stator casing and frame
crevices.
Suggested annual electrical tests
include measuring (a) the polarizaStator core is the magnetic
Annually, or more frequently,
tion index to determine the extent
heart of the stator (2). Air-inlet and
visually inspect for contamination
of moisture and dirt contamination,
outlet temperatures are monitored
or signs of distresssuch as disand (b) winding resistance to detect
to ensure against damage from an
coloration, FOD, etc. Every 100,000
the presence of broken conducover-temperature condition caused
hours Brush recommends an El CID
tors. Also, conduct a reduced hi-pot
by a blockage of cooling-air flow.
(electromagnetic core imperfection
(short for high-potential) test to
Check the instruction book for your
detection) test to ensure that there
insure insulation integrity.
machines temperature limits.
are no shorted laminations.
70

COMBINED CYCLE JOURNAL, First Quarter 2007

GENERATOR OVERHAUL
ule, NEC had to remain flexible on
the start date, as did subcontractor NMS. In fact the project started
about a week later than originally
planned because of delays in completing annual preventive maintenance on Unit 1.
On the second day of the planned
outage, NMS had the generator rotor
secured on the flatbed truck and
ready for the four-day trip to Columbus. The NEC repair shop would
have two weeks to complete its ringsoff inspection and address any issues
identified before shipping it back to
the site.
Once the rotor was out of the
31,176-kVA, 13.8-kV machine, NEC
technicians began removing the old
windings from the 42 stator slots.
They were aided by a new tool,
designed by the NEC engineering
department, to facilitate removal of
the difficult slot filler material in a
cramped 3-ft-diam bore. To be sure
critical field operations proceeded
smoothly and without interruption,
Barton acted as a proactive liaison
between the technicians and the
corporate engineering group when
questions arose. Conference calls and
relay of job-progress photos by e-mail
were frequent.
Onsite personnel were challenged by having to do a perfect job

in resin removal, this to prevent core


hot spots and ensure proper fit-up of
the new coils. Material removed from
the slots, including the old coils, was
scrapped.
The new windings were manufactured well in advance of the outage,
to ensure their availability had there
been a forced outage requiring earlier overhaul of the generator. They
were at the plant before the rewind
specialists arrived. NEC engineers
optimized the design of the new coils
so more copper could be installed
in the slots. Benefits include cooler
operation and a slight increase in
power production.
Also, ripple spring fillers were
used, instead of the conventional
packing originally installed, to hold
the coils in place. The expectation:
They will last for the remainder of
the generators intended service life.
Barton said that felt side packing
often degrades over time, allowing
the coils to move. When this occurs it
can facilitate the onset of corona.
Other generator-related field
work conducted in parallel with the
rewind was cleanup of the generator
bearings by a specialty contractor to
eliminate coking in evidence. Service
included sump cleaning and a new
charge of bearing oil. Such activity is
not typical on this type of project.

While field work proceeded, the


rotor was being serviced in NECs
Columbus shops. The procedure was
standard, one with a proven success
rate. After the rotor was off-loaded
from the flatbed, runout and journals were checked, and following
a thorough visual inspection (with
plenty of photos to record machine
condition), the retaining rings were
removed.
Tests included the standard megger, plus a polarization index or PI,
to measure insulation resistance and
verify coil cleanliness. After testing,
the rotor was cleaned, reassembled,
and balanced at high speed.
Back in the oil patch. The stator rewind was straight-forward
after the old coil was removed; testing verified the electrical integrity of the new field. Hundley said
the machine has been running well
since restart, which went off without
a hitch. He added that the machines
output actually has increased by
approximately 400 kW with the new
stator. Hundley doesnt seem like
he could be happier with the performance of NEC and NMStrue professionals, he said. The accompanying outage photo album presents a
visual record of the project. A rewind
of Unit 1 at the plant is scheduled
for the fall. ccj

Custom reprints
D
INTE
REPR

FRO

Did you read an article in the


COMBINED CYCLE Journal
that youd like your customers and
prospects to have? Use reprints
to help raise the visibility of your
products and services among
buying influences at the plant level
and headquarters locations

ate
estimring
r
e
d
nt
t un
f sto
Don ands o quipme
dem rplant e rm
e
te
pow long-

Call Susie Carahalios


for an estimate
303-697-5009
[email protected]
04.

c, 20

In
edia

o
04. C
ll 20

er
ith p

w
nted

ri
Rep

sio
mis

e CO
m th
n fro

YCLE

ED C

MBIN

SI M
ht P
pyrig

Fa
nal,
Jour

COMBINED CYCLE JOURNAL, First Quarter 2007

71

You might also like