Professional Documents
Culture Documents
Gearbox Failure Investigation
Gearbox Failure Investigation
CONFIDENTIAL
This document contains business and technical information and data is confidential and proprietary to
Portsmouth, Rhode Island. Any unauthorized publication or disclosure of such information to any person, or
unauthorized use of such information without the express written consent of Portsmouth, Rhode Island is
strictly prohibited.
DAR-‐EN-‐20120926-‐revA
Portsmouth
Gearbox
Failure
Investigation
14
October
2012
ii
DAR-‐EN-‐20120926-‐revA
Portsmouth
Gearbox
Failure
Investigation
14
October
2012
iii
DAR-‐EN-‐20120926-‐revA
Portsmouth
Gearbox
Failure
Investigation
14
October
2012
EXECUTIVE
SUMMARY
The
Portsmouth
wind
turbine
gearbox
has
suffered
a
significant,
premature
failure
of
the
first
and
second
planetary
stages.
The
gearbox
must
be
replaced
in
order
to
return
the
turbine
to
service.
The
failure
occurred
after
3
years
of
operation
despite
a
design
life
of
20
years.
Total
operating
time
on
the
turbine
is
approximately
21,000
hours,
and
total
production
in
accordance
with
the
production
meter
is
9,898
MWh.
Depending
upon
future
gearbox
reliability,
the
cost
of
one
or
more
replacements
may
exceed
the
potential
future
profit
for
the
life
of
the
turbine.
Therefore,
a
clear
understanding
of
the
root
cause
of
failure
and
the
reliability
of
the
turbine
moving
forward
is
critical
to
Portsmouth.
Due
to
the
extent
of
the
damage,
replacement
is
most
likely
preferable
to
repair,
depending
upon
repair
vs.
replacement
cost.
The
root
cause
of
failure
was
not
determined
during
this
investigation.
Several
potential
causes
were
ruled
out
as
follows:
x Structural
damage
to
critical
support
structure
or
looseness
of
components
such
as
blades,
hub,
tower
foundation,
and
generator
x Generator
misalignment
x Blade
exterior
condition
such
as
cracks,
delamination,
or
missing
vortex
generators
x High
frequency
of
yaw
error
exceeding
10
degrees
(nacelle
direction
vs.
prevailing
wind
direction
Possible root causes of failure that cannot be confirmed or ruled out are categorized as follows:
x High:
Gearbox
manufacturing
error
with
respect
to
assembly
quality
and
choice
of
materials
and
specificationsȄcomplete
teardown
inspection
is
required
as
well
as
review
of
design
and
build
records
x High:
Gearbox
design
error
with
respect
to
structural
stiffness
of
housing-‐-‐
analysis
verification
required
x Low:
Wind
turbine
support
structural
stiffnessȄanalysis
verification
required
of
unique
ring
mount
design
x Low:
Pitch
asymmetry
due
to
incorrect
blade
offsets
x Low:
Service
quality,
especially
after
May
2011.
However,
failure
to
identify
failure
in
a
timely
manner
resulted
in
unplanned
failure
and
extensive
gearbox
collateral
internal
damage.
x Low:
SitingȄpossible
wind
shear
from
freeway
cut.
Nearby
47
foot-‐tall
water
tank
is
not
a
likely
contributor.
The
gearbox
configuration
is
not
conventional
by
US
industry
standards.
The
make
and
model
of
gearbox
have
a
poor
track
record,
with
at
least
3
of
5
installed
in
the
US
having
gearbox
failures
1
DAR-‐EN-‐20120926-‐revA
Portsmouth
Gearbox
Failure
Investigation
14
October
2012
within
the
first
3-‐4
years
of
operation.
Supply
of
replacement
gearboxes
is
likely
to
be
limited,
and
the
turbine
is
destined
to
become
an
orphan
due
to
bankruptcy
of
the
manufacturer
(AAER)
and
sluggish
support
from
the
designer
(AMSC/Windtec).
AAER
purchased
a
license
to
build
the
1650
kW
AMSC/Wintec
design
but
built
a
1500
kW
machine
instead.
The
AAER
1500/77
60Hz
wind
turbine
or
any
other
AMSC/Windtec
design
with
the
JaKe
PPSC
1290-‐
A120
gearbox
never
received
a
Statement
of
Conformity
or
Type
Certificate
from
a
world-‐
recognized
certification
organization
such
as
Germanischer-‐Lloyd
or
Det
Norske
Veritas.
Therefore,
the
turbines
did
not
receive
the
level
of
verification
and
third
party
engineering
review
other
certified
wind
turbines.
The
water
tank
to
the
North
may
wake
the
lower
arc
of
the
rotor
creating
asymmetric
rotor
loads,
and
the
cut
for
Highway
24
may
create
increased
vertical
shear
across
the
rotor
creating
asymmetric
rotor
loads.
These
conditions
are
dependent
upon
the
frequency
of
operation
and
wind
speeds
in
the
direction
of
the
features.
The
ATM
site
suitability
report
did
not
provide
a
Wind
Rose
as
part
of
the
wind
resource
assessment,
nor
did
it
mention
effects
of
tower
or
highway.
Several
significant
opportunities
to
identify
premature
gearbox
failure
were
missed
from
as
early
as
June
2011
through
March
2012.
While
early
detection
cannot
prevent
failure,
it
can
provide
an
opportunity
to
delay
the
failure,
plan
for
replacement
and
reduce
collateral
damage
from
a
single
component
failure.
Collateral
damage
can
significantly
increase
the
cost
of
gearbox
repair
or
result
in
the
need
to
purchase
a
new
gearbox.
Recommendations:
1. Perform
financial
feasibility
study
for
gearbox
replacement
options,
including
identical
gearboxes
to
failed
unit,
remediated
but
similar
models,
and
alternative
models
2. Perform
forensic
teardown
of
failed
gearbox
to
narrow
down
root
cause
of
failure.
3. Perform
blade
pitch
calibration
and
determine
blade
offsets
as
a
potential
root
cause
4. Issue
RFQ
for
qualified
service
providers
with
skilled
senior
level
troubleshooters
5. Obtain
clear
understanding
of
relationships
with
JaKe
and
AMSC/Windtec
with
respect
to
orphaned
product
support
and
cost.
6. Inspect
blades
and
determine
repair
costs
prior
to
gearbox
replacement
in
order
to
perform
complete
financial
analysis
for
return
to
service
of
the
turbine
2
DAR-‐EN-‐20120926-‐revA
Portsmouth
Gearbox
Failure
Investigation
14
October
2012
INTRODUCTION
The
town
of
Portsmouth,
Rhode
Island
(Portsmouth)
commissioned
a
new
AAER
1500-‐77-‐65
1.5
Megawatt
wind
turbine
on
March
24,
2009.
On
May
18,
2012,
significant
amounts
of
metal
were
found
in
the
gearbox
oil
filter
housing
and
significant
internal
damage
was
observed
with
a
borescope.
The
filter
element
was
replaced
on
May
25,
and
the
turbine
was
returned
to
service.
The
turbine
was
removed
from
service
on
June
18,
2012
after
significant
additional
metal
was
discovered
in
the
filter
housing.
BACKGROUND
The
AAER
1500-‐77
was
manufactured
by
AAER
Wind
Energy
of
Quebec,
Canada.
AAER
declared
bankruptcy
shortly
after
commissioning,
and
all
service
contracts
and
warranties
through
the
OEM
were
lost.
AAER
built
the
turbines
under
license
from
AMSC
Windtec
,
who
has
licensed
similar
designs
to
several
manufacturers
around
the
world.
According
to
Alexander
Groeber,
Director
of
ǡDzWe
are
not
100%
sure
what
AAER
has
finally
built
Ȃ
they
have
bought
a
ͳͷͲͳͷͲͲǥ ǯt
the
whole
detailed
turbine
configuration.dz
Anecdotal
information
indicates
several
thousand
are
in
operation,
mostly
with
a
different
model
of
gearbox.
In
2009
and
2010,
five
similar
AAER
1500-‐77
wind
turbines
with
identical
gearboxes
were
installed
in
the
United
States1.
Since
then,
at
least
3
of
the
5
turbines
have
experienced
gearbox
failures
(one
is
operating
at
1000kW
peak
output
due
to
grid
limitations).
The
design
lifetime
of
the
turbine
is
20
years,
including
the
PPSC
1290-‐A120
gearbox
manufactured
by
Jahnel
Kestermann
(JaKe)
of
Germany.
In
order
to
return
the
wind
turbine
to
service,
the
gearbox
must
be
replaced
or
repaired
at
significant
cost
to
Portsmouth.
The
gearbox
warranty
was
24
months
[from
delivery?],
and
the
failure
was
discovered
at
approximately
38
months.
The
gearbox
manufacturer
has
offered
replacement
gearboxes
at
a
significantly
discounted
price
while
supplies
last.
DA
Roberts,
LLC
(DAR)
performed
a
site
inspection
and
obtained
documents
and
data
for
review
on
August
28
and
29,
2012
to
confirm
the
condition
of
the
gearbox
and
to
look
for
root
causes
of
failure.
VISUAL
INSPECTION
The
purpose
of
the
visual
inspection
was
to
become
familiar
with
the
installation
configuration,
review
blade
pitch
settings,
and
identify
gear
oil
leaks,
loose
hardware,
or
other
indications
as
possible
factors
in
gearbox
failure.
GEARBOX INSPECTION
1
Barstow,
CA
(1,
operating
at
1000kW
peak
output
due
to
grid
limitations),
Tehachapi,
CA
(2),
Templeton,
MA (1)
3
DAR-‐EN-‐20120926-‐revA
Portsmouth
Gearbox
Failure
Investigation
14
October
2012
The
purpose
of
the
borescope
inspection
was
to
inspect
the
internal
condition
of
the
gearbox
and
obtain
representative
photographs
of
each
gear
and
bearing
that
is
accessible.
REPORTING
Upon
return
from
site,
a
draft
letter
report
documenting
findings
and
providing
recommendations
was
expected
to
be
submitted
within
two
weeks.
FINDINGS
Findings
are
presented
in
accordance
with
the
scope
of
work,
and
attempt
to
narrow
down
potential
root
causes
of
failure.
VISUAL
INSPECTION
The
visual
inspection
focused
on
potential
mechanical
and
environmental
influences
into
to
gearbox
failure.
Installation
Configuration
The
installation
of
the
gearbox
is
structurally
unique
from
t he
more
common
models
of
turbines
in
operation
in
the
United
States,
and
worth
investigation
as
a
possible
root
cause
of
failure.
While
it
is
not
possible
within
the
scope
of
this
effort
to
perform
a
structural
evaluation,
a
search
through
public
documents
for
industry-‐accepted
design
review
has
provided
some
useful
information.
4
DAR-‐EN-‐20120926-‐revA
Portsmouth
Gearbox
Failure
Investigation
14
October
2012
In
the
AAER
1500,
the
rotor
is
cantilevered
from
the
gearbox,
which,
in
turn,
is
mounted
to
the
nacelle
frame
as
shown
in
Figure
1.
The
gearbox
housing
incorporates
an
oil-‐lubricated
main
bearing
within
the
housing.
The
nose
casing
is
mounted
to
a
ring-‐type
mount
through
elastomeric
elements,
and
non-‐torque
rotor
loads
and
the
torque
reaction
loads
are
transferred
to
the
tower
top
through
the
ring-‐shaped
weldment.
Figure
1
-‐
Gearbox
Installation
Configuration
Similar
configurations
have
received
statements
of
conformity
or
been
type
certified
by
both
Germanischer-‐Lloyd
(GL)
and
Det
Norske
Veritas
(DNV)
on
similar
designs.2,3,4
However,
there
is
no
published
record
indicating
the
specific
combination
of
the
AAER
1500/77
with
the
JaKe
PPSC
1290-‐A120
nor
any
other
make
of
the
AMSC/Windtec
design
with
the
JaKe
PPSC
1290-‐A120
has
any
published
record
of
final
GL
conformity
or
any
type
of
final
certification.
DAR
is
concerned
that
the
Portsmouth
configuration
is
now
an
orphan
design,
and
without
undue
expense,
a
rigorous
evaluation
to
determine
conformity
with
widely
accepted
industry
standards
will
not
be
forthcoming.
The
following
configuration
influences
contributing
to
gearbox
failure
cannot
be
ruled
out:
x An
increase
in
rotor
loading
and
mass
as
a
result
of
increasing
the
rotor
diameter
from
70
to
77m,
resulting
in
higher
overhanging
moments
on
the
gearbox
and
potentially
higher
non-‐
torque
loading
into
the
gearbox
housing
as
a
result
of
increased
transient
aerodynamic
imbalance.
x A
change
of
gearbox
internal
tooth
and
bearing
loading
and
life
expectancy
due
to
a
ratio
change
to
accommodate
60
Hz
from
the
more
common
50
Hz
configuration
2
AMSC
WT
1650
60
Hz,
GL
Statement
of
Conformity,
DAA-‐GL-‐025-‐2010,
15
September
2009
3
Hyundai
HQ
1650,
DNV,
Type
Certificate,
IEC
TC222301-‐0,
23
September
2011
4
Fuhrlander
FL
1500/77,
GL
(1999),
Statement
of
Conformity
WT-‐00-‐022A-‐2004,
July
14,
2004
5
DAR-‐EN-‐20120926-‐revA
Portsmouth
Gearbox
Failure
Investigation
14
October
2012
Additional
insight
into
the
gearbox
design
process
was
obtained
by
Portsmouth
in
an
interview
with
Martin
Venne,
a
former
Technical
Director
for
AAER
and
responsible
for
the
due
diligence
on
the
1290
gearbox.
At
the
time
Portsmouth
was
procuring
the
turbine,
he
was
in
Germany
at
the
JaKe
factory.
He
was
not
convinced
JaKe
had
solved
the
bending
problem
in
their
gearbox
to
his
satisfaction.
He
was
aware
of
two
1281
(first
generation)
gearboxes
that
had
failed
within
one
year
which
prompted
JaKe
to
reinforce
the
second
generation
(1290)
and
provide
finite
element
analysis
data
to
demonstrate
they
had
solved
the
problem.
He
was
skeptical
that
they
had
made
the
case,
called
into
question
some
of
the
assumptions
in
their
study,
and
quite
pointedly
stated
his
conviction
that
the
root
cause
of
our
gearbox
failure
was
lack
of
reinforcement
in
the
gearbox
to
accommodate
high
bending
loads.
No
other
significant
configuration
differences
in
the
Portsmouth
turbine
were
noted
that
could
potentially
contribute
to
gearbox
failure.
Asymmetric
load
can
impart
bending
moments
into
the
low
speed
shaft,
and
in
the
case
of
the
PPSC
1290
design,
into
the
gearbox.
Shaft
bending
can
result
in
uneven
load
distribution
on
bearings
and
gear
teeth
in
the
1st
planetary
stage
of
the
gearbox
as
a
result
of
bearing
clearances,
and
may
cause
structural
deflections
that
also
affect
bearing
and
gear
load
sharing.
While
verification
of
design
loading
against
operating
loads
is
well
beyond
the
scope
of
this
investigation,
one
potential
cause
for
constant
asymmetrical
loading
is
incorrect
blade
indexing.
When
the
blades
are
initially
installed,
the
pitch
system
must
be
calibrated
by
measuring
blade
offsets
from
zero
using
pitch
index
plates
on
the
blades
and
entering
the
offsets
into
the
controller.
A
visual
check
of
the
current
pitch
settings
was
made
and
recorded
as
shown
in
Table
1.
From
the
data,
it
appears
there
may
be
a
pitch
imbalance;
however,
upon
review
of
commissioning
documents,
a
very
specific
procedure
is
outlined
in
section
110.2
(page
36)
of
the
Installation
Commissioning
document
provided
to
Portsmouth
during
commissioning.
DAR
recommends
Portsmouth
authorize
and
witness
an
experienced
service
team
to
execute
the
pitch
calibration
procedure
per
the
instructions
and
note
the
new
offset
values.
Allowable
offset
variances
should
be
obtained
from
AMSC/Windtec.
As
a
result
of
the
pitch
inspection,
pitch
asymmetry
due
to
pitch
differences
between
b lades
cannot
be
ruled
out
as
a
potential
influence
into
gearbox
failure.
Blade Commissioning Blade Index check 29 Aug 12 Pitch difference Blade position
Offset April 29 Aug 29 Aug
2008
1 -‐1.6 .25 left (CW rotation) of mark -‐ Up left looking
s/n 156 forward
6
DAR-‐EN-‐20120926-‐revA
Portsmouth
Gearbox
Failure
Investigation
14
October
2012
Figure
2
-‐
Blade
Vortex
Generators
Mechanical
Looseness
Inspection
of
the
following
hardware
indicated
periodic
partial
torque
tests
were
performed,
and
no
loose
or
missing
hardware
that
could
induce
abnormal
vibration
into
the
drivetrain
was
found.
7
DAR-‐EN-‐20120926-‐revA
Portsmouth
Gearbox
Failure
Investigation
14
October
2012
Figure
3
-‐
Low
Speed
Shaft
Spark
Gap
(one
of
two
shown)
8
DAR-‐EN-‐20120926-‐revA
Portsmouth
Gearbox
Failure
Investigation
14
October
2012
Figure
4
-‐
GasTOPS
Wear
Particle
Monitor
Vibration
S witch
The
vibration
switch
is
installed
in
the
turbine
protection
system
to
detect
large
vibrations
in
the
nacelle
that
could
lead
to
significant
damage
to
the
turbine.
Due
to
the
acceptable
visual
condition
and
positioning
of
the
switch,
undetected,
extremely
high
vibration
leading
to
gearbox
failure
is
ruled
out
as
a
potential
influence
in
gearbox
failure.
Figure
5
-‐
Vibration
Switch
Turbine
Siting
Most
wind
farms
are
installed
subsequent
to
a
wind
resource
assessment
and
site
suitability
analysis.
Site
selection
was
based
upon
a
report
titled
Town
of
Portsmouth,
RI
Wind
Resource
and
Economic
Assessment
prepared
in
2007
by
Advanced
Technology
and
Management
(ATM).
The
report
was
summarized
by
the
Portsmouth
Economic
Development
Council
(PEDC)
report
Wind
Project
for
the
High
School
and
Middle
School
-‐
A
Feasibility
Study.
According
to
the
Portsmouth
Economic
Development
Council
(PEDC),
ATM
provided
Dza
detailed
wind
resource
analysis,
a
detailed
energy
use
analysis,
an
electrical
interconnection
assessment,
and
a
financial
analysisdz
the
final
site
(the
High
School).
ǡDzWithin
the
scope
of
9
DAR-‐EN-‐20120926-‐revA
Portsmouth
Gearbox
Failure
Investigation
14
October
2012
Figure 7 shows the physical location of the turbine relative to its surrounds.
Figure
6
-‐
Site
of
Portsmouth
Turbine
While
they
were
not
considered
fatal
flaws
for
a
site
evaluation,
two
noticeable
features
stand
out
that
cannot
be
ruled
out
as
potential
causes
of
wind
shear
and
subsequent
asymmetrical
rotor
loading.
A
water
tower,
shown
in
Figure
7,
is
located
within
600
feet
North
of
the
turbine
base.
The
lower
arc
of
the
rotor
is
27.5
meters
(90
ft.)
above
the
ground.
The
height
of
the
tower
is
46
feet.
Tower
waking
can
most
likely
be
ruled
out
as
a
contributing
factor
to
gearbox
failure.
Figure
7
-‐
Water
Tank
Near
Turbine
A
freeway
cut
with
a
steep
embankment
is
located
approximately
600
ft
to
the
Northwest
of
the
turbine.
The
cut
may
induce
wind
shear
into
the
rotor
during
inflow
angles
perpendicular
to
the
cut.
10
DAR-‐EN-‐20120926-‐revA
Portsmouth
Gearbox
Failure
Investigation
14
October
2012
Figure
8
-‐
Highway
24
cut
Northwest
of
Turbine
The
ATM
report
does
not
provide
a
wind
direction
frequency
(Wind
Rose).
In
the
course
of
future
wind
turbine
operation,
additional
wind
resource
assessment
of
the
site
should
be
performed
to
investigate
shear
influences
on
the
turbine,
and
to
determine
the
need
for
curtailment
in
the
event
the
geography
adversely
affects
rotor
loading.
GEARBOX
INSPECTION
A
detailed
gearbox
inspection
report
is
provided
in
Appendix
A .
This
section
summarizes
the
findings.
Figure
9
-‐
Gearbox
Data
Plate
The
gearbox
is
a
three-‐stage
planetary
design
with
two
planetary
and
one
parallel
stage
which
is
offset
to
the
right
looking
towards
the
rotor.
In
total,
there
are
twelve
gears
and
twenty-‐two
bearings
producing
an
increase
ratio
of
120:1
to
convert
the
17
RPM
speed
of
the
rotor
to
the
nominal
generator
speed
of
2000
RPM.
The
oil
capacity
of
the
gearbox
is
685
liters
(181
gallon).
Figure
10
and
Figure
11
show
the
internal
arrangements.
11
DAR-‐EN-‐20120926-‐revA
Portsmouth
Gearbox
Failure
Investigation
14
October
2012
Figure
10
-‐
Gearbox
Internal
Configuration
Figure
11
-‐
Gearbox
Section
Drawing
External
Inspection
The
external
inspection
did
not
reveal
any
damaged,
loose
or
missing
structural
components.
Oil
leaks
were
noted
at
the
output
shaft
seal,
and
oil
residue
is
evident
as
far
away
as
the
yaw
deck.
The
seal
is
partially
dislodged
from
the
shaft
as
shown
in
Figure
12.
According
to
the
service
technician,
the
seals
slip
out
of
their
seats
despite
repeated
reseating.
12
DAR-‐EN-‐20120926-‐revA
Portsmouth
Gearbox
Failure
Investigation
14
October
2012
Figure
12
-‐
Output
Shaft
Seal
The
gearbox
housing
has
a
radial
pattern
of
unused
holes.
This
appears
to
be
by
design,
and
perhaps
the
holes
are
used
in
a
different
installation
configuration,
such
as
a
different
turbine
make.
Examination
of
the
mounting
configuration
revealed
the
gearbox
is
held
in
the
housing
using
a
clamping
arrangement
rather
than
the
through-‐holes
on
the
casing.
The
flange
appears
to
be
solid,
and
there
is
no
possibility
of
fretting
or
breathing
of
the
structure
within
the
flange
that
would
be
associated
with
mating
flanges.
Figure
13
-‐
Gearbox
Housing
Flange
Holes
Internal
Inspection
The
gearbox
internal
components
that
could
be
inspected
without
draining
oil
were
viewed
with
a
GE
XLGo+
borescope
with
a
6.1mm
insertion
tube
with
no
lens
or
a
120
deg,
standard
focus
tip
for
close-‐in
imaging.
Inspection
access
was
limited
to
one
access
cover
on
top
of
the
second
planetary
stage,
and
the
desiccant
breather
fitting
on
the
nose
casing.
The
inspection
revealed
severe
abrasion
and
deformation
of
the
first
stage
planetary
ring
gear
and
severe
Hertzian
fatigue
and
debris
damage
to
second
stage
planetary
ring
gear.
The
ring
gears
are
the
largest
gears
in
the
gearbox.
Failure
of
either
gear
requires
removal
of
the
gearbox
from
the
turbine.
It
is
possible
the
damage
observed
is
a
secondary
failure
mode,
with
the
primary
failure
13
DAR-‐EN-‐20120926-‐revA
Portsmouth
Gearbox
Failure
Investigation
14
October
2012
mode
a
planet
bearing
failure.
The
appearance
of
the
damage
on
the
ring
gear
indicates
extreme
structural
deformation
of
the
housing
due
to
overload,
or
cocking
of
one
or
more
planet
gears
after
complete
failure
of
a
planet
bearing
or
bearings.
Figure
14
-‐
First
Planetary
Stage
Ring
Gear
Damage
The
second
stage
ring
gear
appears
to
have
Hertzian
fatigue
damage
as
indicated
by
macropitting
and
spalling
in
the
lower
portion
of
the
tooth
flank.
The
teeth
also
show
significant
debris
damage
from
hard
particles
entrained
in
the
gear
mesh.
Figure
15
-‐
Second
Planetary
Stage
Ring
Gear
Damage
The
gearbox
is
not
serviceable
in
this
condition,
and
further
operation
of
the
turbine
is
not
recommended
due
to
the
certainty
of
additional
d amage
to
the
gearbox
and
risk
of
gearbox
housing
structural
failure.
14
DAR-‐EN-‐20120926-‐revA
Portsmouth
Gearbox
Failure
Investigation
14
October
2012
Typical
causes
of
premature
bearing
failure
in
wind
turbines
are
due
to
poor
lubrication
or
uneven
load
distribution.
A
forensic
teardown
of
the
gearbox
will
provide
a
much
greater
opportunity
to
investigate
the
primary
failure
mode
of
the
gearbox.
Error
List
ǡDz dzǡ indicates
low
oil
pressure
after
the
pump
is
switched
to
high
speed
(for
better
cooling).
This
error
causes
the
turbine
to
shut
down
using
Fast
Stop
mode
(no
mechanical
brake).
Figure
16
-‐
Top
Faults
by
Frequency
A
time
trace
of
fault
666
shown
in
Figure
17
indicates
faults
occurring
as
early
as
October,
but
significantly
increasing
in
February,
2012.
Service
records
and
the
service
provider
(Solaya)
report
titled
Portsmouth
Rhode
Island
Wind
Turbine
Maintenance
Gearbox
Analysis
(undated,
unsigned)
indicate
technicians
suspected
and
searched
for
a
faulty
sensor,
ultimately
finding
significant
metal
in
the
gear
oil
filter
clogging
indicator.
The
filter
housing
was
cleaned,
the
sensor
changed,
and
the
fault
frequency
decreased
in
late
May
2012.
It
peaked
again
in
early
June.
In
all
likelihood,
the
666
fault
increase
was
due
to
large
amounts
of
metal
in
the
gear
oil
pump
and
filter
housing,
resulting
in
a
restricted
supply
pressure
downstream
of
the
pump.
15
DAR-‐EN-‐20120926-‐revA
Portsmouth
Gearbox
Failure
Investigation
14
October
2012
Figure
17
-‐666,
548,
824,
and
536
Fault
History
Figure
16
and
Figure
17
show
a
high
incidence
of
error
824,
which
is
a
yaw
nuisance
fault,
and
has
no
bearing
on
the
gearbox
failure.
However,
it
does
prevent
the
turbine
from
yawing
into
the
prevailing
wind
during
the
extended
downtime
associated
with
the
gearbox
failure.
The
yaw
fault
should
be
corrected
in
order
to
limit
blade
flapwise
loading,
which
could
cause
blade
damage
in
high
winds.
10-‐Minute
SCADA
This
section
is
limited
to
a
review
of
eight
days
of
data
with
at
least
one
hour
of
winds
6
m/s
or
higher.
Key
tags
indicating
gearbox
health
were
reviewed
for
days
operating
at
or
near
rated
power
in
the
provided
data
set
with
wind
speeds
in
excess
of
12
m/s,
were
December
11,
2011
and
March
27,
2012
only
data
to
date,
with
the
following
results:
x Gearbox
bearing
temperature
average
rose
from
69.8
°C
to
81.5°C
x Gearbox
oil
temperature
average
rose
from
69.3
°C
t o
71.6
°C
x Ambient
temperatures
for
December
11
were
15
°C
and
March
27
near
0°C
In
a
trend
analysis
of
eight
days
of
data
with
power
output
greater
than
350
kW,
gearbox
bearing
temperature
was
providing
predictive
data
indicating
gearbox
health
issues
as
early
as
January,
well
in
advance
of
the
severe
failure
finally
noted
in
May
and
June.
16
DAR-‐EN-‐20120926-‐revA
Portsmouth
Gearbox
Failure
Investigation
14
October
2012
Figure
18
-‐
Gearbox
Bearing
&
Oil
Tempterature
Trend
Turbine
operating
conditions
affecting
drivetrain
loads
could
include
yaw
error,
or
the
difference
between
the
nacelle
direction
and
the
prevailing
wind.
A
review
of
several
days
of
data
indicates
yaw
error
rarely
exceeds
6
degrees.
In
addition,
there
is
no
clear
correlation
of
increased
tower
vibration
or
drivetrain
swinging
error
that
correlates
with
yaw
error
in
the
limited
data
set.
It
is
unlikely
that
the
frequency
of
operation
at
yaw
errors
of
less
than
10
degrees
introduce
loads
that
exceed
the
design
load
spectrum.
Logbook
entries
show
repeated
events
of
error
526,
Gearbox
Oil
Filter
at
High
Speed.
This
error
cannot
be
found
in
the
current
error
listing,
but
it
may
relate
to
the
666
errors
observed
throughout
late
2011
and
2012.
If
so,
17
DAR-‐EN-‐20120926-‐revA
Portsmouth
Gearbox
Failure
Investigation
14
October
2012
An
emergency
stop
at
5
m/s
was
conducted
for
training
purposes
on
September
10,
2009.
This
stop
activates
the
high
speed
brake
and
fast
pitches
the
rotor.
At
5
m/s
the
generator
is
rotating
at
approximately
1250
RPM.
The
likelihood
that
this
event
resulted
in
gearbox
damage
is
low.
FINANCIAL
CONSIDERATIONS
In
order
to
determine
the
profitability
of
continued
turbine
operation
subsequent
to
a
gearbox
replacement,
two
scenarios
are
considered,
a
three-‐year
and
a
ten-‐year
gearbox
life.
The
three-‐year
life
estimate
assumes
a
larger
maintenance
reserve
based
upon
current
replacement
costs
divided
by
three
years.
The
ten-‐year
estimate
divides
replacement
costs
by
10
years.
Based
upon
the
high
potential
for
curtailment
of
future
operations
to
protect
the
gearbox
under
potentially
damaging
conditions,
and
to
protect
the
surrounds
from
potential
ice
throw,
an
availability
estimate
of
92%
is
used.
No
additional
major
corrective
events
are
factored
in.
At
current
power
purchase
prices,
annual
revenue
estimate
of
$160,000
is
grossly
insufficient
to
cover
the
costs
of
operation
in
either
case.
18
DAR-‐EN-‐20120926-‐revA
Portsmouth
Gearbox
Failure
Investigation
14
October
2012
19
DAR-‐EN-‐20120926-‐revA
Portsmouth
Gearbox
Failure
Investigation
14
October
2012
OTHER
OBSERVATIONS
ICE
THROW
The
initial
site
assessment
did
not
adequately
address
the
potential
for
ice
throw,
which
presents
a
significant
risk
to
inhabited
areas
surrounding
the
turbine.
The
recommended
ice
throw
calculation,
1.5
*
(hub
height
+
rotor
diameter),
indicates
a
minimum
safe
distance
of
213
meters
(699
feet).
Per
Figure
6,
Highway
24,
portions
of
the
High
School
and
several
houses
fall
within
this
distance.
BLADES
The
LM
34.5
blades
are
identical
to
blades
used
on
other
1.5
MW
wind
turbines,
which
have
had
a
history
of
manufacturing
defects
and
remediation
for
limited
portions
of
the
production
run
approximately
around
the
time
the
Portsmouth
blades
were
manufactured.
Portsmouth
should
contact
LM
as
soon
as
practical
for
a
blade
inspection
and
serial
number
check
against
blade
serial
numbers
with
known
manufacturing
errors.
YAW
ERRORS
The
824
yaw
fault
should
be
corrected
immediately
in
order
to
allow
the
turbine
to
align
with
the
wind.
High
winds
acting
on
the
blades
in
a
flapwise
direction
could
result
in
blade
structural
damage.
In
discussions
with
Windland
technicians,
the
fault
frequency
can
be
reduced
significantly
by
increasing
yaw
pad
pressure
on
the
slewing
ring
until
Yaw
Power
is
5-‐6
kW
(or
exceeds
the
alarm
threshold
in
the
controller
software).
One
other
operator,
Templeton
Municipal
Light
and
Water,
reports
AMSC
changed
the
parameter
settings
in
the
software.
CONCLUSION
The
following
conclusions
should
be
taken
into
consideration
during
discussions
involving
continued
operation
of
the
turbine:
1. The
AAER
1500/77-‐65m
turbine
is
an
orphan
turbine
with
very
few
operating
in
North
America,
limited
engineering
and
spares
support,
and
a
history
of
gearbox
failures
(3
of
5
installed).
2. The
AAER
1500/77-‐65m
design
never
received
certification
by
any
recognized
certification
organization,
though
similar
designs
licensed
from
AMSC/Windtec
have
been
certified
by
GL
and
DNV.
20
DAR-‐EN-‐20120926-‐revA
Portsmouth
Gearbox
Failure
Investigation
14
October
2012
3. The
gearbox
design
may
be
defective,
specifically
having
insufficient
structural
stiffness
resulting
in
uneven
1st
stage
planet
bearing
load
distribution
and
premature
failure,
or
having
defective
bearings.
4. A
similar
make
and
model
turbines
based
upon
the
AMSC
Windtec
design,
the
Fuhrländer
in
Princeton,
NJ,
has
had
a
similar
gearbox
failure,
though
the
gear
ratios
are
slightly
different.
5. Subsequent
similar
turbine
models
licensed
by
AMSC
Windtec,
the
AMSC
wt1650df
and
the
Hyundai
HQ
1650
appear
to
have
nearly
identical
nacelles,
but
may
incorporate
an
improved
gearbox.
Retrofit
options
should
be
considered.
Attempts
to
obtain
confirmation
of
compatibility
via
AMSC
Windtec
or
Hyundai
to
date
have
been
difficult.
6. Based
upon
a
preliminary
profit/loss
estimate,
it
does
not
appear
the
turbine
can
be
operated
profitably,
even
if
the
gearbox
is
replaced
and
operates
without
failure
for
at
least
10
years.
A
detailed
10-‐15
year
pro
forma
with
anticipated
operating
costs,
debt
servicing,
revenues,
energy
credits,
maintenance
reserves,
and
insurance
should
be
developed
with
assistance
from
a
recognized
authority
such
as
DNV-‐Kema
or
GL
Garrad
Hassan.
7. Due
to
limited
service
experience
in
the
region,
a
collaborative
users
group
should
be
developed.
Potential
collaborators
are
Windland
(operates
3
AAER
1500
in
California),
Templeton
Municipal,
and
Princeton.
8. A
procedure
to
prevent
operation
of
the
turbine
with
ice
on
the
blades
should
be
implemented,
and
a
study
of
anticipated
loss
of
energy
production
due
to
curtailment
should
be
commissioned.
9. The
blades
should
be
inspected
by
LM
Glasfiber,
and
specifically
checked
for
manufacturing
defects
by
serial
number.
10. In
the
absence
of
a
highly
experience
service
provider,
an
Independent
Engineer
with
intimate
knowledge
of
wind
project
operations
should
be
contracted
to
perform
annual
project
monitoring.
11. The
GasTOPS
wear
particle
monitor
should
be
returned
to
service.
This
monitor
will
provide
sufficient
advance
warning
of
a
similar
failure
to
plan
for
repair
or
replacement
at
lower
cost
and
less
downtime
than
the
current
event.
12. A
24/7
vibration
monitoring
system
should
be
installed.
Bachmann
Electronics
is
highly
recommended
for
the
turbine
due
to
the
Bachmann
controller
used
in
the
turbine.
13. Turbine
SCADA
gearbox
bearing
temperature
should
be
more
closely
monitored
for
opportunities
to
obtain
advance
warning
of
bearing
failure.
14. The
blade
pitch
system
calibration
should
be
verified
by
a
qualified
technician,
and
existing
offsets
should
be
recorded
to
determine
whether
the
blade
pitch
is
properly
set.
Incorrect
calibration
may
have
been
a
contributing
factor
to
gearbox
failure.
15. Portsmouth
should
enroll
in
the
free
National
Renewable
Energy
Collaborative
(NREL)
Reliability
Collaborative
to
keep
up
to
date
on
most
pressing
reliability
issues,
learn
methods
to
improve
reliability,
and
network
with
other
owners.
21