You are on page 1of 49

High Speed Auxiliary Fan

Failure and Subsequent Fan


and Electric Motor
Modifications

History

Auxiliary ventilation has been used in underground coal


mining for many years
In the seventies 9m3 fans were the norm, in the
eighties 13m3, in the nineties 18m3 and today 22.5m3
and larger fans are required
With the increase in fan capacity comes the need for
more Kilowatts, more impeller mass and greater
system inertia.

History

13 cubic metre per second

9 cubic metre per second

History

22.5 cubic metre per second

18 cubic metre per second

History

In 1991 MDG3 was initially published to provide


manufactures of auxiliary fans with the construction and
safety features for fans required to be approved under the
Coal Mines Regulation Act of 1982
Industry expectations and the publishing of this document
resulted in the upgrading of fans from having a very basic
onboard explosion protected enclosure housing a basic
isolator, contactor, overload and receptacle configuration to
a more sophisticated control system as we see today

Occurrence
of Fan
Failures
1994 Tower Fan Failure
2006 Austar, Mandalong,
Wambo, Ashton Coal and
Newpac
The failures were similar in all
fans:
Contact between rotor shaft
and the flame path bushing
generated heat due to friction
Flame path bushing and shaft
friction welded

History

As you can see in the significant incident report, in April


1994 an auxiliary ventilation fan failure occurred in the
Southern Coal Fields of NSW resulting in an under ground
fire on the fan. This failure was almost identical to the
catastrophic failure which occurred at Newpac Mine which is
the subject of todays presentation
There have also been many fan failures which have
occurred in the past 20 years and these fans have been
simply repaired, rebuilt and returned to service. It should
be noted that all fan failures recorded occurred on fans
having two pole motors rotating at 2980 RPM
A number of these failures have been very close to
resulting in catastrophic fan failures as a number of you
here today would have witnessed

History

The Department, with assistance from key industry


representatives, published an updated Mines Department
Guideline MDG 3 in 1996, in an attempt to set minimum
design standards for Australian Ventilation Fans and to
reflect current safety expectations.
This guideline was
quickly adopted by the industry as the standard for the
rebuild and overhaul of fans. The key engineering safety
controls which were reinforced in this MDG3 document
were:

To

Vibration monitoring of the Drive Trains


Temperature monitoring of Bearings
Anti Wind Milling Protection
Noise level controls
mention just a few of the key design criteria .

In 2006 a number of fan failures occurred which were as a


direct result of the onboard electric motor failure. The fan
configurations included the impeller directly mounted to the
motor shaft, no carrier shaft was incorporated into the fan
designs

Ashton Coal

Ashton Coal fan taken out of service due to repeated trips


caused by vibration

Impeller balance checked on site and found to be okay

Motor bearings appeared to be source of vibration

Motor stripped and inspected at G & D Donnelly


Drive End
Flame Path
Bushing

Flame path exhibited signs of high temperature


(discolouration)
Flame path housing scored by contact with shaft

Ashton Coal Fan

Flame Path on Rotor Shaft at Drive End

Flame path surface on shaft damaged by friction


welding and scoring by contact with housing

Ashton Coal
Bearing Mounting Bore in
DE End Shield

Bearing mounting bore in DE end shield worn by rolling


outer race
Damage to bearing outer race

DE Bearing Outer Race

Newpac

On 14 August 2006 a catastrophic failure occurred to


a 22.5 cube high speed fan operating underground in
the hazardous area at Newpac Coal Mine in the Hunter
Valley. The impellor was direct mounted to the motor
shaft.
It is this failure and the findings that this session will
cover in detail today

Initial Investigation

On the morning of 14 August we received a phone call


from the acting Engineering Manager at Newpac Mine,
advising that a Hire Fan, Plant Number AF20214, had
catastrophically failed while operating underground in
the hazardous zone
The Department of Primary Industry had been notified
and an underground inspection was pending
After the underground inspection was completed by
the Department of Primary Industry inspectors, we as
equipment OEM, were advised that the fan was being
transported to the surface where the fan was
quarantined, until further investigation into the failure
could be undertaken

Initial Investigation

As the OEM Fan Manufacturer, we were summoned to


a site meeting with colliery representatives, mine site
officials, and DPI investigators. The site meeting
involved interviews with relevant mine staff including
the underground manager, maintenance personnel
and statutory officials, plus OEM representatives
During the meeting every attempt was made to track
and document the operation and maintenance issues
of the fan, in particular the shifts leading up to failure
The maintenance and service documentation/manuals
provided by the OEM were tabled along with Colliery
Maintenance records

Result of Fan failure

Loss of explosion protection methods Exd due to the motor


junction box breaking away from the motor

Motor holddown bolts

Result of Fan failure


Impeller detached from motor shaft
and laying in fan casing

Damaged inlet eye

Result of Fan failure

Bent motor shaft


flame path subjected
to temperatures
above 8500C

Result of Fan failure


Structural and mechanical
damage

Fan Configuration

22 m3 high speed (2850 RPM) centrifugal auxiliary


ventilation fan
Impeller direct mounted to motor shaft
150 kW, 2 pole, 1000 volt Exd motor with side mounted
terminal box

Protection System Fitted

The fan was manufactured in August 2005 and fully


complied with MDG3.
Fan protection included:
Vibration monitoring including an (External Accelerator)
Bearing temperature monitoring on both the D.E. and
N.D.E Bearing
Relay logic control
Anti Wind Milling Protection

Block Diagram

Electrical
System

Control System Audit

On 25 August the fan electrical system was examined jointly


by Vale and DPI personnel with the following issues reported;
Motor overload screwed up above normal trip settings
Vibration system failed to trip (accelerometer on motor foot
inactive)
Accelerometer mounted with its axis incorrectly oriented
Bearing temperature monitoring on double ended housing active
but didnt register on over temperature status

Examination at OEM Factory

On 28 August the fan was systematically disassembled in


the presence of representatives of the DPI, CCI Pope on
behalf of Newpac Mine, WEG, G&D Donnelley, Airgasco
(Vale fan consultant) and Vale Engineering personnel
Key motor components were sent to CCI Pope for detailed
metallurgical analysis
Airgasco were to compile a detailed report in the form of a
Failure Mode Effect Analysis
Finding by all parties were reported to subsequent meetings
at the DPI offices in Maitland
Safety alerts were issued to mines using this type of fan

Most Likely Failure Sequence

Impact damage or uneven build up of particulate on the


impeller caused above normal fan running vibration
(3mm/sec)
Vibration not detected by faulty accelerometer fails to
shut down fan (6mm/sec)
Vibration not detected by mine personnel attending
toolbox talk in crib room
Combination of shock loading from vibration and the
bearing outer race sliding fit causes damage and wear to
bore in DE end shield as the bearing outer race rolls in the
bore
Wear in DE end shield decreases radial clearance in flame
path
The bearing retainer/flame path bushing comes in
contacts with the rotor shaft

Most Likely Failure Sequence (contd)

Friction heat is generated between the bushing and


the rotor shaft
Rapid heat build up destroys the bearing grease
causing the bearing to fail catastrophically
Shaft temperature at the flame path exceeds 850oC
reducing rotor shaft material properties
Shaft reaches critical condition and yields, causing
massive bending and centrifugal forces
Large centrifugal forces cause an unstoppable chain of
events leading to the damage observed in the fan
Damage included detachment of motor terminal box
from motor frame parting stater pig tails which deenergised the motor stopping the fan

Root Cause

Root cause of drive end bearing failure is a combination of


vibration and bearing outer race fit in the end shield.
As the severity of vibration caused by imbalance increases
the impeller develops an orbital motion.
Since the bearing outer race fit is a slide fit the outer race
begins to roll in the end shield bore due to the
circumferential difference between the outer race and the
end shield bore, the bore being greater
Metal removed from the bearing and the end shield by this
process further exacerbates the wear accelerating
deterioration of the bearing and its mounting
Eventually the flame path bushing comes in contact with
the rotor shaft
The heat generated by this contact results in failure of the
bearing and destruction of the shaft in a catastrophic mode

Issues

Large inlet screen size (50 x 50) allows objects


capable of causing damage to enter impeller
Less than ideal vibration transducer location may
not trip fan when large vibrations occur
Failure in single
monitoring system

transducer

disables

vibration

Close clearance between bearing retainer/flame path


housing and shaft makes contact inevitable if wear
occurs between DE bearing outer race and mating
bore in end shield or, if the bearing fails

Issues (contd)

Flame path housing material is not conducive to


acting as a bearing so when it comes in contact with
the rotor shaft at the high relative surface speeds
rapid friction heating occurs
Bearing fit used by WEG does not follow SKF
recommendations
Possibility that fixed (DE) bearing outer race is not
positively clamped on the outer race shoulders by
bearing retainers increasing likelihood that it can
rotate in end shield bore

Considerations

WEG motor bearing L10 life under normal loads is


more than adequate but bearings can fail
prematurely for a variety of reasons usually
lubrication or vibration related
Premature bearing failure will normally result in
detectable vibration before axis constraint (bearing
condition) is seriously compromised
A small but significant increase in bearing clearance
will result in shaft contact with the inboard bearing
retainer/flame path bushing
WEG rotor shaft stiffness is adequate - ratio of
critical speed to actual speed approx 1.4
These are high speed machines with limited
tolerance of large out of balance or foreign body
impact

Considerations (contd)

Out of balance can occur over time due to uneven


build-up of particulate on the impeller
Jack shafts are not free of problems in service either
Alignment between motor and jack shaft to be
maintained throughout life of fan
Coupling between motor and jack shaft must
accommodate angular and radial miss-alignment
Longer drive train must withstand mounting base
deflection
Two additional bearing assemblies to maintain
Skating bearing outer races have also been observed in
jack shaft bearings
No flame path in jack shaft housing

Possible Remedies Hard Barriers

Locate vibration transducers to measure horizontal


vibration at the motor DE on the x axis normal to the
(z ) shaft axis. This should detect vibrations
sufficiently early to prevent major damage
Fit second vibration transducer if possible directly to
DE bearing outer race
Revise DE bearing outer race fit to comply with SKF
recommendation
Rework existing motors to install steel sleeve in DE
end shield (Donnelly modification) to accomplish SKF
fit and provide harder material in which outer race is
mounted

Possible Remedies Hard Barriers (contd)

Ensure bearing retainers apply a positive clamping


force to DE bearing outer race
Fit longer bolts for motor terminal box attachment
to more closely match strength of motor frame
material
On new fans, use alternative motor with better history
of reliability
Fit a smaller mesh screen to prevent large objects
impacting the impeller

Possible Remedies Soft Barriers

Introduce vibration transducer inspection and change


out program similar to used with radio remote
transmitters
Regularly inspect and clean the impeller and remove
debris from the stone drop out box
Adopt predictive maintenance techniques that can
detect bearing deterioration preceding failure allowing
preventive action to be taken regular vibration
survey
Update pre-despatch check list to ensure all key
features are installed and functioning

WEG Motor Bearing Mounting

G & D Donnelly have been recovering motor end


shields by installing a press fit steel sleeve bored to a
closer (but not tighter) machining tolerance
Motors fitted with steel sleeve have subsequently not
been returned as the result of the same failure
Motor on Kestrel fan modified in Dec 2003 sleeve
machined to WEG tolerances (+0.018 / -0.000)

Returned in May 2004 with suspected bent shaft

Shaft found to be not bent, sleeve bore still good

Returned in Dec 2005 sleeve bore oversize 0.060

Bearing Fit
SKF Recommended Fit

Actual Fit Used in Motor

Housing fixed bearing


Fit class - K7
150.000+0.012 / -0.028

Housing fixed bearing


Fit class H5
150.000+0.018 / -0.000

150.012
149.972

150.018
150.000

Outer Race
Fit Class - h5
150.000+0.000 / -0.018

Outer Race
Fit Class - h5
150.000+0.000 / -0.018

150.000
149.982

150.000
149.982

Theoretical fit: 0.028 interference


0.030 clearance

Theoretical fit: 0.000 interference


0.036 clearance

Probable fit:

0.021 interference
0.023 clearance

Probable fit:

0.005 clearance
0.031 clearance

Median fit:

0.001 clearance

Median fit:

0.018 clearance

SKF

MOTOR

Alternative Motor Manufacturer


Bearing Size: 6316 C4

Bearing Size: 6316 C4

Housing DE fixed bearing


Fit class M6
170.000 -0.008 / -0.033
169.992
169.967
Outer Race
Fit Class h6
170.000+0.000 / -0.025
170.000
169.975
Theoretical fit: 0.033 interference
0.017 clearance
Probable fit:
0.026 interference
0.010 clearance
Median fit:
0.008 interference

Housing NDE floating bearing


Fit class closest to G7
170.000+0.030 / +0.055
170.055
170.030
Outer Race
Fit Class h6
170.000+0.000 / -0.025
170.000
169.975
Theoretical fit:
0.030 clearance
0.080 clearance

Comparison Alternate Motor Versus Motor Fitted


Alternative Motor
Bearing Size: 6316 C4
Housing DE fixed bearing
Fit class M6
170.000 -0.008 / -0.033
169.992
169.967
Outer Race
Fit Class h6
170.000+0.000 / -0.025
170.000
169.975
Theoretical fit: 0.033 interference
0.017 clearance
Probable fit:
0.026 interference
0.010 clearance
Median fit:
0.008 interference

Weg Motor Fitted to Fan


Bearing Size: 6314 C3
Housing fixed bearing
Fit class H5
150.000+0.018 / -0.000
150.018
150.000
Outer Race
Fit Class - h5
150.000+0.000 / -0.018
150.000
149.982
Theoretical fit:

0.000 interference
0.036 clearance

Probable fit:

0.005 clearance
0.031 clearance

Median fit:

0.018 clearance

WEG Investigation - Fan Tests

Flakt Woods
Limited clean air tests were conducted on a fan at
Melbourne factory
Run time was short motor not reaching normal operating
temperature
Air flow vanes adjusted to different settings
demonstrating no measurable affect of different flow
volumes on vibration levels
Test results demonstrated the base line operation of new
product prior to shipment on a solid concrete floor base

WEG Investigation - Fan Tests (contd)

Vale SMS
Test setup outside factory with approx 50m of duct
attached to fan intake
The operational results showed acceptable vibration
levels and no indication of bearing defects
After tests similar to those at Flakt Woods were
complete solids were introduced into the air stream at
the duct intake
Sand, coal and small stones were thrown into the intake
duct passing through the fan exiting as dust
Larger rocks were introduced but these did not get past
the screen and were found in the stone drop out box in
front of the inlet vanes
Other debris such as a hard hat, rain coat, rubber hose
and paper were thrown in
Air flow control vanes were also adjusted to various flow
rates but no noticeable changes in vibration levels were
detected

WEG Investigation - Fan Tests (contd)

Mine Tests
Tests were conducted at three mines

Wambo Coal, Newpac Mine and Ashton Coal

A total of five fans tested under normal underground


operating conditions
All fans were operating at acceptable vibration levels
These fans had been in service for between 5 days and
10 months since their last OEM service/overhaul
Four motors showed lubrication deficiencies
In some sites vibration analysis had only been
introduced recently
Vibration levels as high as 16mm/sec had been reported
Vibration sensors were mounted in various locations
from motor feet to control enclosures
Vibration levels indicated by portable equipment differed
from the levels indicated by on board systems

WEG Motor Modification

Up-grade instructions for all existing 150kW 2 pole Exd


motors in direct mounted 22m3/sec auxiliary fans
Initial inspection Upon removal of motor from fan,
disassemble motor and record:

DE and NDE bearing housing diameter


DE and NDE Bearing OD
Record quantity and condition of grease in motor bearing
and caps
Motor serial number
Fan serial number

Flame path inspection

A full inspection of motor flame paths, as per AS/NZS 3800


is required
Proceed with necessary repairs, if required
Report any problems to WEG

WEG Motor Modification (contd)

Up-grade
New end-shield with M6
bearing housing tolerance to
be fitted
The DE bearing will change to
6314 C4
The bearing caps must be
fully packed with Esso Polyrex
EM grease

WEG Motor Modification (contd)


Comparison between old and new bearing fit
Existing fit

Revised fit

Fan Modification

Progressively modify existing fans to add jack shaft


Supply jack shafts on all new high speed fans
Design includes bearing vibration and temperature
monitoring
Vale offers an alternative motor on new fans

Jack shaft

Toshiba motor

Fan Modification (contd)

Mount vibration transducers to the impeller end of the jack


shaft housing and the foot on the motor DE to monitor
lateral axis vibration

Motor transducer

Jack shaft transducer

Fan Modification (contd)

If not already modified, modify WEG motor when fans are returned
to add jack shaft
Replace existing 50 x 50 mesh screen in ADV with 25 x 25 mesh to
catch a higher percentage of objects in the air stream capable of
causing impact damage to the impeller
Strengthen motor jacking bolt blocks to prevent bending as
occurred on Newpac fan

Reinforced motor jacking bolts blocks

ADV screen

The Future?

Are high speed fans with or without jack shafts a


permanent solution to the problem of bearing and
electric motor reliability?

Do we need a new generation of fans?

Smarter maintenance strategies?

Smarter monitoring systems?

Are low speed fans a better option?

You might also like