F-111 EBU TF30 RAM Program

1
F-111 EBUTF30 RAM Program

• John Hall (Quality Manager)
• Greg Mason (Engineering Manager)

2
Presentation Outline

• F-111 EBU Overview
• What is a RAM Program
• What is a Condition Monitoring Philosophy
• What Condition Monitoring Techniques are used
• Performance Indicators
• Reliability / Safety
• Availability / Health
• Maintainability / Cost Effectiveness
• Applicability to other industries
• Nine steps

3
F-111 EBU Overview

• F-111 EBU supports the Pratt Whitney TF30 gas
  turbine fleet that powers the RAAF F-111.
• Support includes Fleet planning, Engineering,
  Deeper Maintenance, and Spares (and RI) inventory
  management
• Engine facts
• 100 plus engines managed (65 in rotable fleet)
• Designed in the 1960s (first Turbo Fan engine
  built)
• Approx 2M each (4M in piece parts)
• Four variants supported (will be one in two
  years)
• Only TF30 operator (USN retired TF30 powered F14s
  in 2005)
• Oldest fleet (time since new increased risk)

4
Engine picture (now removed)

• 16 stages of compression overall compression
  ratio 191
• Length 19 feet, diameter 39 inches, weight 4 160
  lbs, max thrust 21 000 lbs,
• Max temp 2200 ºF, Max pressure 250 psia,
• Airflow 240 lb / sec (by pass ratio 11), max
  fuel flow 55 000 lb / hr (15 lb / sec or 9 L /
  sec)
• Twin spools (9600 rpm and 14400 rpm - single fan
  blade exerts a centrifugal force of 30 000lb)
• Eight combustion cans (arranged in an annulus),
  four fuel nozzles per can (primary and secondary
  flow for each)
• Four stages of turbine
• Variable, five zone, Afterburner

5
TF30 RAM Program

• Why a RAM program?
• To continually improve Reliability, Availability
  and Maintainability (quality, timeliness and
  cost)
• What is a RAM program?
• Essentially a management philosophy based on
  Condition Monitored Maintenance (CMM) principles
  ie use knowledge of individual equipment
  condition (engine in our case) to make decisions
  on each piece of equipment and the fleet at
  large.
• How was the RAM program introduced?
• In 1990 a group of people who had Condition
  Monitoring experience and a strong desire to
  provide cost effective TF30 maintenance, became
  aware of the current engine failure and Test Cell
  rejection rates and decided to act.

6
TF30 RAM Program

• Unscheduled removal rates too high (30 of all
  work performed)
• Review of engine failure modes and reasons for
  engine rejection at Test Cell. Identified as
• Main line bearing failures / oil system leaks
• High engine vibration levels
• Low Turbine Inlet Temperature (TIT) margin
• Engine accessory failures
• Repeated maintenance for same problem (addressing
  symptoms not cause),
• Inefficient maintenance strategies when
  performing major maintenance (fixed interval, one
  size fits all)
• Need recognised by middle managers and the RAAF
  TF30 RAM program was born..followed by 3 years
  of persistence, stealth and self promotion in
  order to prove the concept.
• A few early successes helped and the RAM program
  was formalised in 1993 (senior management
  commitment). From that time to this, it has
  officially had staff, equipment, knowledge and a
  goal.
• More on RAM program implementation later in the
  presentation

7
Condition Monitoring - Philosophy

• An effective RAM / CM program
• must be focused and structured
• Five CM Phases
• Detection, taking readings and obtaining raw CM
  data
• Diagnosis, analysing this data to determine
  failure mode(s) in play
• Prognosis, decide on course of action based on
  failure criticality (time remaining until engine
  removal / repair)
• Prescription, decide what scope of maintenance is
  reqd and perform it
• Post Mortem, gaining feedback during repair and
  using it to improve CAC, build techniques,
  facilities, CM techniques!

8
Condition Monitoring Techniques

• Technique selection
• Chosen based on ability to monitor important
  failure modes
• Not chosen based on - the way we always do it,
  equipment offered by salesman, cheapest or
  easiest
• Five Techniques
• Spectrometric Oil Analysis
• Wear Debris Analysis
• Vibration Analysis
• Gas Path Performance Analysis
• Remote Visual Inspections

9
Condition Monitoring Techniques

• Spectrometric Oil Analysis (approx. 20 monitored
  faults)
• Interval every three flights and every flight in
  warning (same day sample, burn and analysis,
  rapid failures ie lt10 enhrs)
• Automated trending (not just data collection)
• Included Oil Additions and a smoothing algorithm
• Trend corrected concentration and wear rate (not
  uncorrected concentration)
• Automated predictions (time to break warning /
  alarm levels including colour coding and
  messages)
• Not used in isolation (WDA, Maintenance history,
  modification status etc)

10
Condition Monitoring Techniques

• Spectrometric Oil Analysis

11
Condition Monitoring Techniques

• Spectrometric Oil Analysis

12
Condition Monitoring Techniques

• Wear Debris Analysis (faults as per SOA, provides
  confirmation)
• Interval is pre/post maintenance, as indicated by
  SOAP and ?P
• From main engine filter (no mag plugs, particles
  often non magnetic)
• Use 15 micron Pall Dirt Alert filter (high
  efficiency debris recovery)
• Also use 3 micron Dirt Alert for Green Run
  post major maintenance
• Optical microscope with digital camera (initial
  analysis)
• SEM EDX if worthy of further examination using on
  base NDTSL Materials Officers
• Stored electronically for comparison with metal
  map (OEM and RAAF developed)

13
Condition Monitoring Techniques

• Vibration Analysis (approx 20 faults, balance,
  alignment, straightness, concentricity,
  looseness, rubs etc)
• Interval is pre/post maintenance and when removed
  serviceable (slow failures ie gt200 enhrs)
• Transducers (accelerometers) on Fan Inlet Case,
  Diffuser and Turbine
• Data acquired by predominately using RAAF / DSTO
  developed Engine Vibration Analysis System (EVAS)
• Use Run up / down plots (amplitude phase) and
  FFT
• Trending / fault isolation using various DSTO
  software and Vibralog / Entek software
• Current Development of IEVAS (portable version of
  EVAS)

14
Condition Monitoring Techniques

• Vibration Analysis
• Single channel system replaced with PC based 8
  channel system
• Steady state acquisition and one slow run-up /
  run-down rather than 6 to complete vibration
  survey
• Takes 2 minutes instead of 12, providing approx
  200k fuel savings per year
• Provides both Transient and Steady State Peak
  vibration data
• Single page report generated for serviceability

15
Condition Monitoring Techniques

• Gas Path Performance Analysis (approx 15 faults
  monitored)
• Interval is pre/post maintenance and when removed
  serviceable (slow failures ie gt200 enhrs)
• Normal parameters trended (N1, N2, Tt2, Tt4, Tt5,
  Tt7, Wf, thrust, Pt2, Ps3, Ps4, Pt7m)
• Also looking at Tt3, Ps3f.
• Acquired using PC based Engine Data Acquisition
  System (EDAS developed by DSTO, supported and
  maintained by Raytheon)
• Use OEM EPR Plots, OEM influence coefficients,
  plus various cross plots (DSTO)
• Trending using EBU developed Excel program
  (against same engine, average of the serviceable
  band or average engine)
• Fault isolation aided by various DSTO developed
  software
• Approx last 10 years of data used to trend
• Two Compressor wash types performed (water and
  detergent) at specified intervals to recover
  performance, help with corrosion.

16
Condition Monitoring Techniques

• Remote Visual Inspection / Videoscopes (monitors
  corrosion, erosion, cracking, oil leaks, P/N and
  build checks)
• Interval is pre and post maintenance, special
  servicings, as reqd (slow failures ie gt200 enhrs)
• Access to Fan, majority of LPC, front/rear of
  HPC, majority of combustion area, front of HPT
  and rear stages of LPT (access often limited by
  imagination)
• Use latest generation 4, 6 and 8 mm videoscopes
• Features include stereoscopic measurement,
  working channel, digital image / video capture
  and comparison, improved optics / light source /
  portability
• Under investigation in-situ blending and NDT

17
Engine Inductions

• To provide an insight into how this CMM
  philosophy is used on the TF30 a number of
  initiatives will be discussed. The first is
  engine inductions.
• Every engine removed for deeper maintenance is
  prepared for induction into work. This involves
• data collection eg test cell run and existing CM
  data (five techniques), configuration data (ECs
  incorporated) and lifing data (LCF and other
  lifed components).
• Analysis of above data
• The goal is to tailor the maintenance to be
  performed so that
• minimum necessary / optimal work is carried out,
• maximum engine life is produced,
• highest probability of successful engine test
  post maintenance
• reduced support costs
• This work scope is documented in a work
  requirement and is conveyed to maintenance
  supervisors at an induction meeting (chaired by
  RAM staff). NB Scheduled maintenance such as OH
  and HSI have a fairly standard work requirement.

18
Engine Failure Modes

• The second initiative is to use the Condition
  Monitoring and reliability data to target
  Engineering Changes / Build techniques that will
  best help improve TF30 RAM outcomes.
• Essentially we are doing the post mortem phase in
  a systematic manner ie various Tech info /
  reports raised during engine repair are reviewed
  in an attempt to not just repair the engine but
  to eliminate that failure mode throughout the
  whole fleet. Sounds simple, but isnt!
• Challenges are many, some are
• too busy / not enough staff, no dedicated RAM
  focus on long term opportunities (bush fires are
  raging, no time available to build a fire engine)
• concentrate on treating symptoms, forget to
  remove the cause
• time, effort, cost to develop and implement ECs
  is often daunting
• Not enough ECs introduced to RAAF fleet
• Unnecessary ECs introduced to RAAF fleet (fixed a
  problem we didnt have)
• management administrata (distracted from the goal
  of improved RAM outcomes)

19
Engine Failure Modes

• Some TF30 examples are
• ECs, 25 relevant mods introduced during current
  OH series (23 worked only 4 in 7 years during
  late 80s - mid 90s, more to come!)
• Build Improvements, numerous processes to improve
  engine condition eg 5 bearing housing lug
  clearance software
• Facilities, numerous but some are
• upgrade bearing cleaning, inspection, repair
  facilities
• upgrade plasma spray room, welding workshop
• dehumidified engine storage
• Equipment, continual but some are
• purchase of DCCMM
• purchase of air bed run out table
• digital blade moment weighing / automated
  patternising
• bearing flushness indicating system
• Old engine but new processes / equipment /
  facilities

20
Maintenance Interval Extensions

• Maintenance Interval Extension (Scheduled via FLP
  and via MIER)
• In 1990, scheduled intervals were 750 / 1500 enhr
  for HSI and OH. Pre pacer were 550 / 1100.
• We are currently exploring a 1000 / 2000 enhr
  schedule via a Fleet Leader Program that is
  validated by previous PW ASMET testing for the
  USAF and USN.
• Initial goal was 1200 / 2400 but this is looking
  unachievable without JP8100 and TBC NGVs.
  Latest Mission Profile Analysis also shows
  increased AB lights.
• Additionally, all engines approaching scheduled
  maintenance are reviewed for extension. Due to
  variability of build, use repair history it is
  inevitable that many engines can safely operate
  beyond scheduled maintenance. Some can not! This
  process relies heavily on CM, build / lifing data
  etc
• RAAF TF30 DM currently produces 7200 enhr per
  annum using 100 staff and associated spares
  MIER produces 3000 additional enhr with lt5 staff
  and limited spares - saves money, safety not
  compromised! Was 15 OH / yr in early 80s, now 7.

21
Performance Indicators (1990 VS 2005)

• Reliability / Safety
• Availability / Health
• Maintainability / Cost Effectiveness

22
Performance Indicators (1990 VS 2005)

• Reliability Safety
• In Flight Shut Down (IFSD) rate
• In early 1990s was 0.6 / 1000 enhrs or 4.5 per
  annum
• 2005 was 0.07 / 1000 enhrs or lt 1 per annum or
  84 reduction c/w early 1990s
• Compares favourably with F404 fitted to RAAF
  F/A18 which has an IFSD rate of 0.28 / 1000 enhrs
• Oct 96-Jan 01 achieved 4.5 years with no IFSD
• Goal - continually decrease IFSD rate 0.02
  achievable for TF30 within 5 years (97 reduction
  c/w early 1990s

23
Performance Indicators (1990 VS 2005)

• Reliability / Safety
• Unscheduled Engine Removal Rate (UERR)
• In early 1990s was 7 / 1000 enhrs or 55 per annum
• 2005 was 3 / 1000 enhrs or 21 per annum or 65
  reduction c/w early 1990s
• Most removals FOD (6) and engine accessories (5).
• Without FOD UERR was 6 / 1000 enhrs in 1990 and
  2.1 / 1000 enhrs in 2005 or a 70 reduction c/ w
  1990.
• Goal - continually decrease UERR rate focus on
  FOD control initiatives and accessory reliability

24
Performance Indicators (1990 VS 2005)

• Reliability / Safety
• Average Time On Wing (ATOW) NB engines removed
  serv for AF servicings not counted
• was 160 enhrs (or 1 year installed), now 300 (2
  years installed)
• c/w RAAF F/A 18 F404, currently 400 enhrs (2.3
  years installed)
• was 300 three years ago

25
Performance Indicators (1990 VS 2005)

• Reliability / Safety
• Major Repair Arising Rate (MRAR)
• In early 1990s was 1.2 / 1000 enhrs or 9 per
  annum
• 2005 was 0.4 / 1000 enhrs or 3 per annum or 66
  reduction c/w early 1990s
• Approximately half of the Major Repairs have been
  due to FOD
• With FOD removed 0.9 /1000 enhrs (1990) and 0.2 /
  1000 enhrs (2005) or 78 reduction c/w 1990s
• Goal - continually decrease MRAR rate focus on
  FOD control initiatives.FOD very frustrating

26
Performance Indicators (1990 VS 2005)

• Availability / Health
• Good in 1990, slightly better in 2005 - no bare
  firewalls
• Engines serviceable above fit
• average of 8 early 1990s when fleet was 69
  engines and 18 aircraft with engines fitted (EAR
  3.6 1)
• average of 7.5 in 2005 when fleet was 65 engines
  and 22 aircraft with engines fitted (EAR 31)
• Fleet Health (total engine hours available to be
  used until next scheduled maintenance provides a
  contingency buffer and insight into future DM
  workload requirements)
• average of 21 000 enhrs (stable) early 1990s
• average of 24 000 enhrs (increasing) in 2005

27
Performance Indicators (1990 VS 2005)

• Maintainability / Cost effectiveness
• Maintenance cost (labour, spares, etc) was 31M
  per annum and increasing in 1991 dollars (or 46
  M in 2005 dollars)
• Maintenance cost was lt20M in 2005 or 56
  reduction c/w 1991
• OR
• 3.5K per ENHR flown in 1991 (5.1K / ENHR in
  2005 dollars)
• 2.5K per ENHR flown in 2005 or 1.9K per ENHR
  made in 2005
• Maintenance costs should decrease further as RAAF
  F111 approaches PWD (currently 2010-2012)
• Interesting trend - in 1990 RAAF used 12 LSA
  (Engineering, Fleet Planning, Logistic ie TSA /
  RIM / SIM) and 250 DM staff in 2005 we used 40
  LSA and 105 DM staff ie we have found it to be
  extremely cost effective to invest up front in
  planning / analysis and save down track in DM
  costs (labour and spares)

28
Applicability to other industries

• RAM applicable to high cost and / or critical
  process industries (eg civil aviation, mining,
  rail, paper, power, racing, manufacturing etc)
• Must be in a position to implement
  (organisationally, culturally, disciplined
  approach, technically, long term view and
  commitment)

29
RAM Program Implementation

• 1. Justify (Commitment)
• 2. Baseline (KPIs)
• 3. Target (Failure modes that hurt most)
• 4. Select (Appropriate CM techniques)
• 5. Resource (Team of people, equipment)
• 6. Train (CM techniques and CMM philosophy)
• 7. Implement (hard work, discipline, attention
  to detail)
• 8. Monitor (review KPI progress / failure
  modes)
• 9. Adjust (fix the bits of the program that
  arent working)
• Beware of pitfalls, use a facilitator if possible