Qualifying an On-Line Diagnostic and Prognostic Sensor for Fixed and Rotary Wing Bearings and Gears Karen Cassidy, PhD IEEE Aerospace Conference Big Sky, MT - March 2008
Abstract A sophisticated, mature on-line sensor that provides real-time status of bearing and gear health, and monitors fault initiation, progression, and remaining useful life is qualified for field use. The technique is a continuous, full-flow measurement of metallic particles in the lubrication system using an inductive field that quantifies the time-dependent release of wear debris observed during failures. Multiple sources of data are used to support the qualification process including component rig tests, ground-based engine tests, filter debris analysis, and operational data. This paper outlines the qualification process including establishing limits for early warning of fault indication, and guidelines for in-service monitoring of aircraft engines and helicopter transmissions. The primary condition indicators are critical mass loss and rate; also particle size and count are supplemental indicators. Prognostic algorithms have been developed to set warning and alarm limits, and validated by aircraft OEMs and DOD agencies.
Outline Introduction Prognostics research Theories - Bearing failure behavior - Prediction of remaining useful life - Alarm limits Test data - Component rig tests (bearing, gear) - Propulsion system tests (engine, transmission) - In-service aircraft qualification Summary
Introduction Purpose: Diagnosis and prognosis of bearing and gear failure for aircraft applications - Aircraft gas-turbine engines - Helicopter engines, transmissions and gearboxes Goals of condition monitoring: - Assess the current condition of machinery components - Determine severity of damage - Predict remaining useful life
Stranded SeaKing Helicopter Can we fly home safely?
Bearing Prognostics Research US Air Force Research Laboratory (AFRL/PRTM, Dayton) is working on engine bearing prognostics GasTOPS and AFRL have collaborative R&D program for bearing prognostic methods development - Theoretical models of bearing remaining useful life - On-line oil debris monitor (ODM) sensor data - Wear debris analysis using x-ray fluorescence Goal: To predict the remaining life of rolling element bearings under conditions experienced by the main-shaft bearings of an advanced military gas turbine engine
Oil Debris Sensor On-line full-flow inductive sensor fitted in lube oil line Detects 100% of particles above minimum particle size Measures number, size, mass of ferrous & non- ferrous debris Detects spall initiation, progression, rate Can be used to quantify damage severity and remaining useful life
In-Service Aircraft ODM Applications F22 Raptor: F119 engine Joint Strike Fighter: F135 engine & lift fan Eurofighter Typhoon: EJ200 engine SeaKing Helicopter: engine and gearbox Pilatus PC12: PWC PT6A engine F22 F35 Typhoon Sea King PC12
Surface Fatigue Failure Failure mechanism that ODM technology designed to monitor is common failure mode for bearing & gears Rolling contact fatigue manifested as spalling phenomena; can be initiated on surface or subsurface - Surface initiated spall from poor lubrication or pitting - Subsurface initiated spall is associated with stress concentrations Material surface is subjected to static and dynamic loading - Up to a certain critical spall size, dynamic loading is stable Spall detection and diagnosis has been studied by many researchers; still a lack of reliable prognostic methods to predict the remaining useful life
Theory: Regions of Spall Progression L10 theories Probabilistic Change bearing Extend usability Slow, stable Load static + (stable) dynamic Increase mass, temp, vibration Duration varies Fast, unstable Vibration => end M crit Critical stress, spall length, vib End useful life RUL Inst. End
Surface Fatigue Failure: Cause and Effect What factors may directly affect spall propagation rates? - Load (stress), RPM (variable, operator controlled) - Material properties, component geometry, lubrication properties What are direct symptoms? - Spalling: release of component wear debris into lube oil - Increase in vibration, dynamic stresses, and temperature What are key condition indicators? - Wear debris: mass, rate, size, count & composition - Vibration
Test Data Bearing and Gear Component Rigs - GasTOPS and National Research Council small scale bearing rig - GasTOPS and Pratt & Whitney full scale aircraft bearings - AFRL 40mm bearing rigs - test bearing materials and fluids - NASA Glenn Research Center bearing & gear component rigs - Spur Gear component rig - truck transmission model Engine & Transmission Test Stands - NASA OH-58 Kiowa helicopter main rotor transmission test stand - CAF Sea King helicopter engine & gearbox test facilities - DTSO Bell 206 helicopter main rotor transmission test stand - F22 Raptor - F119 engine pre-flight tests - AH-64 Apache helicopter - transmission test stand In-Service Aircraft Qualification - Eurofighter Typhoon / EJ200 - Pilatus PC-12 / PWC PT6A Engine
Characteristic Debris Accumulation 8000 Particle Size > 200 um Number of Particles 6000 4000 2000 0 > 250 um > 300 um > 350 um > 400 um > 500 um > 700 um 0 20 40 60 80 100 120 * Elapsed Time [%] Ref: JL Miller (Pratt & Whitney) and D. Kitaljevich (GasTOPS Ltd.), In-line Oil Debris Monitor for Aircraft Engine Condition Assessment, IEEE 2000
AFRL Bearing Test Program 40 mm bearing test facility in Dayton, OH Bearing life and spall propagation tests on materials, lubricants, load/stress levels Life tests - GasTOPS ODM used to quantify effect of variables on bearing life Spall propagation fests - Uses failed bearings from life test to explore remaining useful life - Measure spall propagation rate for stress in 250-350 ksi range - Measure effect of load and stress on rate - ODM sensor characterizes damage progression; correlated to mass loss
AFRL Rig Effects of Load on Prognostics 200 180 160 140 364 ksi 314 ksi 300 ksi 300 ksi 278 ksi 278 ksi Mass Loss(mg) 120 100 80 Critical Mass Rate Critical Mass Loss 60 40 20 250 ksi 0 0 100 200 300 400 500 600 700 800 900 1000 Elapsed Time(mins) 52100 steel Outer race 278ksi Outer race 278ksi Inner race 364ksi Outer race 314ksi Outer race 300ksi Outer race 300ksi Outer race 250/278ksi Data from an inductive sensor can be processed to obtain prognostic information. Ref: Forster, Thompson, Toms, Horning, ISHM 2005
AFRL Rig with M50 NiL Propagation Rates 180 160 140 52100 278 ksi 52100 278 ksi Mass Loss(mg) 120 100 80 60 40 20 Affect of Stress on Rate During Operation: 52100 250 ksi 52100 278 ksi M50 NiL 300 ksi M50 NiL 350 ksi 0 0 20 40 60 80 100 120 140 160 Cycles (Millions) 250 ksi at 1820 Min. Plus 490 Min. at 278 ksi Ref: Forster, Thompson, Toms, Horning, ISHM 2005
Reduction in Debris Generation Rate with Reduction in Stress after Damage Rate has Accelerated 180 160 Projected damage if stress remained at 278ksi Influence rate 140 Mass Loss(mg) 120 100 80 60 40 Stress 278ksi Mass loss rate increase Operator reduced stress to 250ksi 20 0 0.0 10.0 20.0 30.0 40.0 50.0 Cycles (Millions) Ref: N. Forster & K. Thompson (AFRL-PRTM); A. Toms & S. Horning (GasTOPS), Assessing the Potential of a Commercial Oil Debris Sensor as a Prognostic Device for Gas Turbine Engine Bearings, ISHM 2005
NASA Glenn Test Rigs Goal to quantify debris generation during bearing & gear wear Measured debris progression, counts, mass, particle size Test Methods: Hybrid bearing, tapered roller bearing, spur gear, spiral bevel gear, OH-58 helicopter transmission, and others Quantify failure effects in components and complex systems ODM mass during spur gear failure Ref: Dr. Paula J. Dempsey, Dr. David G. Lewicki (ARL) and Harry J. Decker (ARL), NASA Glenn Research Center, Cleveland, OH
Spur Gear Rig Test 90 80 70 ODM Mass (mg) 60 50 40 30 20 10 0 85 90 95 100 Time on Test (hrs) XRF of debris
F119 Engine / F22 Aircraft Pre-Flight Tests New engine run on test stand ODM detected initial damage Bearing replaced, no secondary damage occurred. Found that: - Damage due to assembly error - Bearing highly over-stressed Debris rate returned to normal
In-Service Aircraft Qualification Application: Eurofighter Typhoon / EJ200 Condition Indicators - Total Mass Accumulation Level and Rate - Large Particle Accumulation Level and Rate EJ200 Debris Database - 3 Bench Test Engines - 7 Flying Development Engines Validation - Bearing rig tests used for initial condition indicator limits - Correlated ODM mass rate to legacy debris monitor limits - Database of wear debris data (ODM, chip detector and oil filters) of healthy and faulted engines used for ongoing limit verification
In-Service Aircraft Qualification Application: Pilatus PC-12 / PWC PT6A Engine Condition Indicators - Level 1 Threshold - total particle count threshold - Level 2 Threshold - short term particle count rate - Level 3 Threshold - medium term particle count rate Validation - Normal engine oil contamination rates evaluated in test cells Over 100 Production Engines and 50 repair/overhaul engines - Over 350 in-service aircraft
AH-64 Apache Helicopter Transmission Application: Naval Air Station at Patuxent River Helicopter Transmission Test Facility Condition indicator: total mass - Right nose gearbox sensor detected high quantity of wear debris Damage verification - X-ray debris analysis showed M50 in right nose gearbox, 100 x left side - Teardown showed one roller over 50% of contact surface had spall; early signs in other rollers and race
Summary Bearing & gear failures complex and affected by multiple variables - Load, speed, & material properties affect the failure rate - Rate may be decreased by changing operation during failure event - Need prognostic information to aid in making operation decisions ODM sensor provides quantitative data to support diagnostics and prognostics for bearings and gears - 15 years of condition indicators verified by military, government & OEMs - Verification of parameters including critical mass loss & mass rate - Validated by engine tear-down and x-ray debris analysis US Air Force and GasTOPS collaborating on research to provide prognostics, prediction of remaining useful life on aircraft engines
Contact Information Karen Cassidy, PhD President, GasTOPS Inc. Application Center, Pensacola, FL Kcassidy@gastopsUSA.com, Tel: (850) 478-8512, www.gastopsusa.com Nelson Forster, PhD Mechanical Systems Branch, AFRL Wright-Patterson AFB, Dayton, OH Nelson.Forster@WPAFB.AF.MIL, Tel: (937) 255-5568 Duka Kitaljevich VP Product Sales, GasTOPS Ltd., Ottawa, ON Dkitaljevich@gastops.com, Tel: (800) 363-8658, www.gastops.com