wind turbine drivetrain condition monitoring (presentation), nrel
TRANSCRIPT
NREL is a national laboratory of the U.S. Department of Energy, Office of Energy Efficiency and Renewable Energy, operated by the Alliance for Sustainable Energy, LLC.
Wind Turbine Drivetrain Condition Monitoring
Shuangwen (Shawn) Sheng Senior Engineer, NREL/NWTC
CREW Seminar Series October 7, 2011
Boulder, CO
NREL/PR-5000-52908
NATIONAL RENEWABLE ENERGY LABORATORY
Outline
Introduction Drivetrain Condition Monitoring Tests Results and Discussions Concluding Remarks
2
DOE
1.5
MW
Win
d Tu
rbin
e /P
IX17
245
NATIONAL RENEWABLE ENERGY LABORATORY
Introduction
3
A typical utility-scale wind turbine
Global Wind Energy – Then and Now
Wind Turbine Gearbox Reliability Challenge
Gearbox Reliability Collaborative (GRC)
A typical Condition Monitoring (CM) system
Benefits of machine CM
CM under the GRC
Cost justification of wind turbine CM
NATIONAL RENEWABLE ENERGY LABORATORY
A typical utility-scale wind turbine
4
Turbine • Foundation • Tower • Nacelle • Hub • Blades
Test
Tur
bine
s at N
REL/
PIX1
8937
Nacelle • Main bearing • Main shaft • Gearbox • Brake • Generator shaft • Generator • Yaw and pitch systems
(not illustrated)
NATIONAL RENEWABLE ENERGY LABORATORY
Wind Turbine Gearbox Reliability Challenge
6
Gearboxes do not always achieve their 20-year design life
Premature failure of gearboxes increases
the cost of energy • Turbine downtime • Unplanned maintenance • Gearbox replacement and rebuild • Increased warranty reserves
The problem • Is widespread • Affects most original equipment manufacturers (OEMs) • Is not caused by manufacturing practices
NATIONAL RENEWABLE ENERGY LABORATORY
Gearbox Reliability Collaborative (GRC)
7
Facilitate dialog among all parties • Designers and consultants • Suppliers and rebuilders • Operations and maintenance organizations
Understand gearbox response to specific loading • Pure torque, bending, thrust (dynamometer) • Turbulence (field)
Understand the physics of premature wind turbine gearbox failure
Identify gaps in design process
Suggest improvements in design practices and analytical tools
NATIONAL RENEWABLE ENERGY LABORATORY
Gearbox Reliability Collaborative (Cont.)
8
Technical Approach • Modeling and
analysis • Dynamometer test • Field test • Failure database • Condition monitoring
Goal • To improve gearbox
reliability and increase turbine uptime, which in turn will reduce the cost of energy
Field Test Dynamometer Test • Test plan • Test article • Test setup & execution
• Test plan • Test turbine • Test setup & execution
Analysis • Load cases • System loads • Internal loads
Test
Tur
bine
at N
REL/
PIX1
9022
NREL Dynamometer/PIX16913
NATIONAL RENEWABLE ENERGY LABORATORY
A typical Condition Monitoring System[2]
9
Physical Measurement: convert measured physical symptoms, typically to electrical signals
Data Acquisition: data transmission from sensors to processing unit
Feature Extraction: extract characteristic information from raw sensor data
Pattern Classification: diagnosis of defect type and severity level Life Prediction: prognosis of remaining service life of the
monitored structure.
Monitored Structure
Physical Measurement
Feature Extraction
Data Acquisition
Pattern Classification
Life Prediction
Vibr
ation
Leve
l Bearing Failure
Alarm Trigger
Healthy ThresholdVibr
ation
Leve
l Bearing Failure
Alarm Trigger
Healthy Threshold
NATIONAL RENEWABLE ENERGY LABORATORY
Benefits of Machine CM
10
Early deterioration detection to avoid catastrophic failure
Accurate damage evaluation
to enable a cost-effective maintenance practices
Root cause analysis to
recommend improvements in component design or equipment operation and control strategy.
Test
Tur
bine
at P
onne
quin
/PIX
1925
7
NATIONAL RENEWABLE ENERGY LABORATORY
Condition Monitoring under the GRC
11
The main deliverable from the GRC: improved gearbox design and modeling practices • Helps the industry as a whole • Provides no information on individual
turbine health condition
Health Monitoring • Helps provide individual turbine
health information and increase turbine uptime
• Condition monitoring (CM) for drivetrain (gearbox, generator, and main bearing) and structural health monitoring (SHM) for rotor
Test
Tur
bine
at N
REL/
PIX1
9022
NATIONAL RENEWABLE ENERGY LABORATORY
Cost Justification of Wind Turbine CM
12
Return on Investment for all three cases less than 3 years
Based on 1.5 MW wind turbine with replacement costs of about €150,000 for gearbox, €38,000 for a generator and €25,000 for a main bearing (DEWI)
Costs for planned repair < 30% for unplanned replacement (DEWI) Cost per CM system approximately €5,000 plus €1,000 per year per wind
turbine (service) Above cost savings do not include loss of production
Operator / Owner
# of Turbines Duration of Service
Costs CMS plus Service in €
Detected Damages Costs unplanned Replacement Costs planned Repair in €
Total Savings in €
enviaM 15 WTG‘s 5 years
150,000 3 x Gearbox 405,000 101,250
303,750 In 5 years
e.disnatur 130 WTG‘s 5 years
1,300,000 12 x Gearbox 40 x Generator bearing
4,620,000 1,155,000
3,465,000 In 5 years
juwi Management
59 WTG‘s 3 years
472,000 20 x Gearbox 1 x Generator bearing 1 x Main bearing
2,811,000 702,750
2,108,250 In 3 years
Experience at Schenck [3]
NATIONAL RENEWABLE ENERGY LABORATORY
Drivetrain Condition Monitoring
13
Downtime caused by Turbine Subsystems
Typical CM Practices
GRC CM Approach and Rationale
NATIONAL RENEWABLE ENERGY LABORATORY
Downtime Caused by Subsystems
14
Data Source: Wind Stats Newsletter, Vol. 16 Issue 1 to Vol. 22 Issue 4, covering 2003 to 2009[4]
Based on the data reported to Wind Stats for the first quarter of 2010, the data represents: about 27,000 turbines, ranging from 500 kW to 5 MW.
Top Three: 1. Gearbox 2. Generator 3. Electric Systems
Consider crane cost: • Main bearing also needs
attention • Electric systems often do
not need an expensive crane rental.
NATIONAL RENEWABLE ENERGY LABORATORY
Typical Drivetrain CM Practices
15
Techniques • Acoustic emission (e.g.,
stress wave) analysis • Vibration analysis • Oil debris analysis
Typical Practices • Integration of vibration or
stress wave with oil debris analysis
Sample Vibration Spectra[5]
Sample Oil Debris Counts[5]
NATIONAL RENEWABLE ENERGY LABORATORY
GRC CM Approach and Rationale
16
Integrated Approach • Acoustic emission (stress wave) • Vibration analysis • Oil debris and condition monitoring techniques • Electric signature-based technique
Rationale • No single technique can detect all possible failure modes seen in
wind turbines • Each technique has its own strengths and limitations • Combine active machine wear detection capability of lubrication
oil monitoring techniques with crack location pinpointing capability of AE and vibration analysis
• Investigate potential technique for direct-drive turbines
NATIONAL RENEWABLE ENERGY LABORATORY
Tests
17
NREL 2.5-MW Dynamometer
Test Articles
Dynamometer Test Setup
Tests Conducted
Damaged GRC Gearbox #1
Lubrication System Diagram
CM System Configuration for Dynamometer Test
NATIONAL RENEWABLE ENERGY LABORATORY
NREL 2.5-MW Dynamometer[6]
18
Configuration • 2.5-MW induction motor • Three-stage epicyclical speed
reducer • Variable-frequency drive with full
regeneration capacity
Capabilities • Performance and reliability tests on wind turbine drivetrain
prototypes and commercial machines • Static, highly accelerated life and model-in-the-loop tests • Bed plate tilts up to 6° to mimic angle between the plane of the
blades and the turbine main shaft, as seen in the field • Rated output torque 1.4 MNm • Speed variation from 0 to 16.7 rpm rated torque, 16.7 to 30 rpm
rated power • Non-torque loads rated up to 880 kN for radial and 156 kN for thrust
NREL 2.5 MW Dynamometer/PIX08581
NATIONAL RENEWABLE ENERGY LABORATORY
Test Articles
19
Two gearboxes rated at 750 kW • One planet stage and two parallel stages • Redesign
Floating sun, cylindrical roller planet bearings, tapered roller bearings in parallel stages, pressurized lubrication, offline filtration and desiccant breather
• Up to 150 channels of measurements for loads, displacements, and temperature
NATIONAL RENEWABLE ENERGY LABORATORY
Tests Conducted
21
Dynamometer test of GRC gearbox #1: run-in Field test of GRC gearbox #1 Dynamometer test of GRC gearbox #2: run-in and non-
torque loading Retest of GRC gearbox #1 in the dynamometer
NREL 2.5 MW Dynamometer/PIX16913
NATIONAL RENEWABLE ENERGY LABORATORY
Damaged GRC Gearbox #1
22
Annulus
Planet
Sun Gear
Gear Pinion
Pinion
Low-Speed Stage
High-Speed Stage
Intermediate-Speed Stage
Low-Speed Shaft
Intermediate-Speed Shaft
High-Speed Shaft
Planet Carrier
1. Completed dynamometer run-in test 2. Sent for field test: experienced two oil losses 3. Stopped field test 4. Retested in the dynamometer under controlled conditions
High-Speed Stage Gear[7]
NATIONAL RENEWABLE ENERGY LABORATORY
A CM Configuration in Dynamometer Tests[7]
24
As a research project, this set up is beyond the typical drivetrain CM configuration seen in the industry.
NATIONAL RENEWABLE ENERGY LABORATORY
Results and Discussion
25
Run-in Test Results • Oil Cleanliness Level
• Oil Debris Monitoring
• Oil Sample Analysis
Dynamometer Retest of the Damaged GRC Gearbox #1
• Stress Wave Analysis
• Vibration Analysis
• Oil Debris Monitoring
• Oil Condition Monitoring
Discussion
NATIONAL RENEWABLE ENERGY LABORATORY
Run-in Results: Oil Cleanliness Level[8]
26
A broad range of particles were found in the oil • Results: GRC
gearbox #1 • All three bins • Similar results: GRC
gearbox #2
Contamination level
• Increases when generator speed is ramped up • Decreases when generator is shut down and with the use of an operating
lubricant filtration system • Potentially useful for controlling the run-in of gearboxes
NATIONAL RENEWABLE ENERGY LABORATORY
Run-in Results: Oil Debris Monitoring[9]
27
0100200300400500600700800900
100011
/2
11/4
11/6
11/8
11/1
0
11/1
2
11/1
4
11/1
6
11/1
8
11/2
0
11/2
2
11/2
4
11/2
6
11/2
8
11/3
0
12/2
12/4
Date
0
50
100
150
200
250
300
11/2
11/4
11/6
11/8
11/1
0
11/1
2
11/1
4
11/1
6
11/1
8
11/2
0
11/2
2
11/2
4
11/2
6
11/2
8
11/3
0
12/2
12/4
Date
Oil debris monitoring conducted throughout the GRC gearbox #2 run-in test period • K1 (left top), K2 (left
bottom) and K3 (right top) • Trends are similar though
particle counts vary
NATIONAL RENEWABLE ENERGY LABORATORY
Run-in Results: Oil Sample Analysis[9]
28
Results: GRC gearbox #2 • Particle counts: important to identify particle types
Analysis Results Reference Limits
• Element identification
NATIONAL RENEWABLE ENERGY LABORATORY
Dynamometer Retest: Stress Wave Analysis[7]
29
Dynamometer retest of GRC gearbox #1 (right) indicated abnormal gearbox behavior
Parallel stages sensor GRC gearbox #2
dynamometer test (left) indicated healthy gearbox behavior
NATIONAL RENEWABLE ENERGY LABORATORY
Dynamometer Retest: Vibration Analysis[7]
30
Intermediate speed shaft sensor
GRC gearbox #2 dynamometer test (left) indicated healthy gearbox behavior
Dynamometer retest of GRC gearbox #1 (right) indicated abnormal gearbox behavior • More side band frequencies • Elevated gear meshing
frequency amplitudes
NATIONAL RENEWABLE ENERGY LABORATORY
Dynamometer Retest: Oil Debris Monitoring[7]
31
0
100
200
300
400
500
600
700
800
9/15 9/15 9/16 9/16 9/17 9/17 9/18
Part
icle
Cou
nts
Date
Particle generation rates: • Damaged GRC gearbox #1: 70 particles/hour on 9/16 • Healthy GRC gearbox #2: 11 particles over a period of 4 hours
NATIONAL RENEWABLE ENERGY LABORATORY
Dynamometer Retest: Oil Condition Monitoring
32
Field test of GRC gearbox #1 (left): • Wild dynamics • Possible damage
Retest of GRC gearbox #1 (right): • Well controlled test
conditions • Possible damage
NATIONAL RENEWABLE ENERGY LABORATORY
Discussion
33
Oil cleanliness level measurement can potentially be used to control and monitor wind turbine gearbox run-in.
Oil debris monitoring, specifically particle counts, is effective for monitoring gearbox component damage, but is not effective for damage location.
If sensor mounting location is appropriate, similar trends in oil debris counts between the offline filter loop and the inline filter loop can be obtained.
When obtaining particle counts through oil sample analysis, attention should be given to identifying particle types.
Periodic oil sample analysis may help pinpoint failed component and root cause analysis.
NATIONAL RENEWABLE ENERGY LABORATORY
Discussion (Cont.)
34
Stress wave amplitude histogram appears effective for detecting gearbox abnormal health conditions.
Spectrum analysis of vibration signal (or stress waves) can, to a certain extent, pinpoint the location of damaged gearbox components.
Damaged gearbox releases particles at increased rates. Oil condition monitoring, specifically moisture, total
ferrous debris and oil quality: • Oil total ferrous debris appears indicative for gearbox
component damage • More data is required to understand oil moisture and quality
Electric signature-based technique did not reveal any gearbox damage in this study.
NATIONAL RENEWABLE ENERGY LABORATORY
Challenges[10]
36
Justification of cost benefits for CM: each wind turbine has a relatively lower revenue stream than traditional power generation
Limited machine accessibility: makes retrofitting of CM systems or taking oil/grease samples difficult
Cost-effective and universal measurement strategy: sensor readings are affected by mounting locations and various drivetrain and gearbox configurations
Diagnostics: variable-speed and load conditions and very low rotor speeds challenge traditional diagnostic techniques developed for other applications
Data interpretation: requires expert assistance for data analysis and maintenance recommendations
Oil sample analysis: sample variations, different lubricant may require different sets of tests or procedures
Additional complexity for offshore: foundation, undersea transmission lines, saltwater and wave influences on turbine, and weather forecast.
NATIONAL RENEWABLE ENERGY LABORATORY
Future R&D Areas[10]
37
Determine cost-effective monitoring strategy Improve accuracy and reliability of diagnostic decisions,
including level of severity evaluation Automate data interpretation to deliver actionable
recommendations Develop reliable and accurate prognostic techniques Improve use of Supervisory Control and Data Acquisition
(SCADA) system data Research fleet-wide condition monitoring and asset
management Improve turbine operation, control strategy, and
component design through root cause analysis
Challenging yet Rewarding
NATIONAL RENEWABLE ENERGY LABORATORY
References
38
1. Global Wind Energy Council. Global Wind Report Annual Market Update 2010. http://www.gwec.net/index.php?id=180
2. R. Gao, R. Yan, S. Sheng, and L. Zhang, “Sensor Placement and Signal Processing for Bearing Condition Monitoring”, Condition-Based Monitoring and Control for Intelligent Manufacturing, Lihui Wang and Robert X. Gao (Eds.), Springer-Verlag, pp. 167-191, ISBN 1-84628-268-3, 2006.
3. Kewitsch, R. “Optimizing Life Cycle Costs (LCC) for Wind Turbines by Implementing Remote Condition Monitoring Service,” presented at the AWEA Project Performance and Reliability Workshop, January 12–13, 2011, San Diego, CA.
4. Wind Stats Newsletter, 2003–2009, Vol. 16, No. 1 to Vol. 22, No. 4, Haymarket Business Media, London, UK.
5. Dempsey, P. and Sheng, S. “Investigation of Data Fusion for Health Monitoring of Wind Turbine Drivetrain Components,” presented at the 2011 American Wind Energy Association WINDPOWER Conference, Anaheim, CA, USA, May 22-25, 2011.
6. National Renewable Energy Laboratory. 2.5-MW Dynamometer Testing Facility Fact Sheet. http://www.nrel.gov/docs/fy11osti/45649.pdf. Accessed July 27, 2011.
7. Sheng, S. “Investigation of Various Condition Monitoring Techniques Based on a Damaged Wind Turbine Gearbox,” 8th International Workshop on Structural Health Monitoring 2011 Proceedings, Stanford, CA, September 13-15, 2011.
8. Sheng, S.; Butterfield, S.; and Oyague, F. “Investigation of Various Wind Turbine Drivetrain Condition Monitoring Techniques,” 7th International Workshop on Structural Health Monitoring 2009 Proceedings, Stanford, CA, September 9-11, 2009.
9. Sheng, S. “Investigation of Oil Conditioning, Real-time Monitoring and Oil Sample Analysis for Wind Turbine Gearboxes,” presented at the 2011 AWEA Project Performance and Reliability Workshop, January 12–13, 2011, San Diego, CA.
10. Sheng, S. and Veers, P. “Wind Turbine Drivetrain Condition Monitoring – An Overview,” Machinery Failure Prevention Technology (MFPT) Society 2011 Conference Proceedings, Virginia Beach, VA, May 10-12, 2011.
NATIONAL RENEWABLE ENERGY LABORATORY
Thanks for Your Attention!
39
Special thanks go to GRC CM partners: CC Jensen, Castrol, Eaton, GasTOPS, Kittiwake, Herguth Laboratories, Lubrizol, Macom, SKF, SKF Baker Instruments, and SwanTech!
NREL’s contributions to this presentation were funded by the Wind and Water Power Program, Office of Energy Efficiency and Renewable Energy of the U.S. Department of Energy under contract No. DE-AC02-05CH11231. The authors are solely responsible for any omissions or errors contained herein.