protest procedures for testing and measuring wind energy ... · protest –procedures for testing...
Post on 26-Apr-2020
8 Views
Preview:
TRANSCRIPT
SITE ASSESSMENT . WIND TURBINE ASSESSMENT . GRID INTEGRATION . DUE DILIGENCE . KNOWLEDGE . CONSULTANCY
PROTEST – Procedures for Testing and Measuring Wind Energy Systems
Drive Train Case Study II
Holger Söker, DEWI GmbH
Page 1PROTEST – Background Information
PROcedures for TESTing and measuring wind energy systems
• Collaborative Project in the EU-FP7 BUDGET: 2.7 Mio €
• Start: March 2007 End: Sept. 2010
• Participants:
ECN (NL) – project co-ordinator,
Suzlon Energy GmbH (DE),
DEWI (DE),
Germanischer Lloyd (DE),
Hansen Transmissions International (BE),
University of Stuttgart (DE),
CRES (GR)
Page 2PROTEST – Motivation
PROcedures for TESTing and measuring wind energy systems
• Focus on mechanical systems:
fail not very often but O&M cost dominated by repair of failed
mechanical systems like
moreoften
more costrelevant
Page 3PROTEST – Motivation
PROcedures for TESTing and measuring wind energy systems
While design procedures
for blades and towers
are detailed
Such for other mechancial
components are rather vague
Potential risk ofcomponent failure
Rapid increaseof turbine size
Shortcomings in component
load measurements
Shortcomingsin load
simulation
PROTEST – Approach
PROcedures for TESTing and measuring wind energy systems
Page 4
WP
1 State Of Art Report
WP
2 Load cases and design drivers
WP
3 Loads at interfaces
WP
4 Prototype measurements definition
6-step approach
WP
5 Case Studies
Drive Train
Pitch System
Yaw System
PROTEST – Approach
IEC61400-13 Approach seemed to be
not adequate
impossible to define a rigid testing
procedure with fixed channels and
sampling rates!
6-STEP Approach allows for
• different design and concepts
• different computational models
• allows flexibility to serve the model
validation task
Page 5
PROTEST- WP5 Case Study on Drive Train
Application of the six-step-approach
for validation of a wind turbine drive train
design & run SIMPACK model
validate SIMPACK model against validated FLEX5
sensitivity study of model input parameters
measurement strategy for model validation
measurement data processing and analysis
methods to estimate model parameters from
measurements
comparison of SIMPACK model results/measurements
Page 6
six-step-approach
STEP 1: failure modes
No special failure mode was chosen
as the main focus was set on testing
the process of model design
measurement set-up
methods of data processing
validation of designed models
Page 7
six-step-approach
STEP 2: design model
turbine model:
28 DOFs
rotation of low speed shaft
two bending DOFs for main shaft
torsion between hub and generator rotor
inertias of rotating parts are modelled in
one body -individual rotating bodies is not
considered
rotation of high speed shaft is defined by
low speed shaft rotation (stiff connection
/gear ratio )
FLEX5 model – 4DOF for Drive Train
Page 8
required parameters: overall torsional drive train stiffness, drive train dampingtransmission ratio
9
six-step-approach
STEP 2: design model
SIMPACK model – stage2 sophisticated model
Page 9
implementation of inertia, stiffness and damping
between rotating bodies
included gear box model:
torsional stiffness of gear box mounts
torsional stiffness of shafts and gear teeth
stiffness of the different gear stages
10
six-step-approach
STEP 4: determine relevant parameter
Sensitivity Analysis on gives information on the effect of
input input parameters uncertainty on simulation results
starting point for measurement campaign (STEP 5)
1st approach:
vary model parameter (e.g. high speed shaft stiffness)
run a modal analysis
observe the change of the resulting eigenfrequencies and eigenmodes
judge what uncertainty is acceptable
2nd approach:
vary model parameter
run load simulation for relevant DLC’s
analyze results i.t.o. load statistics, Rainflow count
judge what uncertainty is acceptable.
Page 10
turbine type :
Suzlon Energy S82 / 1500kW
rotor diameter: 82m
gear box:
Hansen, EH751A
site:
Tamil Nadu/India
Page 11
six-step-approach
STEP 5: measurement campaign
12
Measurement setup
standard IEC 61400-13 load signals
blade loads (root beding edgewise, flapwise)
main shaft loads (bending and torque LSS)
tower top torsion
tower base bending
sampling rate 50Hz
for drive train model validation, load validation
shaft speeds and torques
displacements of gearbox housing
temperatures of bearings and oil
oil pressures
six-step-approach
STEP 5: measurement campaign Page 12
13
Rotational speed measurements
six-step-approach
STEP 5: measurement campaign Page 13
14
Focus on
model validation
manned measurements
to capture specific
measurement load cases
MLC
December 2008,
June 2010
contiuous monitoring to
capture normal power
production MLCs
December 2008,
June 2010
six-step-approach
STEP 5: measurement campaign Page 14
Step Quantity to Check Example for Methods Objective of ValidationStep
1 Documentation Selected Time
Series
Comparison of model dataagainst weighing log
Spectral analysis ofselected time series forvarious operational states(e.g. in partial and fullload)
Main structuralproperties like masses,stiffnesses,eigenfrequencies andcoupled modes
2 CharacteristicCurves
Visual comparison ofcurves of operationalparameters (e.g. speed,power) and loading forseveral environmentalconditions
Validation of basiccontrol characteristicsand rotor aerodynamicsas well as mechanicaland electricalparameters (e.g.losses)
3 Time Series ofvarious operationalstates, like power production start stop emergency stop
Visual comparison of datain time and frequencydomain
Check of statisticalproperties of data
Analysis of decay rates ofoscillations duringstopping procedures
Dynamic behaviour allimportant andassessable operationalstates with focus onaerodynamic mode,controller model andactuator models
Structural andaerodynamic damping
4 Post-ProcessedData
Comparison of loadingspectra like rainflow distribution load duration distributions damage equivalent loads
Final check of turbinebehaviour and dynamicproperties
Check of all previouslyperformed validationsteps
Page 15six-step-approach
STEP 5: measurement campaign
Normal transients in manual campaign
NTA Normal start-up (+ grid connection)
10th Dec, 13:30, 15:08,
15:40, 17:38, 17:48, 19:40
11th Dec, 7:31, 7:44
NTB Normal stop (run to pause)10th Dec, 17:34
11th Dec, 7:41
NTC Grid loss => this case can NOT be provoked, but occurs frequently in India
NTD Pause (idling at low speed => 2 RPM)9th Dec: 17:38 - 18:02
10th Dec, 11:42 - 11:48
NTE Idling at high speed => 16RPM ("waiting for wind without grid connection")
9th Dec: 18:05 - 18:11
10th Dec, 12:20 - 12:35,
13:10 - 13:18,
NTF Stand still without rotor lock 10th Dec: 17:18
NTG Stand still with rotor lock not measured
NTHConstant speed at X RPM - no generator connection = idling (=> X is in the
range [ 100 - 1750 RPM] - generator side)
9th Dec: 18:05 - 18:11
10th Dec: 12:20 - 12:35,
13:10 - 13:18,
Run-up
Page 16
Special transients in manual campaign
STA E-stop with mechanical brake disk (brake program 6 from ground) not measured
STB E-stop without mechanical brake disk (= equivalent to grid loss) 10th Dec: 12:35
STCE-stop after activation of overspeed guard when generator is NOT
connected (e.g. set blade pitch angle to 20 degrees and wait)
9th Dec: 18:12,
10th Dec: 13:18,
STD E-stop after activation of overspeed guard during power production10th Dec: 17:43,
11th Dec: 7:29,
STFConstant speed at Y RPM - power production (=> Y is in the range [ 1500 -
1600 RPM] - generator side => adapted FLEXISLIP)10th Dec: 14:18 - 15:00
STG Slow reverse rotation not measured
STHLow Voltage Ride Through => special container required (not available in
India)not measured
STIShort circuit (in the generator OR grid) => special container required (not
available in India)not measured
STJ Constant power production at Z kW (=> Z is in the range [ 750 - 1500 kW]) 10th & 11th Dec
six-step-approach
STEP 5: measurement campaign
Resonance
Operation at constant power levels
rotational speed[rpm]
exitation fequency[Hz]
generator 1495.2 24.92
rotor 1st order 15.7 0.26
rotor 3rd order 47.1 0.79
rotor 6st order 94.2 1.57
excitation frequencies
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0 0.5 1 1.5 2 2.5 3 3.5 4
Am
plit
ude [
kN
m]
NP, 42kW NP, 200kW NP, 376kW NP, 624kW NP, 984kW Res, 517kW Excitation-Eigenfreq
Rotor 1st Rotor 3rd Rotor 6th 1st Drive Train
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
0.1
0 10 20 30 40 50 60 70 80 90 100
Am
plit
ude [
kN
m]
NP, 42kW NP, 200kW NP, 376kW NP, 624kW NP, 984kW Res, 517kW Excitation-Eigenfreq
Generator Shaft
six-step-approach
STEP 6: data processing
Selected Time Series: Operation at different load levels for Eigenfrequency determination
Page 17
fHSS : high speed shaft rotating
frequency
fISS : intermediate speed shaft rotating
frequency
fLSS : low speed shaft rotating frequency
fROT : rotor–main shaft–planet carrier
rotating frequency
fHSM : high speed stage meshing
frequency (ISS – HSS)
fISM : intermediate speed stage meshing
frequency (LSS – ISS)
fLSM : low speed stage meshing
frequency (planetary stage)
six-step-approach
STEP 6: data processingPage 18
Selected Time Series: Run-up for Eigenfrequency determination
P_WT_cut T_gen_c2_cut S_rot_c2_cut
300
400
500
600
700kW
1
2
3
4
5kNm
100
200
300
400
500kNm
820 846 872 898 924 950
s
Selected Time Series: Deliberate resonance (modified control parameters)
six-step-approach
STEP 5: measurement campaign Page 19
six-step-approach
STEP 6: data processing Page 20
analysed data: resonance case
assumption:
relangle
torquestiffness
Selected Time Series: Stiffness determination - deterministic approach
torque signal and difference in angle between the two
measurement points at the low speed and high speed
shafts
21
six-step-approach
STEP 6: data processing
to_erase: torque_smophs
-2
-1
0
1
2
3
kNm
13:43:00 13:43:05 13:43:10
10.12.08
h:m:s
StiffnessTrange
highhighrotorrangerotor
analysed data:stationary resonance
acceleration and damping assumed to
be neglectible from one interval n to n+1
5 seconds
Range
of torque
of twist
median_stiffness histo_stiffness
0
2
4
6
8
10
counts
30 35 40
kNm/rad
stiffness
30
35
40
45
kNm/rad
13:43:00 13:43:20 13:43:40 13:44:00 13:44:20
10.12.08
h:m:s
Page 21
Selected Time Series: Stiffness determination - Stochastic approach
RFC’s indicate high variations in mechanical torque around rated torque
further investigation is required to determine (1) the impact and (2)
ability of the FLEX5 simulation model to consider this effect
six-step-approach
STEP 6: data processing Page 22
Post- Processed Data: Rainflow Count
23
P_WT_c
0.0
0.5
1.0
1.5
MW
Trot Tgen_c Taxis_c
0.0
0.5
1.0
MNm
0
5
10
kNm
Mhz
-1.0
-0.5
0.0
0.5
1.0MNm
12:40 12:45 12:50
13.12.08
h:m
six-step-approach
STEP 6: data processing
Trot= mech. Torque LSS
Tgen_c = mech. Torque KTR Sensor
HSS
Taxis_c = mech. Torque Telemetry HSS
Page 23
P_WT_c
0.0
0.5
1.0
1.5
MW
Trot Tgen_c Taxis_c
0.0
0.5
1.0
MNm
0
5
10
kNm
0
5
10
kNm
Mhz
-1.0
-0.5
0.0
0.5
1.0MNm
12:48:00 12:48:20 12:48:40 12:49:00
13.12.08
h:m:s
Trot_Mean Trot_Max Trot_Min
0.000
0.133
0.267
0.400
0.533
0.667
0.800
0.933
1.067
1.200MNm
Trot_EQL Trot_StDev
0
20
40
60
80
100
120
140
160
180
200
220
240
kNm
0
50
100
150
200
250
300
350
400
450
500kNm
4 6 8 10 12 14 16
m/s
six-step-approach
STEP 6: data processing
Trot_Max Trot_Mean Trot_Min
0.000
0.100
0.200
0.300
0.400
0.500
0.600
0.700
0.800
0.900
1.000
1.100
1.200
1.300
1.400
1.500MNm
Trot_StDev Trot_EQL
0
50
100
150
200
250
300
350
400
450
500
550
600kNm
0
50
100
150
200
250
300
350
400
450
500kNm
4.00 6.00 8.00 10.00 12.00 14.00 16.00
m/s
2008 monitoring statistics 2009
Page 24
Trot_RP_tu07 Trot_RP_tu11
1.00·102
1.00·103
1.00·104
1.00·105
1.00·106
1.00·107
1.00·108
1.00·109
1.00·1010counts
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40
MNm
Trot_RP_Fatigue_TU07 Trot_RP_Fatigue_TU11
1·101
1·102
1·103
1·104
1·105
1·106
1·107
1·108
1·109counts
0.0 0.2 0.4 0.6 0.8 1.0 1.2
MNm
six-step-approach
STEP 6: data processing
Impact clearly visible in
Torque Load Range Spectra
TU 7%
TU11%
range
counts
2009
2008
Page 25
six-step-approach
STEP 6: data processing
Impact clearly visible in
Time @ Level Plots
Torque level
Trot_TaL_Fatigue_TU07 Trot_TaL_Fatigue_TU11
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
2.0
2.2
2.4
2.6
2.8
3.0
3.2
3.4
3.6
3.8
4.0
4.2%
0.0 0.2 0.4 0.6 0.8 1.0 1.2
MNm
Trot_TaL_tu07 Trot_TaL_tu11
0.00
0.10
0.20
0.30
0.40
0.50
0.60
0.70
0.80
0.90
1.00
1.10
1.20
1.30
1.40
1.50
1.60
1.70
1.80
1.90
2.00
2.10
2.20
2.30
2.40
2.50%
0.00 0.20 0.40 0.60 0.80 1.00 1.20
MNm
2009
2008
Rela
tive fre
quency
Page 26
Conclusions
sensitivity studies to define relevant model input parameters
(like inertia, stiffness, damping) are important to setup measurement
campaigns
measurements for model parameter validation have been carried out
different methods to determine eigenfrequencies, stiffness, damping and
intertia have been developed and applied
stiffness values are reproduced
all methods show similar trends with respect to inertia and damping values
further investigations on the applied methods are needed
Page 27
Page 28Authors
Birte-Marie Ehlers, Florian Stache
SUZLON Energy GmbH, Rostock, Germany, +49 381 203578 592,
birte.ehlers@suzlon.com, florian.stache@suzlon.com,
Kris Smolders, Joris Peeters
Hansen Transmissions International nv, Lommel, Belgium, +32 11 54 9412,
KSmolders@HansenTransmissions.com, peeters@hansentransmissions.com
Thomas Hequet,
Stiftungslehrstuhl Windenergie, Stuttgart, Germany, +49-711-685 68240,
thomas.hecquet@ifb.uni-stuttgart.de
Holger Söker, Oscar Monux,
DEWI GmbH, Wilhelmshaven, Germany, +49 4421-4808-825,
h.soeker@dewi.de, o.monux@dewi.de
Page 29
Thank you!
top related