PROTEST Procedures for Testing and Measuring Wind Energy ... · PROTEST –Procedures for Testing...

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

PROTEST – 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)

PROTEST – Motivation


• Focus on mechanical systems:

fail not very often but O&M cost dominated by repair of failed

mechanical systems like

moreoften

more costrelevant

PROTEST – Motivation


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


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


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 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

six-step-approach


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

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


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


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


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


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


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


Impact clearly visible in

Torque Load Range Spectra

TU 7%

TU11%

range

counts

2009

2008

Page 25

six-step-approach


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

Authors

Birte-Marie Ehlers, Florian Stache

SUZLON Energy GmbH, Rostock, Germany, +49 381 203578 592,

[email protected], [email protected],

Kris Smolders, Joris Peeters

Hansen Transmissions International nv, Lommel, Belgium, +32 11 54 9412,

[email protected], [email protected]

Thomas Hequet,

Stiftungslehrstuhl Windenergie, Stuttgart, Germany, +49-711-685 68240,

[email protected]

Holger Söker, Oscar Monux,

DEWI GmbH, Wilhelmshaven, Germany, +49 4421-4808-825,

[email protected], [email protected]

mailto:[email protected]








Thank you!

PROTEST Procedures for Testing and Measuring Wind Energy ... · PROTEST –Procedures for Testing...

Documents

Transcript of PROTEST Procedures for Testing and Measuring Wind Energy ... · PROTEST –Procedures for Testing...