Advances in Condition Monitoring Linking the Input to the Output Martin Jones Insensys

1
Advances in Condition Monitoring Linking the
Input to the OutputMartin JonesInsensys
2
Turbine Monitoring Todayzxcz

• Many advanced turbines continuously acquire and
  transmit measurement data to a remote location
• Wide range of parameters often measured covering
• Inputs
• Wind conditions, yaw angle, blade pitch angle,
  etc
• Outputs
• Power, bearing wear, etc
• Measurements are used for performance monitoring,
  condition monitoring and fault protection
• Providing analysis of turbine performance
• Indicating when intervention is required

3
SKF Condition Monitoring Systemszxcz

• Data from a wide range of sensors gathered in iMU
• Data reduction in iMU prior to onward
  transmission over ethernet
• Alarms flagged via SMS and email
• Data sent to remote client in control room
• for display and interpretation

Control Room Client
Ethernet
Ethernet
iMU
ProCon Server
Email SMS
Alarm messages
4
Drive Train Monitoring Todayzxcz

• Monitoring of drive train outputs
• Unbalance
• Alignment errors
• Bearing problems
• Damage on gear wheels
• Shaft bending
• Mechanical looseness
• Tower vibrations
• Electrical problems
• Resonance problems

5
Drive Train Monitoring Methodszxcz

• Drive train vibration monitored by accelerometers
• Characteristic frequencies identify source of
  vibration
• But these are measurements of the output, effect
  or result of degradation
• The major input or cause of these parameters is
  the loads from the blades

6
Insensys Blade Load Instrumentation zxcz

• Benefits
• Realtime blade load measurement as input to
  cyclic pitch control
• Long term measurement for structural health
  monitoring
• System configuration
• 4 optical fibre strain sensors located in the
  blade root to measure flapwise and edgewise
  bending moment
• Optical signals converted to digital electronic
  data in OEM-1030 instrument located in the hub
• Measures the key input loads to the drive train

7
Insensys Data Reduction Algorythms zxcz

• For condition monitoring, large volumes of blade
  load data need to be reduced for analysis,
  display and interpretation
• 1 minute blocks of data processed to a single
  summary file
• Powerful on board DSP performs analysis
• Statistical analysis of all raw measurement
  performed
• Max, min, average
• 1F and 3F amplitude and phase

8
Insensys Blade Load Data Interpretation zxcz

• However blade load measurements can infer other
  parameters about the blades and the rotor
• Key blade parameters are calculated including
• Blade bending moment
• Blade cumulative fatigue at 4 locations around
  each blade root
• Key rotor parameters are calculated including
• Drive torque
• Resolved horizontal vertical shaft load
• Load on tower
• All derived parameters are summarised
• For local black box storage
• Or transmission to a condition
• monitoring system

9
Insensys / SKF System Integrationzxcz

• Insensys instrument located in hub
• Performs data interpretation and first stage data
  reduction
• Is polled for data like any other sensor
  connected to the iMU
• Can be retrofitted to turbines already containing
  iMU
• SKF iMU located in the nacelle
• Further data processing and onward transmission

iMU
10
Blade Service History zxcz

• Flapwise and edgewise bending moment histories
  are processed to blade load service histograms

Blade 1 EdgeWise
Blade 2 EdgeWise
Blade 3 EdgeWise
200 180 160 140 120 100 80 60 40 20
200 180 160 140 120 100 80 60 40 20
200 180 160 140 120 100 80 60 40 20
Hours at Given Load 000
0 10 20 30 40 50 60 70 80
90 100 110 120
0 10 20 30 40 50 60 70 80
90 100 110 120
0 10 20 30 40 50 60 70 80
90 100 110 120
Measured Blade Bending Moment kNm
Measured Blade Bending Moment kNm
Measured Blade Bending Moment kNm
Blade 1 FlapWise
Blade 2 FlapWise
Blade 3 FlapWise
200 180 160 140 120 100 80 60 40 20
200 180 160 140 120 100 80 60 40 20
200 180 160 140 120 100 80 60 40 20
Hours at Given Load 000
0 10 20 30 40 50 60 70 80
90 100 110 120
0 10 20 30 40 50 60 70 80
90 100 110 120
0 10 20 30 40 50 60 70 80
90 100 110 120
Measured Blade Bending Moment kNm
Measured Blade Bending Moment kNm
Measured Blade Bending Moment kNm
11
Blade Fatigue zxcz

• Rainflow counting of stress time histories
• Fatigue characteristics of blade stored in
  Insensys instrumentation
• Fatigue calculated at 4 locations around the root
  of each blade
• 4 cumulative fatigue values continually updated
  per blade to minimise data transfer and storage
  requirements

Blade 3
Blade 2
Blade 1
0 10 20 30 40 50 60 70 80
90 100
0 10 20 30 40 50 60 70 80
90 100
0 10 20 30 40 50 60 70 80
90 100
12
Rotor Drive Torquezxcz

• Calculated by resolving individual blade loads
• Time domain summary statistics generated
• Frequency components analysed
• 1F amplitude and phase provide measure of rotor
  balance
• A fully balanced rotor has zero 1F amplitude
• Phase of 1F component identifies unbalanced blade
• 3F amplitude and phase provide measure of drive
  variability in different parts of blade sweep
• Wind shear, tower shaddowing

Components at 1F and 3F
13
Resultant Rotor Offset Load

• Calculated by combining individual blade loads
• Time domain summary statistics generated
• Frequency components analysed
• 3F amplitude and phase provide measure of
  resultant offset load on drive shaft in
  differrent parts of blade sweep
• Wind shear, tower shaddowing

Main component at 3F
14
Linking Input And Ouput

• Simple to analyse cause and effect by logging all
  parameters via a single system
• Can correlate blade and rotor loads with changes
  in other turbine parameters

15
Acting on Information

• Condition monitoring
• Event driven alarms generated in iMU
• Based on threshold values stored in iMU
• Alarm flagged via email or SMS
• Alarm event information transmitted to SKF client
  in control room
• Provides the opportunity for planned intervention
  and maintenance
• Load reduction
• Improve understanding of blade load variations on
  drive train degradation
• Implement improved control algorythms to reduce
  drive train wear

16
Summary

• Blade load data is interpreted in the load
  monitoring system to generate key parameters for
  both the blades and the rotor
• Large volumes of data are compressed using time
  and frequency domain algorythms and statistical
  analysis
• Blade load monitoring system integrates with
  either new or existing SKF turbine condition
  monitoring system
• Linking the drive train input loads to the drive
  train health monitoring output parameters via
  single system
• Enabling retrofit installation
• Further data reduction and alarm generation by
  iMU in nacelle
• Enables cause and effect analysis of turbine
  degradation and performance optimisation through
  scheduled intervention and load reduction

17
AcknowledgementsHarry Timmerman, SKFFredrik
Sundquist, SKF