SOM Based Condition Monitoring of Wind Turbines

1
SOM Based Condition Monitoring of Wind Turbines

• Simon Catmull
• Project Analyst
• 17/03/11

2
CONTENTS

1. Introduction
2. Motivation
3. Method
4. Results
5. Conclusion

3
MOTIVATION FOR CONDITION MONITORING
4
MOTIVATION

• In an ideal world...
• No manual resets.
• No component failures.
• Longer service interval.

All this results in downtime and lost revenue!
5
CONDITION BASED MAINTENANCE
Condition Monitoring
WTG Warning
WTG Alarm
6
TURBINE INSTRUMENTATION
Anemometers
Thermocouples
Voltage Probes
Current Transformers
Vibration Sensor
Accelerometers
Rotary Encoders
Strain Gauges
7
RESOURCE INTENSIVE ANALYSIS
8
KOHONENS SELF ORGANISING MAP (SOM)

• Fantastic pattern detectors.

• A capability that can be mimicked.

9
PRINCIPLE OF OPERATION

10
RESULTS SENSOR ERROR
11
RESULTS GENERATOR TEMPERATURE FAULT
4 Years
7 days
6 Hours
12
RESULTS GENERATOR TEMPERATURE FAULT
13
RESULTS REACTIVE POWER COMPENSATION
25 Hours
3 Years
6 Hours
14
CONCLUSIONS

• The Self-Organising Map is a broadly applicable
  tool to the problem of condition monitoring of
  wind turbines.
• Although it might indicate that something is
  wrong it does not show what is wrong. However,
  there are ways in which this can be done.
• Accuracy of the distance as an indicator is
  improved by grouping relevant parameters
  together.
• The distance from the SOM is a measure of novel
  behaviour and has proven to be successful in RES
  investigations of operational wind turbine data.

15
FUTURE WORK
16
QUESTIONS?
Simon.Catmull_at_res-offshore.com