Co-authors Klaus Haselgruber Franz Langmayr Simon Watson Contents - - PowerPoint PPT Presentation

A Practical Approach to the Use of SCADA Data for Optimised Wind Turbine Condition Based Maintenance

Presenter Christopher Gray Co-authors Klaus Haselgruber Franz Langmayr Simon Watson

• Motivation
• Failure Mode Assessment
• SCADA Data Analysis
• System Response
• Physics of Failure
• Workflow
• Summary

Contents

EWEA Offshore, 1 December 2011 2

Motivation

• Significant contribution from several systems, range of failure modes
• Monitoring system required to assess overall turbine health status
• Specialized CMS for each system not financially viable
• SCADA data readily available, high level of detail in modern turbines

3 EWEA Offshore, 1 December 2011 Source: EWEC 2010 “Methodology and Results of the Reliawind Reliability Field Study“

Failure Mode Assessment

• System analysis  identify failure root causes, damage drivers and model parameters
• Expert input to identify key issues
• Input for model development & fault diagnostic algorithms  link to available data

4 EWEA Offshore, 1 December 2011

Limitations of SCADA 10-minute Logs

5

• Information loss for dynamic signals (Nyquist theorem)  missing features
• Non-linear damage kinetics, error in remaining life estimates
• Improvement: on-line analysis, improve data aggregation

EWEA Offshore, 1 December 2011

Log N Log S

Shock Loads

S-N Curve (simplified example)

Signal Noise

Blade Pitch Activity

10-min log 10-min log ….

SCADA Data Analysis: Quick Wins

• Large number of turbines, large data volumes  automated analysis and statements
• High data quality required, check for errors (sensors, signal form, drift, cross checks)
• System health check  simple analysis techniques (statistics, correlations, trends)

6 EWEA Offshore, 1 December 2011

Early Warning of Overheating, Turbine 4

System Response Modelling & Monitoring

7

• Library of models for various systems, as analysed in Failure Mode Assessment
• On-line comparison of measured vs expected behaviour

Yaw Operation Pitch Operation Gbox Oil Temperature Generator Winding Power Performance Rotor Aerodynamics Bearing Temperatures

Part load, elevated temperature Controller alarm limit Early warning

November 2011 | 7

Prognostics, Physics of Failure

• SCADA log defines load history for various systems and components
• Models created to describe the relationship between load and damage accumulation
• Transfer functions generated where damage driver is not directly measured

EWEA Offshore, 1 December 2011

Generator Winding

Thermal Aging

Gearbox Bearing

High Cycle Fatigue

Rotor Blade

Laminate De-bonding

Drive Shaft Flange

Fretting

Yaw Drive Ring Gear

Adhesive Wear

Pitch Motor Housing

High Cycle Fatigue

Foundation

High Cycle Fatigue

Damage driving events

99.9% 99.9%

Gearbox Temperature Residual [°C] Active Power Residual [kW]

Statistics, Reporting, Actions

9

PHYSICS STATISTICS SOFTWARE

EWEA Offshore, 1 December 2011

Statistics, Reporting, Actions

10

ALARMS

• Response anomaly
• Performance deficit
• High failure probability

ACTIONS

• Turbine Inspection
• Spare parts order
• Maintenance schedule

DIAGNOSIS

• Expert system
• Failure mode identification

Feedback

EWEA Offshore, 1 December 2011

Windpark: Demo 1 Date From: 01.08.2010 Date To: 31.08.2010

Conclusions

• Holistic, complete life cycle approach required for WEC reliability
• SCADA data analysis a valuable & cost effective technique for condition monitoring
• Keys to Success
• Automated

data validation, analysis, reporting

• Comprehensive

all components and relevant failure modes

• Independent

all WEC types

• Scalable

single monitoring solution

• Effective

integration into asset management program

11 EWEA Offshore, 1 December 2011

Thank You!