Wind Turbine Main Bearing Fault Prognosis Based Solely on SCADA Data.

Area of Science:

• Engineering
• Renewable Energy Systems
• Predictive Maintenance

Background:

• Main bearing failures are a critical issue in the wind industry, leading to high repair costs and significant downtime.
• Existing prognosis methods often require specialized sensors or extensive fault data, limiting their applicability.

Purpose of the Study:

• To develop a data-based methodology for wind turbine main bearing fault prognosis.
• To demonstrate the feasibility of using only readily available supervisory control and data acquisition (SCADA) data for fault prediction.

Main Methods:

• A data-driven approach utilizing historical SCADA data from wind turbines.
• The methodology requires only healthy operational data for training and application.
• The algorithm is designed to function effectively under diverse operating and environmental conditions.

Main Results:

• The proposed method successfully predicts main bearing failures months in advance using only SCADA data.
• The system was validated on a real-world wind farm with 12 turbines.
• The prognostic system enables proactive maintenance planning for wind turbine operators.

Conclusions:

• Advanced prognostic systems based solely on SCADA data are viable for predicting wind turbine failures.
• This approach enhances wind turbine reliability and availability by enabling early detection and planned interventions.
• The methodology offers a cost-effective and broadly applicable solution for wind farm maintenance.