Intelligent system for fault detection in wind turbines gearbox

Abstract

New generations of turbines have lower repair and maintenance costs than the previous generation. This is justified by the development of new components and materials. As the power of newer turbines is usually substantially larger, it is possible to get an economy of scale and lower maintenance costs per kW of rated power. This is simply because it is not needed to service a large turbine more often than a small one. New methods of earlier detection of faults are needed. The use of all information from SCADA (Supervisory Control and Data Acquisition) system can be useful, but it is necessary to develop tools to deal with bigger amount of information. Neural networks can help and turn possible new maintenance and operation schemes.