Condition monitoring of wind turbine gearbox using electrical signatures

For reducing effect of energy production on the climate change, renewable energy sources play a very important role in the global energy mix. Wind energy has the largest share in the renewable energy technologies with approximately 433 GW of globally installed capacity as of 2015. Share of power production through Wind Turbine (WT) technique is growing day by day and there is an urgent need to control the cost of operation and maintenance of these systems. For early detection of failures or faults we use Condition monitoring (CM) technique for WT in order to maximize productivity and minimize downtime. In this paper we present a method for CM of WT gearbox using vibration data. We have used Hilbert-Huang transform (HHT), Empirical mode decomposition (EMD) technique for feature extraction and Neural network train tool which is based on MLP (Multilayer Perceptron) model for classification. The focus is on the gearbox, as it is typically one of the most crucial components in terms of long average down time and high failure rates. The opportunities and challenges are identified to help in conducting future research in enhancing the ability and accuracy of CM and prognosis systems for WT gearboxes.