Early fault prediction of wind turbine gearbox based on temperature measurement

Gearbox fault is widely received as the leading issue of wind turbine, and early fault prediction of gearbox is very meaningful for ensuring reliable running and reducing maintenance costs. For this reason, the relation between the gearbox temperature and the gearbox faults was researched and a new method for early fault prediction of wind turbine gearbox was presented. By this prediction method, running condition of gearbox could be monitored by comparing the real time value of a detecting function which can indicate the deviation between the actual running condition of gearbox and the predicted condition of gearbox with the chosen threshold. And this can be shown in the warning curve clearly. This prediction method avoids identifying the nominal model, relaxes the requirements to model precision and improves its robustness through the modified vector-valued function. Theoretical analysis and simulation results verify the effectiveness and the robustness of the proposed method.