Abstract:

There are disadvantages in the diagnotic method of wind rotor imbalance fault based on the vibration signal, which are in great demands for sensors, of high cost and poor reliability. Matlab and Simulink are used to build the models of units under normal and unbalanced conditions respectively. Taken the second wind speed data measured in a real wind farm as simulation input, output power is obtained. Results of frequency domain analysis indicates that output power signals of wind rotor includes first frequency component during imbalance fault. After empirical mode decomposition (EMD), output power contains different energy values in different frequency bands. The amplitude of first frequency component and the EMD energy entropies are used as the evidence of fault. Relationship between fault severity and two evidences mentioned above are obtained by numerical fitting ，which realizes the fault diagnosis without requirements for expensive sensors.

Key words:

wind power generation, wind rotor, imbalance fault, fault diagnosis, numerical simulation