Wind turbine state evaluation and early-warning method and system based on similarity statistics

The invention provides a wind turbine state evaluation and early-warning method and system based on similarity statistics. The method comprises the following steps: step 1, generating a plurality of normal state models and a safety datum line of a wind turbine according to historical operating data of the wind turbine; step 2, acquiring real-time operating data of the wind turbine, comparing the real-time operating data of the wind turbine with the normal state models to determine the similarity between the real-time operating data of the wind turbine and the normal state models, and comparing the real-time operating data of the wind turbine with the safety datum line of the wind turbine to warn when the real-time operating data are abnormal; and step 3, analyzing the real-time operating data to evaluate the fault of the wind turbine. According to the embodiment of the invention, the normal state models can be constructed through the historical operating data of the wind turbine, so that warning and evaluatio