Abstract: Based on the relevant operating data from a wind farm in Northwest China, the reliability of wind turbines is studied. Considering the correlation and redundancy among operating data, Principal Component Analysis (PCA) method was used to reduce dimensionality, and some key reliability indicators were selected. Subsequently, a reliability evaluation model is constructed utilizing fuzzy theory in conjunction with the selected key indicators from operational data. A subset of wind turbine operating data is selected for model verification. The research results show that the PCA method extracts the principal component with a cumulative variance contribution rate of 87.585%, which can comprehensively express the reliability information of wind turbines. In the evaluation of a single reliability index, the availability of B02 is as high as 98%, but its total power generation is the lowest. Conversely, the availability of A04 is the lowest in the available area yet relatively higher power generation, indicating potential discrepancies when assessing reliability based on single reliability index. All indicators of A05 and B03 are high, the actual operation state of which is good, and the power generation is high, indicating that their comprehensive reliability is higher, which is consistent with the research results. Therefore, the fuzzy theoretical evaluation model (FSEM) of wind turbine reliability established based on PCA and fuzzy theory is in line with the actual operating state and has guiding significance for quantitatively evaluating the reliability of wind turbines.

Key words: wind turbine, PCA, reliability, FESM