本研究探討風力發電機驅動鏈架構中之行星齒輪的故障特徵分析與壽命預測，動軸系行星齒輪系於風力發電機驅動鏈之應用極為重要，以於狹小空間傳遞高動力負載。然其振動行為與定軸系齒輪相較更為複雜，欲掌握其故障特徵之難度亦相對提高。本研究以行星齒輪不同劣化程度之振動時域訊號進行分析，探索齒輪故障特徵頻率幅值的變動狀況、包絡分析(Envelope Analysis)之影響、及邊頻的差異。行星齒輪於劣化狀態下，就其振動時域訊號進行頻譜分析時，可發現其故障特徵主要以齒輪嚙合頻率兩側間距為行星齒輪故障特徵頻率之邊頻呈現，其幅值隨劣化程度加劇而成長。而經由包絡分析則能於頻譜中有效追蹤其齒輪故障特徵頻率，若齒輪之劣化程度加劇，行星齒輪故障特徵頻率幅值隨之成長，且其倍頻及邊頻皆亦可清晰呈現並隨持續之劣化而增長。經由訊號處理萃取之故障特徵於不同劣化狀態下之差異，可作為建立特徵指標之重要參考依據，經由特徵指標之運用可據以進行剩餘使用壽命之預測，有效監控風力發電機機艙驅動鏈之行星齒輪系操作狀態，以供風機維運之重要參考。

This study explores the analysis of the planetary gear fault characteristics and life prediction in wind turbine drivetrain. The application of the moving shaft planetary gear train in wind turbine drivetrain is extremely important to transmit high power loads in a small space. However, its vibration behavior is much more complicated than that of the fixed-axis gear train, and it is relatively difficult to grasp its fault characteristics. This study analyzes the time-domain vibration signals of the planetary gears with different degrees of deterioration, and explores the changes in the amplitude of the gear fault characteristic frequency, the influence of envelope analysis, and the difference in sideband frequencies. When the planetary gear degrades, the time-domain vibration signal of the planetary gear is analyzed by spectral analysis. It can be found that the fault characteristic is mainly the appearance of the sidebands around the gear meshing frequency, with the spacing distance being the planetary gear fault characteristic frequency, and the associated amplitudes grow with the degree of deterioration. Through envelope analysis, the fault characteristic frequency of the gear can be effectively tracked in the spectrum. If the deterioration of the gear is intensified, the amplitude of the fault characteristic frequency of planetary gear will grow accordingly, and its harmonics and sideband frequencies can also be clearly observed and grow with the proceeding deterioration. The differences in the fault characteristics extracted by signal processing under different degradation states can be used as an important reference basis for establishing characteristic indexes. Subsequently, the remaining useful life can be estimated such that the operating conditions of the planetary gear train of the drivetrain can be effectively monitored and serve as important reference for wind turbine operation and maintenance.