由於齒輪加工情形複雜,製造與裝配誤差在所難免,這些誤差皆會影響高轉速下齒輪的振動情形,但產生振動因素太多,僅由最終之振動情形找出誤差成因卻有一定的難度。 本文主要以齒輪對為研究對象,利用線上齒輪嚙合振動檢測系統。以傅利葉頻譜分析擷取訊號特徵,由於齒輪嚙合時,其嚙合頻與其倍頻之間、嚙合頻與其邊頻之間、嚙合頻倍頻與其邊頻之間有一定之關連性,所以可應用頻譜分析於齒輪振動檢測技術上;設定齒輪對運轉時不同狀況的模擬,提取振動訊號進行判斷,觀察在不同條件下齒輪對運轉之情形,如齒數不同時對背隙的變化、不同齒數之齒輪增加小鐵屑的變化、不同齒數齒面上增加小質量的變化或具偏擺,歸納實驗所得到的結果並進行討論,以了解各種製造及組裝誤差之成因。
Due to the complexity of gear manufacturing process, the machining and assembly errors are two main factors in gear vibration and the associated noise. This thesis focuses on the investigation of gears vibration applying signal processing techniques. The signals play an important role in detecting the gears faults. Owing to the relations among spectral characteristics of the gears, rotational harmonics, the Gear Mesh Frequency(GMF), and the GMF harmonics and their sidebands, spectrum analysis can be applied in the gear vibration detection technologies. Different conditions are set in the experiment to study the vibrations. In the eperiment, effects of different teeth numbers of the gears, backlash, small iron shaving or mass attached on the gears surface, are studied by the FFT spectrum.