本篇研究主要探討端铣刀在切削過程因磨損，進而造成銑床轉軸的加速力與端铣刀切削力頻率訊號的改變，藉由分析此頻率的變化去得知端铣刀受損前後的特性。在此 研究將使用時頻分析-黃鍔 (Hilbert-Huang Transform, HHT)進行銑床轉軸的加速力以 及端铣刀切削力頻率的分析，找出端铣刀受損前後的特性，並再將分析的結果與傅立 葉轉換(Fourier Transform, FT)做比較。 端铣刀的切削力之量測方法在本篇研究裡，其運用動力計(dynamometer)進行 測， 量測結果發現端铣刀在切削過程中因磨損而改變端铣刀尖端之幾何 狀，造成磨損前 後切削力的改變。而由HHT 以及FT 分析銑床轉軸 速力頻率的變化，皆發現端铣刀的磨損(tool fault)並不會對銑床轉軸加速力頻率造成明顯的影響。但若端铣刀只有輕微 的受損(tool wear)，則會對銑床轉軸加速力頻率造成明顯的影響，並可用加速力頻率的變化來診斷端铣刀是否有輕微受損之狀況(tool wear)。且藉由分析加速力頻率在x 軸、y 軸以及z 軸所產生的變化，其顯示加速力頻率在x 軸(垂直於進给速度的向量)上的改變最為明顯。 時頻分析-黃鍔法(HHT)優於調和分析的好處在於HHT能分解端铣刀磨損特性，並可辨識這類非穩態及非線性之訊號，但若使用調和分析則無法觀察出端铣刀磨損訊號之真實特性。 The presented study contains research related to cutter tool fault detection in machining process. Acceleration and cutting force signals were subjected to analysis in order to find characteristic feature for cutter tool fault detection. The new time – series analysis technique Hilbert-Huang Transform (HHT) was applied to find a recognition pattern for successful tool fault or tool wear recognition. The results were compared to Fourier Transform (FT) which has been massively used to monitor cutting processes. The cutting force signal obtained from dynamometer measurements has shown its good characteristic in order to recognize a cutter tool geometric change which is actually a cutter tool fault. The acceleration signal has shown to be insensitive to cutter tool fault detection in both FT and HHT investigations. The sensitivity of both signals has shown to be a good tool for tool wear diagnostic. The sensitivity of different components has shown that axis which is perpendicular to feed rate is the most suitable for cutter tool conditions monitoring. The advantage of the HHT apart from processing of non-stationary and non-linear signal is absence of harmonics which has been very confusing in cutter tool fault or tool wear recognitions.
