應用振動診斷法對軸承破壞之偵測

Bearing Fault Detection Using Vibrations Diagnosis

設備診斷技術是防止設備故障及確認故障源的利器，然而國內對於設備診斷技術的實施並不普遍。本文所討論的案例，是應用振動診斷技術於國內一家大型造紙廠的抄紙機上。抄紙機長約23.5公尺，高約7.8公尺（抄紙機煙囪的高度不計）。振動量測過程中使用雙頻道頻譜分析儀及加速規，對選取的9個量測點，進行水平，垂直及軸向等三個方向的振動量測。另一方面，並將量測的振動訊號，經由8頻道的訊號記錄 器儲存於磁帶中，以便於實驗室中做詳細的分析。由於量測的振動訊號中有軸承破壞的現象，無法直接分析，因此對於振動訊號先行使用包絡線處理後，再進行分析。由於軸承破壞的訊號很明顯，因此進行包絡線處理時不需經過前置濾波處理（傳統包絡線處理，需先進行前置濾波處理〉。經分析後發現該機器的自動對心軸承的外環有明顯的損壞現象。同時，抄紙機上的毛毯也已損壞。當更換損壞的軸承之後，抄紙機振動的現象皆已消除。然而，由於新毛毯的選用不當，使得毛毯的低頻振動卻仍然存在，並且振動頻率由0,2Hz升高至0.5Hz。整體而言，抄紙機的生產速度已由800m/min提高至1050m/min，並可提昇至1200m/min，而不會發生斷紙的現象。

Machinery diagnosis technique is a powerful tool in preventing machine failure and identifying sources of machine failure. However, this technique has not been widely accepted in Taiwan. The case discussed in this study deals with fault detection of a mill machine in a paper manufacturer using vibration diagnosis technique. The paper mill machine is about 23.5 meters long and 7.8 meters high excluding the chimney which is at the top of the paper machine. Nine measuring points were selected at various locations on the machine to measure the horizontal, vertical and axial vibrational signals. Dual-channel frequency analyzer with PCB accelerometers were used to measure and analyze vibration signals from the machine. In addition, an 8-channel instrumentation cassette recorder was used to record the vibration signals which were played back later for detailed analysis. Enveloping technique was applied to the signal analysis process due to the fact that signals were difficult to analyze directly. The measured data showed that the pre-filtering of the signal was not necessary as long as the vibration signals were high. The deterioration of outer race of one self-alignment spherical bearing was clearly seen from the frequency domain signal after enveloping technique was conducted. In addition, the low frequency response from felt deterioration was also detected. Severe vibration which caused machine frame vibration and broken papers disappeared after replacing the self-alignment spherical bearing. However, the low frequency vibration from the felt still existed due to wrong choice of new felt. It was also found that the felt frequency was changed from 0.2 Hz to 0.5 Hz after the felt change. Nevertheless, the paper mill machine has been operating at 1050 m/min with the potential to increase to 1200 m/min along with very rare chance of broken papers. The validity of using vibration analysis to detect bearing fault of a complex machine is presented in this case study. Signal pre-filtering is not always necessary when enveloping technique is employed. In addition, it is also proven that the felt frequency depends on the linen direction when the felt is knitted.