透過您的圖書館登入
IP:3.235.243.45
  • 學位論文

建立高壓反應器軸承振動檢測手法提升預保效能之研究

A Study on the Bearing Vibration Diagnosis System of Autoclave Reactors for Higher Predictive Maintenance Efficiency

指導教授 : 魏乃捷
共同指導教授 : 薄喬萍(Chiao-Pin Bao)

摘要


利用振動監控系統監視設備運轉狀況以預知保養為最有效率,此技術發展歷史較為悠久,最常見該技術使用於;如石化廠、鋼鐵廠及造紙廠等。振動監控診斷技術往往為週期性定時監控診斷,不僅需要專業儀器設備及人員具備判讀能力外,且對於非轉動設備本身所導致之突發狀況影響到設備之運轉,亦無法線上即時監控診斷。本研究對象某石化廠反應器因製程操作考量,反應器本體爐壁厚度約13公分,除了常見診斷手法如;振動頻譜分析檢測或線上振動趨勢監測外,本研究因顧及反應器本體先天設計,再進行振動檢測同時可能因為爐壁厚度導致振動傳導過程衰減。因此本文研究對象為高壓反應器之振動異常檢測,以三種手法為: 1.頻譜分析、2.線上趨勢監控、3.金屬燒檢。本研以上三種診斷手法針對反應器之軸承建立使用壽命,藉由即時線上趨勢振動總值(Over all)監控外,並配合每周一次定期振動頻譜分析診斷。若由即時線上監控系統得知振動總值出現增幅之趨勢,或是反應器操作參數異常時則會增加振動頻譜量測次數。本研究除了借助頻譜分析針對軸承軸承損壞特徵頻率以及軸承元件之損壞外,則會檢測剛上線使用之正常軸承訊號及軸承故障訊號,利用瀑布圖呈現,其目的為了將正常軸承與使用一段時間之軸承進行頻譜比對,助於往後分析之結果判斷,透過既有設備技術可提供預測維護基礎更能達到預知保養最終目的與預知保養之精神。

並列摘要


Regarding the vibration monitoring system to equipment operational control, predictive maintenance enjoys the highest efficiency. This long-developed technique is often utilized in factories such as petrochemical plant, steel mill and paper mill. In most cases, the vibration monitoring diagnosis technique involves regular monitoring diagnosis on a periodic basis. This technique not only necessitates professional instruments and equipment, but also requires operators' expertise in identifying the situation. Furthermore, this technique cannot execute real-time monitoring diagnosis when sudden occurrences caused by non-rotating equipment itself influence equipment operation. In this research, the research subject (i.e. a petrochemical plant) has reactors with wall thickness in 130mm due to its special manufacturing process. Instead of using common diagnostic techniques, such as the detection based on vibration spectrum analysis and on vibration tendency, the research subject adopts the following three approaches to detect vibration anomalies for autoclave reactors because the original design of reactors and the 130mm-thick wall may reduce vibration in the conductive process while performing vibration detection: 1. spectrum analysis;2. on-line vibration tendency monitoring;3. metal burning detection. Based on these diagnosis approaches, the research established the useful lifetime for reactor bearings by monitoring the real-time overall values of vibration tendency in combination with regular vibration spectrum analysis and diagnosis on a weekly basis. When the real-time monitoring system shows the overall vibration value has an incremental tendency, or when the operational parameters of reactors show irregular values, additional times will be taken to measure vibration spectrum. In addition to adopting spectrum analysis to monitor the damages of bearing components and the characteristic frequency of damaged bearings, this research also detects the normal/breakdown signals of new/defected bearings. As shown in the waterfall curve, we can compare the spectrum between normal bearings and the bearings used for a period of time. The comparison results, coupled with the existing equipment and techniques, can provide meaningful insights for understanding the analysis results in the future and serve as the basis for predictive maintenance, and further achieve the ultimate goal and spirit of predictive maintenance.

參考文獻


1.王皇評(2015)。晶圓廠自動倉儲裝載臂異常判斷之研究。逢甲大學資訊電機工程碩士在職專班碩士論文。
2.全全科技有限公司(2009)。轉動機械預知保養。
3.邱祈民(2017)。風力發電機軸承損壞訊號分析。國立中央大學機械工程學系碩士論文4. 洋洋實業有限公司(1995)。預知保養診斷分析巧。
4.林朝乾(2004)。應用振動量測於紙廠設備之預知保養。國立屏東科技大學機械工程系碩士班碩士學位論文。
5.邱高興(2012)。SKF振動Training。

延伸閱讀