織造系統於運動狀態下織針磨耗之評估研究

Evaluation of the needle wear under motion conditions for a weaving system

黎文龍(Wenlung Li) ； 廖承瑋(Chengwei Liao)

《中華民國振動與噪音工程學會論文集》 2018 (2018/06) Pp. 48-53

摘要

紡織產業為台灣貿易順差貢獻了60幾個年頭，時至今日仍維持在產業類別出口貢獻度前10名，然而面對未來競爭產業需要升級，以順利銜接下一個工業世代，其中設備智慧化至關重要。本報告針對圓編針織機中的織造系統智慧化做前期準備特別是對於運動狀態下的織造訊號擷取，利用統計手法對時域訊號做平均數（mean）、標準差（Standard Deviation）、偏態（Skewness）及峰度（Kurtosis）的特徵淬取，並透過離散傅立葉（DFT）進行頻域特徵的轉化，經實驗組與對照組的特徵數值比對，發現織造系統於中低轉速下，織針運動的頻域基頻有高達3到7倍的強弱差異，其次是中低轉速時的時域偏態特徵為1.7倍。本報告經綜合效率與後續量產硬體成本考量，建議未來織造系統智慧化時，選用即時的時域統計量與DFT為主要訊號因子，將更能貼近織造系統磨耗程度之預測與判別。

關鍵字

圓編針織機 ； 旋轉 ； 磨耗

並列摘要

The textile industry has contributed to Taiwan's economics for more than 60 years. It still maintains the top 10 on the list of the export goods. However, anticipating the future competition, the industry needs to upgrade or add more values to textile machines by applying AI to them. This report is a pilot study for circular knitting weaving systems that are to include AI. The report mainly extracts the motion signals from the knitting needles at work. In addition, the feature statistics which are extracted from the time domain like means, standard deviations, skewness and kurtoses are adopted in the report, in connected with the DFT of the frequency domain. The comparison between feature values of the experimental and control groups shows that, at the low-medium running speed tends to exhibit more profound signals. It may reveal values as high as between 3 to 7 times to others. Considering the overall efficiency and subsequent costs of the hardware for mass-production, this report recommends to choose the more real-time statistics of time domain. And, the DFT signals of the weaving systems may be applied for predicting the system conditions of the weaving machines.