Please use this identifier to cite or link to this item: http://hdl.handle.net/11455/2092
標題: 疊層金屬複合板材鑽削品質監測
Monitoring on Metal Matrix Composites Drilling Process
作者: 吳文祺
Wu, Wen Chi
關鍵字: hole quality;孔品質;drilling;monitoring;鑽削;監測
出版社: 機械工程學系
摘要: 
由於結構強度的需求與減少燃料使用與載重的原因,使得在材料上的選擇有了重大的改變。擁有高強度、低重量、耐蝕性高的新合金與複合材料大量地取代了傳統的鋼材。其中,在航空方面多層金屬複合材料及高強度合金之重量/強度比(weight/strength)較傳統的航空材料小,為目前航空工業所意屬之材料。而飛行器的結構大多以鉚釘接合方式將各部為接合,因此在多層金屬疊層材料的鑽削中如何顧慮不同材質加工以維持孔徑品質,進而確保結構的壽命與使用上的安全性則為重要的議題。
而為符合自動化生產與積極性預防的觀念,本研究朝線上監測的方向著手。選擇特定切削參數進行實驗並擷取切削力訊號,觀察各疊層材料間毛邊生長變化與切削力之間的關係。並以實驗所得結果進行類神經網路訓練,以建立鑽削毛邊高度與刀腹磨耗量預測網路。
實驗結果可從切削力訊號中建立出鑽削品質劣化的監測指標,而類神經網路訓練的結果,其誤差函數平均值在10﹪以下。經由驗證實驗,證實可由監測指標成功的完成監測鑽削品質的工作。而在類神經網路預測的結果上,對於毛邊高度預測亦有不錯的效果。

For structure strength needing and carry more payload with lower fuel consumption, there are critical changes in material choice. The new material and composites which have greater strength, lower weight and corrosion resistant replace the conventional material. In the aerospace industry, the multi-layered metal material are concerned at present. And the aircraft structure is conjunction with rivets. Hence, how to maintain the drilling hole quality of the multi- layered material for ensuring structure life and safety become the important topic.
Conforming with automatic process and motivational prevention, the research is processed as on-line monitoring, choice specific parameter for cutting experiments and collecting cutting force signal. To observe the relationship between interlayer burr height of layered material and cutting force. Then we use experiment results for neural network training and build up the prediction model of burr height and flank wear of drill.
Experiment results, we can establish the index of monitoring for interior hole quality from the change of cutting force. And the predicting results of artificial neural network(ANN) , the average of error function is under 10﹪. In exam experiment, verify that the index of drilling monitoring can successfully approach the quality monitoring. In addition, the model of ANN has satisfactory result in burr height predicting.
URI: http://hdl.handle.net/11455/2092
Appears in Collections:機械工程學系所

Show full item record
 

Google ScholarTM

Check


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.