板材沖壓彎曲胚料展開之預測研究

摘要本研究探討板材沖壓彎曲胚料展開之預測，改善胚料展開設計之準確與效率上的問題，論文內容是以田口法規劃板材沖壓彎曲之相關因子，再依規劃之安排進行有限元素模擬分析及測量各變化後之尺寸數據，並將其所得數據結果藉由類神經網路學習並預測，再以彎曲實驗進行驗證，讓展開之數據更具有精確性及效率，有效解決板金設計上的精度及效率之問題，且可將此研究之結果提供板材沖壓彎曲設計的依據及相關系統引用開發。研究結果發現，以針對板材沖壓彎曲模擬分析，其計算展開的數據，與實驗比較證實了有限元素模擬分析可得到精準的展開數據，再配合類神經網路之應用，可滿足效率問題，得到令人滿意的成效。希望藉由本論文之板材沖壓彎曲胚料展開之分析與預測途徑，得到一套符合目前沖壓彎曲加工的需求的模式，並期望未來可將此沖壓彎曲胚料展開方法整合於CAD/CAM相關軟體或系統，產生更具規模的效益，提昇板材沖壓彎曲加工之水準及競爭力。

關鍵字

沖壓彎曲；胚料展開；有限元素法；類神經網路

並列摘要

Abstract Predicting the expansion of plate press bending billet and improving the accuracy and efficiency of billet expansion were studied in the research .It uses the factor about plate press bending with Taguchi method in the thesis. The simulation analysis of finite element and the measurement of the variational size were proceed in accordance with the plan .The data is acquired to learn and predict by the neural network .It proves the data through the bending experiment further . The design let the expanded data be more accurate and rapid in order to solve the problem of accuracy and efficiency effectively in designing sheet metal . The basis of plate press bending design and the development and quotation of related system are proved by research consequence. For the simulation analysis of plate press bending, the expanded data is compared with the experiment and proved that the data can be obtained precisely from the simulation analysis of finite element which is used with the application of neural network to slove the problem of efficiency and obtain the satisfactory.. It is to be hoped that the press bending model can be obtained by analyzing and predicting the expansion of plate press bending billet to fit with the requirements . In order to make quite a big-scale benefit and improve the level and competitiveness of the press bending model, we look forward to integrating the expansion of plate press bending billet with CAD/CAM software or system in the future.