| 摘要: | 數值模型模擬是在傳統製造過程中實施工業4.0的關鍵部分，以加速實現成功的製造過程並最大限度地降低成本。本論文探討了數值模型模擬在商業製造過程的兩個真實案例中的實現。一種是鑄鋼正齒輪的移動感應淬火工藝，另一種是鋁合金工件的近淨形鍛造製程。該研究透過合併額外的模型參數來增強現有的數值軟體，以提高製程模擬的準確性。對於移動式感應淬火過程，模擬包括掃描速度和熱量產生等參數。該研究探討了關鍵變數（包括輪齒尺寸、掃描速度和氣隙）對硬化結果的影響。結果顯示準確度很高，在各種實驗驗證情境中預測誤差範圍為 3.02% 至 4.05%。多變量非線性迴歸分析強調了製程參數對淬火品質的顯著影響，特別是掃描速度的主導作用。研究結果表明，降低掃描速度、氣隙和側面長度可以提高硬化質量，從而獲得更長的硬化側面、更深的硬化深度和最小化邊緣效應。在鋁合金的近淨成形鍛造過程中，根據 AA7050 減震器、AA7075 曲線切割訂書機和 AA6082 輪的實驗結果評估了附加參數，如活化能、Zener-Hollomon 參數和加工圖。研究表明，這些參數對於準確預測鍛造產品的鍛造缺陷和微觀結構演變至關重要。綜合分析需要考慮所有三個參數，以確保更可靠的預測和製程最佳化。總體而言，這項研究強調了增強數值模型在改進製造流程方面的重要性，為優化生產和確保高品質結果提供了寶貴的見解。;Numerical model simulation is a pivotal part of implementing Industry 4.0 in traditional manufacturing processes to accelerate the realization of successful manufacturing processes and minimize costs. This dissertation explores the implementation of numerical model simulation on two real cases of commercial manufacturing processes. One is a mobile induction hardening process on a cast steel spur gear, and the other is a near-net-shape forgings process on an aluminum alloys workpiece. The study enhances existing numerical software by incorporating additional model parameters to improve the accuracy of process simulations. For the mobile induction hardening process, the simulation includes parameters such as scanning speed and heat generation. The research examines the impact of key variables, including gear tooth size, scanning speed, and air gap, on hardening outcomes. Results indicate high accuracy, with prediction errors ranging from 3.02% to 4.05% in various experimental validation scenarios. Multivariable nonlinear regression analysis highlights the significant influence of process parameters on hardening quality, particularly the dominant effect of scanning speed. Findings suggest that reducing scanning speed, air gap, and flank length enhances hardening quality, resulting in a longer hardened flank, deeper hardening depth, and minimized edge effects. In the near-net-shape forging process of aluminum alloys, additional parameters such as activation energy, the Zener-Hollomon parameter, and processing maps were evaluated against experimental results for AA7050 shock absorbers, AA7075 curve cutter staplers, and AA6082 wheel. The study demonstrates that these parameters are essential for accurately predicting forging defects and microstructure evolution in forged products. Comprehensive analysis requires considering all three parameters to ensure a more reliable prediction and process optimization. Overall, this research underscores the importance of enhanced numerical models in improving manufacturing processes, providing valuable insights for optimizing production and ensuring high-quality outcomes. |
