| 摘要: | 雙螺桿空氣壓縮機為工業領域中廣泛應用之關鍵設備，憑藉其穩定性、可靠性及高效率之性能，在各類製程與氣體壓縮應用中均占有重要地位。隨著產業對壓縮機性能提升的需求日益增加，轉子導程設計之優化已成為提升雙螺桿壓縮機效能的重要課題。尤其是採用多段變導程轉子設計，能有效調整壓縮腔體的容積變化速率，優化氣體閉合行為並降低內部洩漏，進而提升整體性能。基於此，本研究針對雙螺桿空氣壓縮機轉子進行多段變導程曲線之設計，並探討其對壓縮機性能之影響。在齒廓型式、轉子螺旋部長度及包封角相同的條件下，設計多種導程變化曲線，並分析其對轉子幾何特性之差異。接著，將導程曲線分為三段，並依據定導程工作角度，分別以吸氣角、壓縮角及排氣角作為設計基準，作為三段變導程曲線之參數設定依據。所設計之導程曲線將經幾何特徵分析及效率模組計算，找出性能表現較佳的變導程趨勢。確立設計趨勢後，進一步將導程參數以定導程之導程值及包封角作為基準值進行無因次化，並結合均勻實驗法及 ISIGHT 軟體內建之徑向基函數（RBF）建立預測模型方程式。有了預測模型後，利用 Mathematica 數值軟體之 Global Optimization 模組設定目標值，並以最小平方法進行優化，使目標值與預測模型數據間的差異降至最小。最後，驗證所獲得之最佳化變導程參數，並探討其幾何特徵及於不同轉速下之容積效率與等熵效率表現，作為轉子導程優化設計之重要參考
依據。;Twin-screw air compressors are essential equipment widely used in various industrial applications, owing to their stability, reliability, and high efficiency. With the increasing demand for improved compressor performance, the
optimization of rotor lead curve design has become a key factor in enhancing the efficiency of twin-screw compressors. In particular, the application of multi-stage
variable lead designs enables effective adjustment of the chamber volume variation rate, optimizes gas sealing behavior, and reduces internal leakage, thereby improving overall performance. Based on this, this study focuses on the design of multi-stage variable lead curves for the rotors of twin-screw air compressors and investigates their effects on compressor performance. Under the condition of identical rotor profiles, helical section lengths, and wrap angles, multiple types of variable lead curves are designed and analyzed to identify differences in rotor geometric characteristics. Furthermore, the lead curve is divided into three segments, with the suction angle, compression angle, and discharge angle defined as reference values based on the constant lead working angles, serving as the basis for parameter setting of the three-stage variable lead curves. The designed lead curves are then evaluated through geometric feature analysis and an efficiency module to determine the trends that yield better performance. After establishing the design trend, the lead curve parameters are
normalized using the constant lead value and wrap angle as the baseline. A predictive model is then constructed using a radial basis function (RBF) approximation method embedded in the ISIGHT optimization software, combined with a uniform experimental design. Once the predictive model is obtained, the Global Optimization module in Mathematica is used to set target values and perform optimization using the least squares method to minimize the difference between the target values and the model outputs. Finally, the optimized
variable lead parameters are validated, and their geometric features as well as the volumetric efficiency and isentropic efficiency at various rotational speeds are analyzed. The results serve as an important reference for future rotor lead curve optimization design. |
