逆解法用於幾何外型設計之研究.docVIP

逆解法用於幾何外型設計之研究 Study on geometrical profile design using inverse method 研究生：賴晉圓 Lai, Jin-Yuan 指導教授：楊榮華 Yang, Jung-Hua 朱曉明 Chu, Hsiao-Ming 【摘要】 本文針對活塞環幾何外型設計提出一套演繹法，此法只需要知道負荷與邊界條件即可估算活塞環外型的平滑曲線以及壓力分佈之情形。此演繹法運用雷諾積分式 Reynolds integral 與力平衡方程式，透過最小均方根誤差法 least-squares error method 、變分法 variation 、高斯-西德爾法 Gauss-Seidel method 及牛頓-瑞佛遜法 Newton-Raphson method 疊代法，即可估算出活塞環最佳幾何外型及壓力分佈值。結果顯示，活塞環幾何外型多項式函數之冪次愈低，使用本文所提出之方法估算的壓力分佈及活塞環幾何外型愈準確，誤差值可控制在3.0%以內。當冪次及點數增加，會使求得之壓力分佈與活塞環幾何外型分佈值更接近目標值。初始猜值之活塞環外型係數 及油膜最大壓力位置 皆對估算之結果有顯著的影響。就以估算之最大壓力值 及最小油膜厚度 值而言，初始猜值會比可容許較大的誤差量。當的初始誤差猜值介於-20%與20%之間，的初始誤差猜值介於-33.33%與20%之間，本文所提出之演繹法可將修正至4.0%內，活塞環幾何外型修正至5.0%內。 本文繼而將視為一未知數，並加入力矩平衡條件，將此演繹法運用於滑塊軸承幾何外型設計。結果顯示，初始猜值除了滑塊軸承外型 ，油膜最大壓力位置 以外，出口油膜厚度 也對此法估算之結果有顯著的影響。並且，初始猜值之可容許誤差量的大小依序為、及。使用本文所提出之演繹法估算壓力分佈之準確性會比估算油膜厚度之準確性較高，因此本文所提出之演繹法確實能夠準確獲得活塞環與滑塊軸承幾何外型與壓力分佈值。This study aims to develop an algorithm for designing the optimum shape of piston ring and pressure distribution using an inverse method. The proposed algorithm only needs to obtain the load and boundary conditions for estimating not only a smooth curve in the piston ring shape, but also in the pressure distribution. The algorithm is developed from Reynolds integral and force balance equations. The least-squares error method, variation method, Gauss-Seidel method and Newton-Raphson method are employed to calculate the optimum shape of piston ring. Simulation results reveal that the greater the degree of the polynomial function used, the greater the maximum pressure , load and friction coefficient are, and the smaller the minimum film thickness is. On the other hand, the lower the degree of the polynomial function used, the more accurate the estimated piston ring shape and pressure distribution are. As the degree of the polynomial function and the number of grid points increase, the estimated piston ring shape and pressure distribution become more accurate. The initial guessed values of the coefficients of the estimated piston ring