本論文之研究目的在於採用有理式B仿線(rational B-spline)取代三次仿線(cubic spline)，以建構葉片外型，並使其符合多軸加工法。研究步驟依序如下：根據泵的操作條件決定比速率(specific speed)與淨正吸入揚程(NPSHR)以避免空蝕現象發生。利用歐拉理論結合速度三角型及葉片參數設定法決定流場參數。採用有理式B仿線數值曲線進行合成，以曲線的平滑性增進流場平滑性，限定流場參數並結合共軛創成加工法建構葉型的幾何實體。由刀具的干涉分析以設計分隔流板，強化流場特性。藉田口式實驗法之直交表，由初始葉型參數搭配不同水準值以產生足夠數目的設計範例，供類神經網路的訓練。以CATIA軟體建構包含隔板的三維流道，由計算流體力學軟體TASCflow進行流場分析。藉由引入隔板的設計，期能進一步優化離心泵葉片之效率，並擴大其工作區間。
本研究論文的結論如下：(一)由有理式B仿線進行合成的葉型具有多解性、多變性、平滑性，並可無因次化以求得曲線趨勢，可以應用在多軸加工機，縮短加工時程，而且可參數化。(二)由體積掃掠法建立葉型與流道形狀，以之建立格點進行流場模擬，可使葉片設計與模擬相互結合。(三)設計分隔流板並進行最佳化，可提升離心泵葉片之效率並增大其工作區間。

The aiming of this thesis is to approve the reflux action and substitute the cubic spline with rational B-spline for deciding the outward scale of blade, and make the blade can be manufactured by multi-axis milling. The process of this thesis would be written below. Decide the specific speed and NPSHR with the operation condition, which can avoid the generating of cavitation. Choose the flow parameter with the Euler theory combined with the velocity triangle and blade parameter designing method. Rational B-spline would be taken into curve designing for its smoothness, which can improve the smoothness of the flow, and create the blade body by concentrating the flow parameter and the conjunction generating machining method. Design the deflector and its length with the interference of tools to improve the flow characteristics. According to the experiment method, we generate enough samples for the training of Artificial Neural Network by the original blade design and different levels. Three-dimensional stream-way including the clearance would be build with CATIA, and its stream-way would be analyzed with computing flow dynamics software, TASCflow. Search the best parameters with the Pressure rising.
The advantages of this thesis were written below. The blade made by rational B-spline, which has the characteristics such as smoothness, multi-solution, proteiform, and can be normalized to search the tendency of spline is able being applied in multi-axis processor, shorten the manufacturing period, and has the advantage of being parameterized. Build the blade body and the stream-way with the conjunction generating machining method for forward blade appraising and flow analysis. Concerning the clearance make our flow analysis approach the actually impellor working situation combining with the shell which always cover it. Design the deflector in order to avoid of generating reflux. Expected issue is to build a high efficiency centrifugal impellor geometry without generating reflux and cavitation.

[1] Lewis, R. I., Turbomachinery Performance Analysis, John Wiley & Sons, Inc., New York, 1996.
[2] Demeulenaere, A., Leonard, O., and Van den Braembussche, R., “Application of a Three-Dimensional Inverse Method to the Design of a Centrifugal Compressor Impeller,” Transactions of the ASME Journal of Turbomachinery, Vol. 127, 1998.
[3] Chen, S. L., and Wang, W. T., “Computer Aided Manufacturing Technologies for Centrifugal Compressor Impellers,” Journal of Materials Processing Technology, Vol. 115, Issue. 3, pp. 284-293, 2001.
[4] Al-Zubaidy, S. N., “A Proposed Design Package for Centrifugal Impellers,” Computers & Structures, Vol. 55, No. 2, pp. 347-356, 1995.
[5] Morishige, K., and Takeuchi, Y., “5-Axis Control Rough Cutting of an Impeller with Efficiency and Accuracy,” IEEE Paper Robotics and Automation, Vol. 2, pp. 1241-1246, 1997.
[6] Lepine, J., Guibault,F., and Trepanier, J. Y., “Optimized Nonuniform Rational B-Spline Geometrical Representation for Aerodynamic Design of Wings,” AIAA Journal ,Vol. 39, No. 11. November, 2001.
[7] 林博正、杜黎蓉與吳建鋒，1999，" NURBS合成與對凸輪機構之影響"，中國機械工程學會，第十六屆學術研討會。
[8] Dawes, W. N., Dhanasekaran, P. C., Kellar, W. P., and Savill, A. M., “Reducing Bottlenecks in the CAD-to-Mesh-to-Solution Cycle Time to Allow CFD to Participate in Design,” Transactions of the ASME Journal of Turbomachinery, Vol. 123, 2001.
[9] Su, S. P., Chen, S. H., Lee, L. C., and Hwang, T. Y., “The Use of CFD in Turbomachinery Applications,” Transactions of the Aeronautical and Astronautical Society of the Republic of China, Vol. 32, No. 1, pp. 1-24, 2000.
[10] Zhang, M. J., Pomfret, M. J., and Wong, C. M., “Three-Dimensional Viscous Flow Simulation in a Backswept Centrifugal Impeller at the Design Point,” Computers & Fluids, Vol. 25, No. 5, pp. 497-507, 1996.
[11] Pak, E. T., and Lee, J. C., “Performance and Pressure Distribution Changes in a Centrifugal Pump Under Two-Phase Flow,” Proceedings of the Institution of Mechanical Engineers-A-Journal of Power and Energy, Vol. 212, pp. 165-171, 1998.
[12] 李海鋒，“利用三維紊流數值模擬進行離心葉輪設計比較”，流體機械，第29卷第9期，第18-22頁，2001.
[13] 黃福居，全三維軸流風扇葉片最佳化設計， 成功大學機械工程研究所碩士論文，2001.
[14] Zurada, J. M., Introduction to Artificial Neural Systems, West Publishing Company, Singapore, 1992.
[15] Cichocki, A., and Unbehauen, R., Neural Networks for Optimization and Signal Processing, John Wiley & Sons, Inc., New York, 1994.
[16] Oh, H. W., and Chung, M. K., “Optimum Values of Design Variables versus Specific Speed for Centrifugal Pumps” Proceedings of the Institution of Mechanical Engineers-A-Journal of Power and Energy, Vol. 213, pp. 219-226, 1999.
[17] Pierret, S., “Turbomachinery Blade Design Using a Navier-Stokes Solver and Artificial Neural Network,” Transactions of the ASME Journal of Turbomachinery, Vol. 121, 1998.
[18] Visser, F. C., Dijkers, R. J. H., and op de Woerd, J. G. H., “Numerical Flow-Field Analysis and Design Optimization of a High-Energy First-Stage Centrifugal Pump Impeller,” Computing and Visualization in Science, Vol. 3, Part. 1-2, pp. 103-108, 2000.
[19] 李生丕，應用類神經網路於軸流風扇葉片設計，成功大學機械工程研究所碩士論文，1999.
[20] Oyama, A., Liou, M. S., “Multiobjective Optimization of Rocket Enngine Pumps Using Evolutionary Algorithm” Journal of Propulsion and Power, Vol.18,No.3,pp.528-535,May-June 2002.
[21] Zierke, W. C., Straka, W. A., “Flow Visualization and the Three-Dimensional Flow in an Axial-Flow Pump”Journal of propulsion and power, Vol.12, No. 2,pp.250-259,March-April 1996.
[22]黃盈瑞，具貼壁式與非貼壁式加熱凸塊渠道流之熵產生分析，成功大學機械工程研究所碩士論文，2000.
[23] Bejan, J. Advanced Engineering Thermodynamics, John Wiley & Sons, Inc. 1988.
[24] Bejan, J. Convection Heat Transfer, John Wiley & Sons, Inc. 1984.
[25] Bejan, A. Entropy Generation Minimization, CRC Press, New York, 1995.
[26] Yilbas, B.S., Shuja, S.Z., and Budair, M.O., “Second low analysis of a swirling flow in a circular duct with restriction” Int. Journal of Heat and Mass Transfer Vol. 42, pp 4027-4041, 1999.
[27] Tuzson, J., Centrifugal Pump Design, John Willy & sons, Inc. 2000.
[28] Dufour , J. W.,and Nelson, W. E., Centrifugal Pump Sourcebook, McGraw-Hill, Inc. 1992.
[29] Brennen, C. E., Hydrodynamics of Pumps, Concepts ETI, Inc. 1994
[30] Volk, M. W., Pump Characteristics and Applications, Marcel Dekker, Inc. 1996.
[31]葉怡成，類神經網路模式應用與實作，儒林出版社，2003.
[32]劉惟信，機械最佳化設計，全華科技圖書，1996.