本研究的目的是展示一套最佳化垂直軸風力發電機（VAWT）功率的半自動 化的流程。其方法是利用基因演算法，設計參數及約束條件根據半經驗坐標系統 翼型參數化(PARSEC)方法定義，目標是最大化功率係數。為了提高優化效率， 該架構結合了機器學習，神經網路模型使用元素空間卷積神經網路(ESCNN)，資 料集由Qblade軟體中的理論模型計算並生成，資料的選定是設定某一區間的風 速及轉速。本篇以NACA0015為範例進行最佳化，為了比較優化前後的功率係 數差異與獲得較準確的結果，所以將優化前與優化後的葉片皆以Ansys Fluent軟 體進行數值模擬，包括在不同轉速下進行功率比對。

The purpose of this study is to demonstrate a semi-automated process for optimizing the power output of a vertical axis wind turbine (VAWT). The methodology involves using a typical genetic algorithm, with design parameters and constraints defined according to the PARSEC method, aiming to maximize the power coefficient. To improve optimization efficiency, the framework integrates machine learning. The dataset is calculated and generated by theoretical models in Qblade software, with data selection based on specified ranges of wind speed and rotational speed. The optimization is exemplified using the NACA0015 airfoil. To compare the differences in power coefficients before and after optimization and to obtain more accurate results, numerical simulations of the blades before and after optimization are conducted using CFD software to determine the power coefficient.

[1] K. Fujii and G. S. Dulikravich, Eds., Recent Development of Aerodynamic Design Methodologies: Inverse Design and Optimization, vol. 65. in Notes on Numerical Fluid Mechanics (NNFM), vol. 65. Wiesbaden: Vieweg+Teubner Verlag, 1999. doi: 10.1007/978-3-322-89952-1.
[2] P. M. Kumar, S. R. Rashmitha, N. Srikanth, and T.-C. Lim, “Wind Tunnel Validation of Double Multiple Streamtube Model for Vertical Axis Wind Turbine,” Smart Grid Renew. Energy, vol. 08, no. 12, pp. 412–424, 2017, doi: 10.4236/sgre.2017.812027.
[3] R. Bravo, S. Tullis, and S. Ziada, “Performance Testing of a Small Vertical-Axis Wind Turbine”.
[4] “A Cascade Theory for the Aerodynamic Performance of Darrieus Wind Turbines,” 2024.
[5] T. J. Baker, “Mesh generation: Art or science?,” Prog. Aerosp. Sci., vol. 41, no. 1, pp. 29–63, Jan. 2005, doi: 10.1016/j.paerosci.2005.02.002.
[6] A. Rezaeiha, H. Montazeri, and B. Blocken, “On the accuracy of turbulence models for CFD simulations of vertical axis wind turbines,” Energy, vol. 180, pp. 838–857, Aug. 2019, doi: 10.1016/j.energy.2019.05.053.
[7] H. Liu and F. Ave, “An Airfoil Aerodynamic Parameters Calculation Method Based on Convolutional Neural Network”.
[8] W. Peng, Y. Zhang, E. Laurendeau, and M. C. Desmarais, “Learning aerodynamics with neural network,” Sci. Rep., vol. 12, no. 1, p. 6779, Apr. 2022, doi: 10.1038/s41598-022-10737-4.
[9] J. H. Holland, Adaptation in Natural and Artificial Systems - An Introductory Analysis with Applications to Biology, Control, and Artificial I. Cambridge: The MIT Press, 2019.
[10] C. S. Ferreira and B. Geurts, “Aerofoil optimization for vertical‐axis wind turbines,” Wind Energy, vol. 18, no. 8, pp. 1371–1385, Aug. 2015, doi: 10.1002/we.1762.
[11] T. Majid and B. W. Jo, “Comparative Aerodynamic Performance Analysis of Camber Morphing and Conventional Airfoils,” Appl. Sci., vol. 11, no. 22, p. 10663, Nov. 2021, doi: 10.3390/app112210663.
[12] G. Bangga, S. Hutani, and H. Heramarwan, “The Effects of Airfoil Thickness on Dynamic Stall Characteristics of High‐Solidity Vertical Axis Wind Turbines,” Adv. Theory Simul., vol. 4, no. 6, p. 2000204, Jun. 2021, doi: 10.1002/adts.202000204.
[13] L. Wang et al., “A deep learning-based optimization framework of two-dimensional hydrofoils for tidal turbine rotor design,” Energy, vol. 253, p. 124130, Aug. 2022, doi: 10.1016/j.energy.2022.124130.
[14] I. M. Mirzana, S. R. Ahmed, and T. Junaid, “Fabrication Methodology of Small Scale Horizontal Axis Wind Turbine,” 2016.
[15] A. L. Maas, A. Y. Hannun, and A. Y. Ng, “Rectifier Nonlinearities Improve Neural Network Acoustic Models”.
[16] B. T. Polyak, “Some methods of speeding up the convergence of iteration methods,” USSR Comput. Math. Math. Phys., vol. 4, no. 5, pp. 1–17, Jan. 1964, doi: 10.1016/0041-5553(64)90137-5.
[17] M. H. G. Syafei, R. T. Indrawati, and T. A. Farhan, “Implementation of the Airfoil parameterization PARSEC Method in Python,” J. Rekayasa Mesin, vol. 17, no. 3, p. 485, Dec. 2022, doi: 10.32497/jrm.v17i3.4000.
[18] F. R. Menter, “Two-equation eddy-viscosity turbulence models for engineering applications,” AIAA J., vol. 32, no. 8, pp. 1598–1605, Aug. 1994, doi: 10.2514/3.12149.
[19] S. Michael, “Applied Aerodynamics Group, University of Illinois Urbana-Champaign, Department of Aerospace Engineering, UIUC Airfoil Coordinates Database”, [Online]. Available: https://m-selig.ae.illinois.edu/ads/coord_database.html
[20] “QBLADE: AN OPEN SOURCE TOOL FOR DESIGN AND SIMULATION OF HORIZONTAL AND VERTICAL AXIS WIND TURBINES,” Int. J. Emerg. Technol. Adv. Eng..
[21] K. Dimitrios, “parsec-airfoils”, [Online]. Available: https://github.com/dqsis/parsec-airfoils
[22] mbodmer, “libairfoil”, [Online]. Available: https://github.com/mbodmer/libairfoil/tree/master
[23] L. Zi, “CNN-for-Airfoil,” 2018, [Online]. Available: https://github.com/ziliHarvey/CNN-for-Airfoil
[24] 林銘洋/Lin M.-Y., “基於深度學習之翼型優化及流場預測/Airfoil Optimization and Flow Field Prediction Based on Deep Learning”.
[25] Y. A. Çengel and J. M. Cimbala, Fluid mechanics: fundamentals and applications, Third edition. New York: McGraw Hill, 2014.
[26] L. Zhang et al., “Influences of trailing edge split flap on the aerodynamic performance of vertical axis wind turbine,” Energy Sci. Eng., vol. 9, no. 1, pp. 101–115, Jan. 2021, doi: 10.1002/ese3.818.
[27] “Hi-VAWT Technology”, [Online]. Available: http://www.hi-vawt.com.tw/tw/tw_why_vaswt.html
[28] Suprayitno, J. C. Yu, Aminnudin, and R. Wulandari, “Airfoil aerodynamics optimization under uncertain operating conditions,” J. Phys. Conf. Ser., vol. 1446, no. 1, p. 012014, Jan. 2020, doi: 10.1088/1742-6596/1446/1/012014.
[29] “MSc Thesis Application of automotive alternators in small wind turbines”.