Horizontal Axis Tidal (HAT) turbines can be used to power RO (reverse osmosis) desalination systems. The greatest weakness of these turbines is the high price of design, development, and manufacturing. Traditionally, optimization of turbine geometry is achieved by running several numerical models of the turbine which can become time consuming and expensive. The Taguchi-CFD (Computational Fluid Dynamics) approach has recently been introduced as an inexpensive and rapid tool for optimizing industrial devices. This technique can be used as a straightforward solution for optimization of geometry of HAT turbines. In this work, a conceptual design of a tidal power reverse osmosis (TPRO) desalination unit was proposed. Subsequently, the geometry of the HAT turbine, which can power the whole desalination system, was optimized with combination of only 16 CFD simulations using the Taguchi method. The effects of blade size, number of blades, hub radius, and hub shape were studied and optimized. The Taguchi results revealed that the most important parameters influencing the power output of HAT turbine are the number of blades, size of blade, hub radius, and hub shape respectively. Moreover, the results of the superposition model showed that the minimum signal-to-noise ratio (SNR) is 21% less than the amount achieved in the Taguchi approach. The power coefficient (C_p) of the optimized HAT turbine was 0.44 according to the results of CFD simulations, which was 10% higher than that of the baseline model (0.40) at tip speed ratio (TSR) of 5. The weight of the optimized model was less than the baseline model by 17%. The results of this study provide a comprehensive guidance for horizontal turbine optimization process.

水平轴潮汐（HAT）涡轮机可用于为RO（反渗透）淡化系统提供动力。这些涡轮机的最大缺点是设计，开发和制造的高昂价格。传统上，通过运行多个涡轮机数值模型来实现涡轮机几何形状的优化，这会变得既费时又昂贵。Taguchi-CFD（计算流体动力学）方法最近作为一种廉价且快速的工具被用于优化工业设备。该技术可用作优化HAT涡轮机几何形状的直接解决方案。在这项工作中，提出了潮汐反渗透（TPRO）海水淡化装置的概念设计。随后，HAT涡轮的几何形状可以为整个海水淡化系统提供动力，使用Taguchi方法仅通过16个CFD模拟的组合进行了优化。研究并优化了叶片尺寸，叶片数量，轮毂半径和轮毂形状的影响。Taguchi的结果表明，影响HAT涡轮机功率输出的最重要参数分别是叶片数量，叶片尺寸，轮毂半径和轮毂形状。此外，叠加模型的结果表明，最小信噪比（SNR）比Taguchi方法获得的信噪比小21％。功率系数（和轮毂形状。此外，叠加模型的结果表明，最小信噪比（SNR）比Taguchi方法获得的信噪比小21％。功率系数（和轮毂形状。此外，叠加模型的结果表明，最小信噪比（SNR）比Taguchi方法获得的信噪比小21％。功率系数（根据CFD模拟的结果，优化的HAT涡轮的C _p）为0.44，比叶尖速比（TSR）为5的基线模型（0.40）高10％。优化模型的重量较小比基准模型高出17％。这项研究的结果为水平涡轮机优化过程提供了全面的指导。