The necessity of this optimization stems from the fact that the output and stability of downhole turbine generators are profoundly influenced by their turbine hydrodynamics. Specifically, the blade profile is the primary factor governing critical parameters such as pressure loading, torque generation, rotational speed, and overall energy conversion efficiency. While previous studies have addressed various aspects of turbine design, such as structural mechanics or turbine-level performance, there remains a substantial gap in the systematic optimization of sectional hydraulic airfoils, particularly under the low-Reynolds-number conditions characteristic of compact downhole turbine applications. This study addresses this gap by employing a robust optimization strategy that integrates geometric parameterization with evolutionary computation.

The foundation of the design lies in the selection of the NACA66 (mod) profile as the baseline hydraulic airfoil. To allow for flexible and precise modification of this profile, the researchers employed high-order Bezier curves, specifically utilizing eight movable control points. Bezier and B-spline representations are highly valued in engineering because they allow for the description of smooth geometries using a relatively small number of variables, which is crucial for effective optimization. By formulating the airfoil’s curve using these Bezier functions, the design space was made flexible enough to allow the genetic algorithm to explore numerous geometric configurations while ensuring the resulting airfoil maintains a smooth and continuous surface, a factor essential for predictable hydraulic performance. The optimization process was executed using MATLAB, which managed the Bezier parameterization and the genetic algorithm, while XFOIL was utilized as a rapid computational tool to estimate sectional lift and drag.

The core of the optimization involved defining a clear objective function and a set of physical constraints. The primary goal was to maximize the energy efficiency of the blade, which is directly correlated with maximizing the lift-to-drag ratio ($C_L/C_D$). The objective function was defined as $f(X) = ext{max } C_L/C_D$, where $C_L$ and $C_D$ represent the lift and drag coefficients, respectively. The design variables ($X$) consisted of the coordinates of the eight control points on the Bezier curve. To ensure the resulting airfoil was practical and structurally sound, several constraints were imposed. For instance, to maintain a wide range of laminar flow and ensure machinability, the maximum relative thickness position ($L/c$) was constrained to be less than 5%. Furthermore, constraints were placed on the curvature ($x_t/c$) and the position of the bending point ($f/c$) to balance the need for high lift with the necessity of maintaining structural strength and preventing excessive camber.

The optimization process followed a standard genetic algorithm protocol: determining design variables, objective functions, and constraints; generating an initial population of candidate airfoils; calculating the fitness value of each candidate; preserving the best-performing individuals; and iteratively selecting, crossing, and mutating the population until the objective function and constraints were satisfied. This iterative process, which involved coupling the Bezier parameterization with the XFOIL solver, allowed the algorithm to effectively navigate the complex, constrained design space. The principal optimization was conducted at a Reynolds number ($ ext{Re}$) of 3000, covering an angle-of-attack range from $-10$ to $20$ degrees. Crucially, the researchers recognized that while XFOIL is a computationally efficient tool for rapid comparative optimization, it is not a complete representation of the complex liquid-flow turbine environment. Therefore, the optimized profile was rigorously assessed using both liquid-flow CFD simulations and physical turbine-level experiments to ensure its real-world viability.

The results of the sectional optimization demonstrated that the optimized airfoil exhibited superior hydrodynamic characteristics compared to the original NACA66 (mod) profile. Specifically, the lift-to-drag ratio achieved a 26.9% improvement at an angle of attack of 15 degrees. Visually and mathematically, the optimized profile showed a reduction in the thickness of the upper surface, and the position of the maximum thickness shifted rearward relative to the initial design. Furthermore, sensitivity analyses conducted across a range of Reynolds numbers—300, 3000, and 30,000—confirmed that the optimized profile maintained stable hydrodynamic performance across this investigated low-to-medium Reynolds-number range, with the most favorable performance occurring at angles of attack between 5 and 10 degrees. This improved sectional performance was further validated by CFD simulations, which showed better pressure distribution and lift-drag characteristics under liquid-flow conditions, and by physical experiments.

The translation of sectional airfoil improvements into macroscopic turbine performance was perhaps the most compelling aspect of this study. Experimental testing within a controlled hydraulic circuit compared the standard NACA66 airfoil against the optimized design. The results showed that the optimized turbine generated significantly higher mechanical output. At the highest tested flow rate, the optimized turbine increased its torque from 0.997 N m to 1.297 N m, representing a substantial increase of 30.2%. Concurrently, the rotational speed of the optimized turbine increased from 712 to 904 revolutions per minute, an increase of 27.1%. When evaluating the overall efficiency, the tip-speed ratio (TSR) of the optimized turbine was consistently higher across the entire tested range, increasing from approximately 0.33 to 0.45, compared to the original turbine’s range of 0.25–0.36. Although the normalized power coefficient ($C_{ ext{power}}$) decreased with increasing flow rate for both turbines—a trend expected because the reference kinetic power term increases with the cube of the inflow velocity—the optimized turbine consistently exhibited a higher $C_{ ext{power}}$ than the original turbine at all investigated operating points, confirming that the gains in sectional hydrodynamic efficiency were successfully reflected in enhanced mechanical power output at the turbine level.

The comprehensive nature of this study, which bridged the gap between rapid 2D sectional optimization and high-fidelity 3D turbine-level verification, underscores the effectiveness of the proposed Bezier–GA–XFOIL framework for downhole generator applications. The methodology successfully demonstrated how precise geometric modification, guided by evolutionary computation and constrained by physical limitations, can yield significant performance enhancements. However, the researchers also provided a candid assessment of the study’s limitations. They acknowledge that the use of XFOIL is a rapid, two-dimensional comparison tool, and while Reynolds-number similarity is used, it cannot fully capture all the physical complexities of liquid flow, such as the specific density, viscosity, and rheological properties of actual drilling fluids. Furthermore, the study did not explicitly model cavitation phenomena, nor did it account for fluid–structure interactions, assuming rigid blade structures. The optimization was also conducted at the two-dimensional wing shape level, meaning three-dimensional blade effects, guide-vane interactions, and casing constraints were not fully incorporated into the model.

Looking ahead, the research team plans to expand this framework into a comprehensive turbine stage design method. Future work will involve introducing cavitation modeling and fluid–structure interaction (FSI) analysis to evaluate blade deformation and stress under real operating loads. Extending the optimization from two dimensions to three dimensions will allow for the consideration of cross-blade chord geometry, hub, and casing effects, as well as the interactions between blades and guide vanes. Moreover, more in-depth experimental studies are planned across a broader spectrum of flow rates, pressures, and drilling fluid characteristics to further quantify mechanical power and hydraulic efficiency. By systematically comparing various population-based optimization methods under identical variables and constraints, the researchers aim to further enhance the robustness and accuracy of this powerful optimization framework for critical energy applications in deep subsurface environments.

Subject of Research: Downhole turbine generator airfoil optimization using genetic algorithms.

Article Title: Optimization and experimental analysis of turbine airfoil of downhole generator based on genetic algorithm

Article References: Qu, J., Xue, Q., Tan, X., Li, Y., & Zhou, P. (2026). Optimization and experimental analysis of turbine airfoil of downhole generator based on genetic algorithm. Mechanical Sciences, 17(2), 935-949. https://doi.org/10.5194/ms-17-935-2026

Image Credits: AI Generated

DOI: 10.5194/ms-17-935-2026

Keywords: Genetic Algorithm, Downhole Turbine, Airfoil Optimization, Hydraulic Power, Fluid Dynamics, Bezier Curves, Mechanical Engineering, Optimization, experimental, analysis, turbine, airfoil