基于物理信息神经网络与模态旋转法的大柔性机翼静气动弹性分析

Static aeroelasticity analysis of very flexible wings based on physics-informed neural network and modal rotation method

  • 摘要: 高空低雷诺数太阳能飞机呈现出展弦比大、结构轻、气动力非线性强、气动弹性变形大的特点。然而,低雷诺数下的复杂流动(如层流分离泡)与大柔性结构的大变形相互耦合,导致传统气动力模型和线性结构分析方法难以准确、高效地预测其静气动弹性响应。针对这一挑战,本文提出一种物理信息神经网络(PINN)与模态旋转法耦合的大柔性机翼非线性静气动弹性响应计算方法。首先,考虑黏性效应,建立基于PINN的翼型流场和气动力求解方法,计算低雷诺数下大展弦比机翼每个剖面的气动力;其次,建立基于模态旋转法的大柔性机翼结构力学模型,通过气动力与结构响应的插值和双向耦合,分析其静气动弹性响应特性。结果表明,与试验计算结果相比,PINN计算的翼型流场相对误差在8%以内;小变形情况下,本文方法计算的静气动弹性变形与线性理论结果误差小于5%。在此基础上,进一步分析了不同刚度下的大变形机翼静气动弹性响应规律:翼尖挠度随刚度降低呈对数型增大,展向扭转角分布显著改变;当弯曲刚度减小一个数量级时,翼尖有效迎角变化超过2°,平均升力系数下降14%以上,表明低刚度会显著加剧几何非线性效应并诱发载荷重分配。本研究可为高空太阳能飞机的气动弹性分析与设计提供有益参考。

     

    Abstract: High-altitude low-Reynolds-number solar-powered aircraft are characterized by high aspect ratios, lightweight structures, strong aerodynamic nonlinearities, and large aeroelastic deformations. However, the complex low-Reynolds-number flows, such as laminar separation bubbles, are tightly coupled with large deformations of very flexible structures, making it difficult for conventional aerodynamic models and linear structural analysis methods to accurately and efficiently predict the static aeroelastic responses. To address this challenge, this paper proposes a nonlinear static aeroelastic analysis method for very flexible wings by coupling a physics-informed neural network (PINN) with the modal rotation method. First, considering viscous effects, a PINN-based solver is established to compute the flow field and aerodynamic forces around airfoil sections, providing spanwise aerodynamic loads for high-aspect-ratio wings at low Reynolds numbers. Second, a structural model based on the modal rotation method is developed for the very flexible wing. Through interpolation and bidirectional coupling between aerodynamic loads and structural deformations, the static aeroelastic response characteristics are analyzed. Results show that, compared with experimental data, the flow fields predicted by the PINN yield relative errors within 8%. Under small deformation conditions, the static aeroelastic deformations computed by the proposed method agree with linear theory results within 5%. Furthermore, parametric studies on wings with varying bending stiffnesses reveal that the wingtip deflection increases logarithmically as stiffness decreases, accompanied by a significant redistribution of spanwise twist angles. When the bending stiffness is reduced by one order of magnitude, the effective angle of attack at the wingtip varies by more than 2°, and the mean lift coefficient decreases by over 14%, indicating that low stiffness substantially intensifies geometrical nonlinear effects and induces load redistribution. This study provides a useful reference for aeroelastic analysis and design of high-altitude solar-powered aircraft.

     

/

返回文章
返回