论文部分内容阅读
针对抖振载荷难以直接测量的问题,提出了一种由抖振加速度响应逐点识别翼面分布抖振载荷的频域方法。根据抖振载荷空间和时间的分布特性,将抖振载荷表示成一组空间正交函数与一组时间函数的线性组合,从而把频域内的抖振载荷识别问题转化为时间函数的识别问题。通过垂尾结构的气动弹性运动方程推导出加速度响应功率谱密度与广义力功率谱密度之间的关系式,再由谱分解理论得出广义力功率密度与抖振载荷功率谱密度的关系式,最后根据空间分布函数的正交性逐点识别出时间函数。为了解决上述逐点识别过程中遇到的不适定问题,提出了一种新的正则化处理方法,并用牛顿迭代法选取最佳正则化因子。对一个垂尾模型,先用计算流体力学(CFD)仿真软件计算出垂尾模型上的抖振载荷,然后将这些抖振载荷施加在垂尾结构上,并计算出垂尾结构的抖振加速度响应,利用计算出的加速度响应识别出抖振载荷,并与计算的抖振载荷进行比较,从而验证了本文所提出的抖振载荷逐点识别方法具有很好的识别精度。
Aiming at the problem that the chattering load can not be directly measured, a frequency-domain method of identifying the chattering load of the airfoil distributedly by the chattering acceleration response is proposed. According to the distribution of chattering load space and time, the chattering load is expressed as a linear combination of a set of spatial orthogonal functions and a set of time functions, so that the chattering load identification in the frequency domain can be transformed into the identification of time functions. The relationship between the power spectral density of acceleration response and generalized power spectral density is deduced from the equation of aeroelastic motion of the vertical tail structure. The relationship between the power density of generalized force and power spectral density of the chattering load is derived from the theory of spectral decomposition. Finally, the time function is identified point by point according to the orthogonality of the spatial distribution function. In order to solve the ill-posed problems encountered in the point-by-point identification process, a new regularization method is proposed and the best regularization factor is selected by Newton’s iteration method. For a vertical tail model, CFD simulation software is used to calculate the chattering load on the vertical tail model, and then the chattering loads are applied to the vertical tail structure, and the chattering acceleration of the vertical tail structure is calculated Response, the chattering load is identified by using the calculated acceleration response and compared with the calculated chattering load, which verifies that the chattering load point-by-point identification method proposed in this paper has good recognition accuracy.