Measuring the rotational dynamic behaviour of a structure is important for many areas of dynamics such as passive vibration control, acoustics, and model updating. Specialist and dedicated equipment is often needed, unless the rotational degree-of-freedom is synthesised based upon translational data. However, this involves numerically differentiating the translational mode shapes to approximate the rotational modes, for example using a finite difference algorithm. A key challenge with this approach is choosing the measurement spacing between the data points, an issue which has often been overlooked in the published literature.ududThe present contribution will for the first time prove that the use of a finite difference approach can be unstable when using non-exact measured data and a small measurement spacing, for beam-like structures. Then, a generalised analytical error analysis is used to propose an optimised measurement spacing, which balances the numerical error of the finite difference equation with the propagation error from the perturbed data. The approach is demonstrated using both numerical and experimental investigations. It is shown that by obtaining a small number of test measurements it is possible to optimise the measurement accuracy, without any further assumptions on the boundary conditions of the structure.
展开▼
机译:测量结构的旋转动力学行为对于动力学的许多领域都很重要,例如被动振动控制,声学和模型更新。除非基于平移数据合成旋转自由度,否则通常需要专业和专用设备。但是,这涉及在数值上区分平移模式形状以近似旋转模式,例如使用有限差分算法。这种方法的主要挑战是选择数据点之间的测量间距,这一问题在出版的文献中经常被忽略。 ud ud目前的贡献将首次证明使用有限差分法可以解决问题。对于类似梁的结构,当使用不精确的测量数据和较小的测量间隔时,它不稳定。然后,使用广义分析误差分析来提出优化的测量间隔,该测量间隔使有限差分方程的数值误差与来自被摄动数据的传播误差之间取得平衡。使用数值和实验研究都证明了该方法。结果表明,通过获得少量的测试测量结果,可以优化测量精度,而无需进一步假设结构的边界条件。
展开▼