5 Effective Use of Scanning Laser Doppler Vibrometers for Modal Testing 53 Fig. 5.11 (a–g) FE modes 10, 12, 14, 15, 19, 23, 26, respectively. These FE modes are given as examples of the high off-diagonal mode pairs in the autoMAC matrix in Fig. 5.10. The reduced geometry which aliased these mode shapes is given in Fig. 5.9b, from which it is clear that whilst the SLDV allows adequate description of these high-order mode shapes on the wing, critically, the sides of the box are inadequately characterised with some high off-diagonal terms in the MAC matrix (Fig. 5.12) caused by aliasing of modes in areas the SLDV could not measure. The FE and EMA modes were paired based on a minimum MAC of 50 %, with no repeated pairings (i.e., each mode could only be paired once) and no restriction between the difference in frequency of the modes. This relative lack of constraint and relatively low minimum MAC is acknowledged to be potentially sub-optimal, but it can be instructive to see how well the FE and EMA agree with minimal intervention. The UB100X BCC structure was observed to exhibit effectively no local modes; the box has inherent symmetries, and as a welded assembly is assumed to have a fairly uniform mechanical impedance (i.e., distributed stiffness). Additionally, considered individually both the box and wing feature many modes in the frequency range of interest (0–250 Hz), and so it is unsurprising that when considered as a system (B attached to C) that all modes see some participation from both components. In the case of the UB100X BCC assembly, the lack of local modes instead saw the spatial resolution advantage of the SLDV translated into being able to match a large number of EMA global modes to the FE, namely 20 pairs. 5.3.3 Model Updating In order to explore the influence of the SLDV measurements on the model updating process, multiple model updates were trialled with the complete EMA dataset, and a reduced EMA dataset representative of the spatial resolution which could be reasonably achieved with accelerometers. In the reduced EMA dataset there were six SLDV DOFs left on the wing (corners plus mid-points), which gave a total of 31 DOFs (which including a force reference on the driving point would populate a 32 channel acquisition system). The model updating was performed in several different ways, as is often the case when trying to understand the sensitivity of a model to changes in parameters, as explained under Sects. 3.3.1 and 3.3.2. The common settings between all updates were mass-density (¡) and Young’s modulus (E) were parameterised, with no constraint on the amount they could be varied. The responses were a target mass, and to optimise the frequencies and MAC for the paired modes. Updating was stopped when the correlation criterion fell below a certain level, or no improvement was seen between model iterations.
RkJQdWJsaXNoZXIy MTMzNzEzMQ==