46 B. Weekes and D. Ewins this case if sufficiently many measurement locations are defined. The limitation of this approach is that a single view-point of the laser will for non-2D (i.e., non-planar) test objects often fail to adequately capture the dynamics of the structure, necessitating supplementary transducers to provide additional DOFs out of the line-of-sight of the SLDV. The caveat in the use of additional transducers in such a manner is that the metrics on which the correlation is based will (by default) be significantly weighted by the many SLDV points, and so the relative weight of the supplementary DOFs (which are clearly of importance to be warranted) is reduced. The capture of many non-coincident measurement locations by SLDV inherently increases capture of local modes. The identification of local modes is usually an advantage of using the SLDV, since it is often desirable to update the model based only on the lower order global modes, discarding the modes identified as local. This is because the low-order global modes are more likely to be a structural concern, and local modes often feature only subassemblies, which are likely to be subject to a boundary condition comprised of joints which are difficult to model (and therefore likely to be updated in a spurious way, to the detriment of the more important global modes). In the model updating process the model is iteratively updated multiple times, making it easier to discard the EMA local modes and rely upon the FE local modes not correlating against the EMA global modes and therefore not affecting the model update. 5.2.2 Import of SLDV Data: The Universal File Format Modal analysis of structures which behave in a strongly linear manner is highly developed, with a number of software suites available in which to perform capture1 and analysis2 of the relevant data. However, these capture suites tend to require an array of transducers such that concurrent capture can be performed at all DOFs. Since capture is performed on discrete transducers by means of a simple voltage input, these systems can easily accommodate various transducers such as accelerometers and strain gauges. Such software suites currently lack the means to acquire data from SLDVs, which is unsurprising since interfacing with SLDVs requires calibration of the scanning mirrors to the video feed (test specific), defining the scan geometry on the video feed, measurement sequencing, etc., with the bespoke hardware/software giving interfacing difficulties. The software which comes with SLDV systems is usually adequate to perform a test and to review the results as FRFs, ODSs and sometimes as mode shapes, damping values, etc., but for more detailed analyses it is often desirable to output the captured time-histories or FRF data to perform the modal analysis in proven modal analysis software. 5.2.2.1 Origin of the Universal File Format The standard means of conveying modal test data is the universal file format (often .unv, .uff, .asc), an ASCII (i.e., text) based format defined by the Structural Dynamics Research Corporation (SDRC) [5] in the late 1960s and early 1970s to permit transfer of data between early computer design and test systems. Such is the legacy of the format that the data fields are typified by 80 character limits to fit 80 column punch card records. As an ASCII-based format, the files can be opened with any text editor, and with the formatting guide the files can be understood by a human operator. 5.2.2.2 Important UFF Datasets for Modal Analysis Universal files comprise datasets of various types [5]. There is provision for time history and FRF data in dataset type 58, and mode shape data in dataset type 55. Also present in universal files are datasets containing the header information (metadata), units, geometry, and coordinate systems. In exporting data from one software package to another using universal files it is often necessary to adjust the data manually if there is some disparity between the universal file interpretations for the software packages. Despite the supposed standardisation offered by the format, there is some ambiguity and variation in how the UFF interpreters are written, e.g., translation and rotation matrices may be defined to map from global to local coordinate systems, or vice versa. There are also legacy dataset types which can lead to difficulties with incomplete support, e.g., geometry dataset type 2,411 vs. type 15. Units are often imported incorrectly, especially in the case of units of acceleration, e.g., g, instead of mm/s2 when the units dataset specifies SI units. The author has found cases of FRF plots scaled correctly in the modal 1e.g.: LMS Test.Lab, DataPhysics SignalCalc, MCP International SmartOffice. 2e.g.: LMS Test.Lab, DataPhysics SignalCalc, Spectral Dynamics STAR Modal, HDM nCode, MCP International SmartOffice.
RkJQdWJsaXNoZXIy MTMzNzEzMQ==