Modal Assessment of Wind Turbine Blade in Preparation of Experimental Substructuring
As an active member of the International Modal Analysis Conference (IMAC) Substructuring Focus Group, AWE has agreed to participate in an international effort in the development of analytical and experimental substructuring techniques. The field of analytical and experimental substructuring in highly complex, especially the coupling of the two methods, and deservedly attracts an international audience. A large and diverse working group directly benefits this development activity, in order to make significant progress over the next few years. The overall aim of the programme is to engage parties interested in dynamic substructuring to attack and conquer the practical problems limiting the use of substructures. In response to the IMAC XXIX, Jacksonville, Florida, experimental substructuring focus group meeting, the Structural Dynamics team at AWE have conducted modal analysis work on a test bed, in preparation for future experimental substructuring efforts to be conducted by the focus group.
Following the first meeting of the technical focus group at IMAC XXIX, it was unanimously agreed by all stakeholders that the mechanism to carry out this development would be via the substructuring of a wind turbine, specifically the Ampair 600 Wind Turbine, manufactured by a UK based company. The longer term deliverable of this work stream would be to successfully model this wind turbine system where the turbine blades were modelled from experimentally derived substructures and other components were modelled as analytical substructures. This offers a unique capability to better model complex structures with complex material properties in an assembly configuration.
Three Ampair 600 Wind Turbine blades were procured with the aim to understand the characteristics of their structure. The blade is divided into sections, a Blade and an Anchor to the rotor. The Blade is believed to be made from a composite of glass fabric and polypropylene matrix, and the Anchor, from a foam of thermoplastic based syntactic polymer.
To derive high precision blade geometry, a non-contact scanning and reverse engineering approach was used. For complex freeform geometry, conventional contact measurement methods are unsuitable due to their low point density and rate of capture. For this reason a Romer Absolute articulated arm with Perceptron V5 laser scanner was used. This system was specifically used due to its tolerance of reflective surfaces and those with differing contrast, as present in the turbine blades. All three blades were measured and the resulting stitched surface meshes were put out in standardised CAD format. A master (average) mesh was constructed from the combinations of the individuals by averaging the errors between the three datasets over their complete geometry. A two-stage best fit alignment was used comprising of a coarse alignment by minimising deviations between a small number of sample points followed by a fine alignment using a more dense point sample. Mesh averaging the surfacing was done using Rapidform XOR3 and the final model generation was completed in Unigraphics NX5.
A full modal assessment was carried out by the authors, consisting of Modal free-free testing and Modal free and fixed analysis, together with an analytical model update and full correlation between the experimental and analytical results. Finite element models (FEM) for each of the blades, tuned to their relative mass, were generated in the ANSYS analysis code. Material models in these analytical models were further optimised following correlation with test data.
The three turbine blades were each scanned in a Free-Free condition using a Polytec 3D Scanning Laser Doppler Vibrometer. Using this technique it was possible to get a relatively high spatial density of measurement points compared to traditional techniques such as accelerometers. In addition to the high spatial density, translational DoFs were measured in all three axes. The measurements were post-processed and its modal parameters were fitted using the ICATS software package. The results from the physical tests were mapped to a reduced FEM. This allowed like-for-like, test-to-test and test-to-FEM correlation using a number of standard correlation metrics.
As an immediate follow-on to this piece of work, the authors suggest that fixed-based modal tests of the three blades are carried out including to-test and to-FEM correlation for completeness at the least. It is also proposed that further material testing is carried out on blade material samples, including visual inspection of the fibre layup and Dynamic Material Analysis to help better characterise its mechanical response.