Chapter 34 Experimental Model Validation of an Aero-Engine Casing Assembly D. Di Maio, G. Ramakrishnan, and Y. Rajasagaran Abstract This work presents an experimental model validation of an aero-engine casing assembly. The objective is to show how to perform an experimental model validation of a casing assembly with bolted flanges. One of the major challenges was given by the unknown geometrical dimensions and material properties of the structure at the outset. This goal could be reached by using modal updating so as to identify the best geometrical and material parameters for the model. Experimental testing was a key contributor in this process, since the test data were used to benchmark the updating process. The updating process highlighted that the model could not be correctly updated until the joints were included. Therefore, the FE model required to be upgraded, by inclusions of joint flexibility, before it could be updated. This paper will present both the upgrading-updating process and the experimental testing carried out on the casing assembly. Keywords Upgrading • Updating • Experimental model validation 34.1 Introduction This research work was developed using an aero-engine casing assembly the finite model of which was not available at the beginning. The hardware was used in past researches [1] where the model was developed and supported by the project sponsor. Unfortunately, this information was not available in this present research. So, the objective and challenge was to exploit the potential of the model updating to experimentally validate a FE model. Ewins et al. [2] suggested that any FE model is good as much as it includes the correct physics which enables the right calculation of the dynamic response. Ewins also suggests that a model must be upgraded if physical elements are not yet taken into account for the calculation of the dynamic response. So, any FE model can be updated as much as possible but it will inevitably fail to simulate the correct dynamics until all important flexible elements are included, for instance joints. Therefore, under these circumstances, a model should be upgraded and then updated. For example, irregular changes of thickness in the physical structure can be included into the model updating process where the mass and stiffness are corrected for better values. However, if a joint flexibility is not included the model must be upgraded before any updating process takes place. To prove this assumption, the cases presented in here will discuss about a FE model created with and without the inclusion of the joint flexibility. Considering that both geometrical and material properties were unknown the paper will show that the model can be updated up to find the best values for predicting some of the modes measured during the experimental campaign. But, only the inclusion of the joint flexibility into the model will deliver the correct response modes. The process might look trivial but the model updating software, used in this project, requires the complete modelling of the bolts in order to perform the correct updating. And, this type of modelling is not a trivial task when approx. 150 bolts must be included, which creates redundancy and increases the computational cost. Despite the entire work is based on linear dynamic response, so nonlinearity is not even attempted, the process to upgrade and update a linear model showed not to be such a straightforward process. D. DiMaio ( ) • G. Ramakrishnan • Y. Rajasagaran Department of Mechanical Engineering, University of Bristol, University Walk, BS8 1TR, Bristol, UK e-mail: dario.dimaio@bristol.ac.uk © The Society for Experimental Mechanics, Inc. 2017 R. Barthorpe et al. (eds.), Model Validation and Uncertainty Quantification, Volume 3, Conference Proceedings of the Society for Experimental Mechanics Series, DOI 10.1007/978-3-319-54858-6_34 339
RkJQdWJsaXNoZXIy MTMzNzEzMQ==