Theoretical and practical implementation of sensor optimization for ground vibration testing

Kurzfassung

Making ground vibration tests more efficient is an ongoing request from aircraft manufacturers in order to enhance aircraft availability late in the development phase. Reducing the number of sensors to be used for a ground vibration test has been identified as one possibility to improve the overall efficiency of such a test. Mathematical methods have been developed and published in the last decades, which exploit the features of the mode shape matrix of the aircraft finite element model in a given frequency range. Examples of such methods include the Effective Independence Distribution Vector method by Kammer [1] and the QR decomposition method by Schedlinski [2]. The methods look for response degrees of freedom of the underlying finite element model that provide the most linear independent contribution to the mode shape matrix. The resulting set of optimized degrees of freedom determined by these methods is then considered as an optimal set of measurement degrees of freedom to be used for observing mode shapes. The application of sensor optimization methods is presented in several publications. The validation of these methods is, however, mostly limited to the verification of the linear independence of the mode shapes of the finite element model of a given analytical system observed with the optimized set of degrees of freedom. This paper is dedicated to the validation of sensor optimization methods from a more practical point of view. In contrast to other publications, the quality of sensor optimization shall be discussed by comparing modal analysis results of a large aircraft structure. In more detail, modal analysis has once been performed using a rich set of more than six hundred sensors. Afterwards, modal analysis has been repeated using a significantly reduced set of optimized sensor positions which comprises no more than one hundred measurement degrees of freedom. The work presented here was embedded in a “Research Ground Vibration Test” that has been performed recently on an A340-600. The results of sensor optimization applied to the A340-600 FE model will be presented. Finally, the validation of the sensor optimization tools is discussed in terms of modal analysis results obtained from the full and the optimized set of sensors. Intermediate results (e.g. identification of eigenvalues), practical aspects (e.g. interpretation of identified modes), and the pros and cons of sensor optimization will be summarized in the paper.