Hendrik Verdonck, Reik Thormann, Hans Bleecke, Bernd Stickan
DOI Number: N/A
Conference number: IFASD-2019-025
During flutter analyses of aircraft a large parameter space has to be analyzed covering not only the extended flight envelope but also a large set of mass variations and failure cases. The generalized aerodynamic forces are computed for a reference mass case applying a linearized frequency domain method based on the Reynolds-averaged Navier-Stokes equations. The aerodynamic matrices for different mass cases are approximated using a least-square approach. While this approach is working for most normal mass cases, larger deviations in flutter predictions can occur for discrete structural failures. Error criteria are developed to estimate if the approximation is within bounds without calculating the aerodynamic response of the actual modes. The modal assurance criterion is compared to a local error, weighting the modal displacements with their corresponding aerodynamic forces, as well as a norm of the generalized aerodynamic force matrices. The three error estimators are benchmarked for a discrete structural failure case showing a similar performance. Furthermore, if the maximal error exceeds a certain threshold, the reference basis used in least-square approach has to be augmented to improve the prediction. While a suitable residual mode can be found by engineering judgment, an automatic process is preferable. A proper orthogonal decomposition is applied to find the most dominant error mode. This approach can also be applied to define a better reference mode set directly in the beginning of the process. Instead of using the mode shapes of a reference mass case, the modes of several mass cases including failure cases are used as snapshots. Results are presented comparing frequency and damping progressions as function of speed for a transport aircraft at nominal and failure conditions.