Real-time identification of flutter boundaries using the discrete wavelet transform

Real-time analysis of an airframe's flutter boundaries during flight testing can help ensure safety and reduce costs. One method of identification is to perform correlation filtering using a set of singlet functions. The method is able to identify accurately the frequency and damping coefficient of the system to excitation, but the computational time required can be too significant to implement in real-time. An alternative method is presented for correlation filtering that employs a multiple-level discrete wavelet transform. The wavelet transform decomposes the response signal into a set of subsignals that correspond to different frequency bands. The same operation is applied to each entry in a dictionary of singlet functions. The transform results in a considerable reduction in the data and, thus, to a reduction in the computational time needed to calculate the correlation. We demonstrate that our approach is able to identify accurately frequency and damping characteristics of the impulse response of both a synthetically generated test signal and actual flight-test data. As a result, real-time identification of flutter boundaries during flight testing may be possible with relatively low-cost computational resources.