Real-Time Identification of Flutter Boundaries Using the Discrete Wavelet Transform
Jeffrey D. Johnson, Jun Lü, Atam Prakash Dhawan, Richard C. Lind · Journal of Guidance Control and Dynamics · 2002
Real-time analysis of an airframe’s eutter boundaries during eight testing can help ensure safety and reduce costs. One method of identiecation is to performcorrelation elteringusing a set of singlet functions. The method is abletoidentifyaccuratelythefrequencyand dampingcoefe cientof thesystemtoexcitation, but the computational time required can be too signiecant to implement in real-time. An alternative method is presented for correlation e ltering that employs amultiple-level discretewavelet transform. The wavelet transform decomposes theresponse 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 eight-test data. As a result, real-time identie cation of eutter boundaries during e ight testing may be possible with relatively low-cost computational resources.