Research On Aircraft Flutter Modal Parameter Identification

Flutter modal parameters identification plays an important role in flight flutter testing field. The traditional modal identification methods suffering from the low SNR (signal to noise ratio), close mode and nonlinear structure can't provide an accurate estimation of modal parameter (especially the damping ratio). For this reason, some new methods based on advanced signal processing and system identification theory are presented to meet the demand of new aircraft flight testing. The research results are partly applied to new aircraft flutter testing, and achieve great military and economy benefit. The main results of this thesis are described as follows:1. A new wavelet denoising method is introduced for flight flutter test excited by rocket impulse. In this method, the testing data is first preprocessed with a gradient inverse weighted filter to initially lower the noise. The redundant wavelet transform is then used to decompose the signal into several levels. A "clean" input is recovered from the noisy data by level dependent thresholding approach, and the noise of output is reduced by a modified spatially selective noise filtration technique. The advantage of the wavelet denoising is illustrated by means of simulated and real data.2. A novel denoising method based on least-square support vector machines (LS-SVM) is proposed for flutter test excited by rocket impulse. The denoising procedure is implemented by time series prediction using LS-SVM, which is applicable to nonstationary and nonlinear impulse response data with short duration. Since SNR of impulse response varies with amplitude for the decaying property, the beginning data points with high SNR is used for training and prediction of the subsequent data with low SNR. Finally, the simulations and experiment on real flight test data demonstrate effectiveness and efficiency of this approach.3. The control surface oscillated by sweep command is a new form excitation in our flight test and two time-frequency filtering methods are developed for new excitation. The common of the two methods is that they all utilize the...