Analysis of in-flight cabin vibration of a turboprop airplane by proposing a novel noise-tolerant signal decomposition method

Abstract

Vibration in passenger cabins of turboprop airplanes is a serious challenge. One of the essential steps in studying the cabin vibrations is to determine the contributing sources of vibration. The vibration signals are highly nonstationary and noisy. Therefore, one may require a noise-tolerant signal processing method for decomposition of the signals. In this article, the wavelet-based empirical mode decomposition is introduced for the first time to improve the performance of the traditional empirical mode decomposition in dealing with noise. Unlike the traditional empirical mode decomposition that extracts the signal trend by averaging the upper and lower envelopes intersecting local maxima and minima of the signal, the wavelet-based empirical mode decomposition directly extracts the signal trend by applying the multilevel wavelet decomposition of the consecutive approximations within the sifting process. Numerical studies are undertaken to evaluate the effect of noise on the performance of the empirical mode decomposition and wavelet-based empirical mode decomposition. Also, comparisons are made between the methods at dissimilar noise powers based on the orthogonality, integral, and energy decomposition criteria. The results indicate that both methods generate similar results in the absence of noise. Considering the number of obtained intrinsic mode functions, decomposition quality criteria, and computational cost, however, the wavelet-based empirical mode decomposition outperforms the classic method at higher noise levels. In this article, the wavelet-based empirical mode decomposition is used for analysis of in-flight airplane cabin vibration. A 52-passenger turboprop aircraft is equipped with eight triaxial piezoelectric accelerometers, and several flight tests are performed to acquire in-flight vibration signals within the passenger cabin. The proposed wavelet-based empirical mode decomposition is applied to the experimental data. Then, the amplitudes and frequencies of the intrinsic mode functions are examined. Finally, the probable vibration sources are identified based on the intrinsic mode functions characteristics.