INVESTIGATION OF EXHAUSTING PHENOMENA OF GAS TURBINE ENGINE ON THE BASIS OF WAVELET ANALYSIS

Abstract

The method for research of a gas turbine engine (GTE) surging phenomena based on wavelet analysis was
developed. GTE DG-90 was chosen for research. A wavelet decomposition of the air pressure signal was carried out using a high pressure compressor and Dobeski, Symlet, Coffle and Meyer's discrete wavelet. The advantages and disadvantages of filtering properties of these wavelets were graphically represented in the form of amplitude-frequency characteristics. Based on the results obtained, a wavelet transform parameter selection scheme based on the analysis of the frequency response of wavelet filters was constructed, and the optimum sampling time of the wavelet filtering of the investigated signal was calculated for determination of the propagation and rotational breakdown in the GTE turbocompressor. According to this scheme, among the wavelets of Dobeski, Simlet, Koefleta, and Meyer's discrete wavelet, for studying the air pressure signal at the compressor during bursting processes, the first were selected by the Dobeches and Symmetes wavelets of the 2nd order according to their filtration rate. Considering high speed and characteristics of the description of the occurrence of excite phenomena by means of a fast Fourier transform to the distribution of the wavelet coefficients of the air pressure signal by turbocompressor, for the diagnosis of unstable flows occurring in the gas path of the gas turbine trajectory (surging and rotary breakdown), the Dobeski wavelet was selected 2nd order. At the same time, to monitor the wavelet coefficients behavior, the detail of the investigated signal is necessary at the 7th level of decomposition. Thus, the obtained results allow to carry out further analysis of breakdown processes using low-order wavelets.