Abstract Multiresolution wavelet analysis of pressure variations in a gas turbine compressor reveals the existence of precursors of stall and surge processes. Signals from eight pressure sensors positioned at various places within the compressor were recorded and digitized in three different operating modes in stationary conditions with a recording interval of 1 ms during 5–6 s. It has been discovered that there exists a scale of 32 intervals over which the dispersion (variance) of the wavelet coefficients shows a remarkable drop of about 40% for more than 1 s prior to the development of the malfunction. A shuffled sample of the same values of the pressure does not show such a drop demonstrating the dynamical origin of this effect. Higher order correlation moments reveal different slopes in these two regions differing by the variance values. The log–log dependence of the moments does not show clear fractal behavior because the scales of 16 and 32 intervals are not on the straight line of monofractals. This is a clear indication of the nonlinear response of the system at this scale. These results provide a means for automatic regulation of an engine, preventing possible failures.
[1]
S. Thurner,et al.
Multiresolution Wavelet Analysis of Heartbeat Intervals Discriminates Healthy Patients from Those with Cardiac Pathology
,
1997,
adap-org/9711003.
[2]
I. J. Day.
Stall and surge in axial flow compressors
,
1992
.
[3]
V. A. Nechitailo,et al.
Wavelets and their uses
,
2001
.
[4]
I. Dremin,et al.
SCALING LAWS FOR DENSITY CORRELATIONS AND FLUCTUATIONS IN MULTIPARTICLE DYNAMICS
,
1995,
hep-ph/9508325.
[5]
L. Amaral,et al.
Multifractality in human heartbeat dynamics
,
1998,
Nature.
[6]
H E Stanley,et al.
Scale-independent measures and pathologic cardiac dynamics.
,
1998,
Physical review letters.
[7]
E. Greitzer.
Review—Axial Compressor Stall Phenomena
,
1980
.
[8]
Ingrid Daubechies,et al.
Ten Lectures on Wavelets
,
1992
.
[9]
Alexander G. Wilson,et al.
Stall Inception and Development in an Axial Flow Aeroengine
,
1993
.