Abstract
A semi-classical signal analysis (SCSA) method is presented in this paper to show the relevance of its use for the detection of compressors surge precursors. This approach provides some new spectral parameters from physical properties of flow that might be interesting for signal analysis of turbomachines unsteadiness in general and surge detection in particular. Spectral parameters are obtained by considering the studied temporal signal as a trap-potential for a semi-classical particle and are represented by discrete energy levels given by the discrete spectrum of a Schrödinger operator. In this paper, it is shown that the new parameters are very sensitive to rapid and very small disturbances and present actually an alternative or new indices for the analysis of unsteadiness in turbomachinery. To illustrate its potential and its usefulness, the SCSA method is applied for the analysis of surge early signs and operating range of a centrifugal compressor.