In order to identify the surge, multifractal spectrums are investigated in the analysis of dynamic pressure at the outlet of centrifugal compressor with 800 kW. At first, the time series of dynamic pressure, including stable and unstable states, are obtained. Then, the relationships between characteristics of multifractal spectrums and dynamic pressure series are studied and explained in detail. The results show that the multifractal spectrum parameter Δf(α) is narrow as the centrifugal compressor is working in stable state. However, Δf(α) becomes wide, as it is working in unstable states. Finally, probability analysis of multifractal variation tendency is used to describe the predictability of the variation. The studies indicated that the associated probability between the data of adjacent periods is highest and the predictability is the strongest in transition period. Thus, the method can be used to predict and avoid surge for centrifugal compressor in early time.
Full Text Finder