Efficient and reliable surge prevention algorithm for centrifugal compressor

Abstract

Purpose The paper aims to present an outline of the technology of the active anti-surge algorithm based on high-frequency pressure measurement. The presented system is fast, inexpensive and reliable and does not limit the machine-operating range. Many contemporary anti-surge systems are based on theoretical surge margin. This solution limits machine operating range by about 10-15 per cent in the region of the highest pressure ratios. It is also often sensitive to change in external conditions such as temperature or density, as the system reacts to limits calculated theoretically. Design/methodology/approach This paper presents results of pressure measurements obtained on the low-speed centrifugal blower DP1.12. The pressure signals were presented in the form of phase diagrams, and conclusions were drawn from their phase portraits to develop the surge indication parameter. Findings The presented safety system uses the signal to develop the so-called (rate of derivative fluctuation) RDF parameter. In nominal working conditions, this parameter keeps the value close to 1. When RDF reaches values over 3, the anti-surge procedure should be implemented. Experimental studies have shown that this algorithm assures enough time to incur actions suppressing unstable phenomena. Originality/value The system reacts to real machine working conditions and is hence reliable. The RDF algorithm could also be used to identify local flow instabilities, as well as off-design operation.

Aerodynamic instabilities detection via empirical mode decomposition in centrifugal compressors

“…Pressure derivatives and resulting phase portraits can also be used as indicators of instabilities. Despite the big potential of this method, it requires thorough filtering of the signal and is limited to surge and stall detection [24,25]. Recent research shows that the use of Singular Spectrum Analysis (SSA) allows for detection of inlet recirculation [26], as some of the reconstructed components clearly capture this instability.…”