Approximate entropy identification of flow unsteadiness and surge precursor in both subsonic centrifugal and axial compressors
Ben Zhao, Bingjie Li, Weiwei Cui, Chen Huang, Teng Zhou
Abstract
Surge poses a significant challenge to the safe operation of compressors. Identifying surge precursors is a valuable approach for surge warning and has been extensively investigated, though it remains unresolved. This paper experimentally examines the flow unsteadiness in centrifugal and axial subsonic compressor setups. To quantitatively assess flow unsteadiness and identify surge precursors, approximate entropy (ApEn) is applied to pressure and sound signals. The signals are collected under two compressor operating conditions: (1) stable operating points, where the valve at the compressor outlet remains fixed, and (2) the transient process during which the valve opening is gradually reduced until surge occurs. In the stable operating region, the lowest ApEn values are observed at medium flow rates, indicating high signal regularity and low flow unsteadiness. This behavior is attributed to the optimal incidence angle at the inducer inlet, corresponding to highly stable internal compressor flow. As the compressor operating point transitions from a stable region to surge onset, ApEn initially decreases due to the weakening of high-frequency unsteady flow, then increases just before surge occurs. The ApEn assessments reveal a similar trend in both centrifugal and axial compressors. These findings suggest that the variation in flow unsteadiness can make ApEn a promising method for quantifying the flow unsteadiness and identifying the surge precursor.