1 January 2025 in Physics of Fluids
Centrifugal compressors play a crucial role in energy and power systems, and compressor surge significantly impact their safe operation. Monitoring typical surge inception characteristics in a compressor can effectively serve as an early warning indicator, thereby enhancing safety. This paper presents the results of an acoustic experiment conducted on a centrifugal compressor setup. Four microphones were positioned in front of the compressor inlet to record the noise emitted as the compressor entered its surge region from a near-surge state. An entropy-based method is proposed to extract surge inception from acoustic signals by evaluating the informational complexity. Four types of entropy, including approximate entropy, sampling entropy, fuzzy entropy, and information entropy, were calculated from the acoustic signals. Their time-domain curves exhibit a clear increasing trend prior to surge inception. Further analysis reveals differing growth rates of the entropy curves among the four microphones, confirming the impact of microphone position on the capture of flow information and, consequently, on surge warning accuracy. Eventually, an entropy-weight method is proposed to eliminate the effects of positions and enhance surge warning capabilities by accounting for multiple acoustic signals.
1 November 2025 in Physics of Fluids
We present an image-based approach for early detection of surge precursors in a centrifugal compressor. One-dimensional pressure signals are mapped to Gramian angular field images, from which Gray Level Co-occurrence Matrix correlation is computed as the primary indicator using fixed windowing and gray-level quantization. On a single test rig at 60 000 and 70 000 rpm, the correlation trace rises prior to measured surge onset and follows the approach–surge–recovery evolution (e.g., a surge interval near 19.41–19.51 s at 60 000 rpm), enabling a baseline-referenced threshold for monitoring. To stabilize short-window fluctuations, a lightweight exponentially weighted moving average (α = 0.1) produces smoother yet responsive trajectories without resorting to black-box training. In the stable operating map, correlation exhibits a V-shaped variation with decreasing mass flow—coherent at overload, weakest near peak efficiency, and strengthening again as surge is approached—providing a physically interpretable trend. Within this scope, the study is a compressor-focused proof-of-concept; extension to other unsteady-flow problems is plausible but will require problem-specific tuning and independent validation.
1 March 2025 in Physics of Fluids
Surge waves are commonly modeled as one-dimensional plane waves in the understanding of compressor surge. However, this assumption does not adequately represent the physical phenomena observed under specific conditions. This study performed surge experiments on a centrifugal compressor, measuring pressure changes around surge onset with fast dynamic pressure transducers arranged at the same impeller inlet section but at different circumferential positions. The pressure signals are analyzed by the Hilbert–Huang transform, the ensemble empirical mode decomposition method, and the phase locking value concept. With them, instantaneous phase information was extracted, surge signal components were successfully isolated, and non-zero phase differences were identified. Those results demonstrate that surge waves exhibit significant circumferential propagation characteristics, challenging the conventional one-dimensional plane wave assumption. The proposed analytical approach was further validated with experimental data from two pressure-field microphones in an axial compressor experimental setup, thus confirming its validity in analyzing surge waves and the correctness of conclusions drawn from the centrifugal compressor. In conclusion, the novelty of this study is to offer experimental evidence and theoretical insights into the non-axially symmetric propagation of surge waves and an understanding of the fluid physics underlying compressor surge.
2 June 2026 in Proceedings of the Institution of Mechanical Engineers, Part C: Journal of Mechanical Engineering Science
Centrifugal compressors are critical components in diverse power systems. However, surge phenomena pose severe risks to both the thermodynamic stability of the cycle and the mechanical integrity of the compressor. Early and accurate detection of surge inception is therefore essential for effective control and protection. This study introduces a novel surge precursor identification method based on Rényi entropy analysis of acoustic emissions. Experiments were performed on an automotive turbocharger compressor, with four microphones mounted upstream of the inlet to record acoustic signals during transient operation approaching the surge limit. Conventional Rényi entropy was found to be insufficiently sensitive to surge onset due to overfitting. To overcome this limitation, an improved Rényi entropy method was developed by reconstructing the probability function. The improved entropy exhibited a clear increasing trend prior to surge inception, effectively indicating the compressor’s approach to instability. Comparative analyses demonstrate that the improved Rényi entropy surpasses conventional entropy-based indicators—such as approximate, sample, and fuzzy entropy—in both sensitivity and robustness. To verify the general applicability of the approach, a second experiment was conducted on a fuel cell centrifugal compressor equipped with 13 inlet microphones. The improved Rényi entropy method consistently detected surge inception in this configuration as well, confirming its reliability across different compressor types. These findings highlight the potential of the improved Rényi entropy as a powerful early-warning metric for surge detection in centrifugal compressors, marking an important step toward noise-based instability monitoring and control.
3 September 2025 in Proceedings of the Institution of Mechanical Engineers, Part C: Journal of Mechanical Engineering Science
One-and-a-half-stage counter-rotating turbines (CRTs) consist of one stator and two rotors; the blade rows rotate in opposite directions. The aerodynamic design process typically involves primary aerodynamic design, throughflow design, and three-dimensional geometry optimization. However, the process lacks an inversion model for propagating information in the opposite direction—from the throughflow design results back to the primary aerodynamic parameters, which is important for improving the accuracy of selecting the primary design parameters. This paper proposes a mathematical model based on Euler’s equation and the first law of thermodynamics, which establishes a connection between the primary design and throughflow design steps. This linkage enables reverse iteration from the throughflow design back to the primary aerodynamic design step to correct the primary design parameters, such as blade load coefficient, degree of reaction, and other input parameters of the throughflow design. The model was validated by using Reynolds-averaged Navier-Stokes (RANS) equations simulations with the Spalart-Allmaras (S-A) one-equation turbulence model, demonstrating a relative error ranging from 0.36% to 4.86%.
6 July 2026 in International Journal of Engine Research
In the actual operation of fuel-cell systems, providing the most accurate centrifugal compressor aerodynamic performance prediction in real-time is crucial for system efficiency and control. One key factor influencing the prediction is choke, which causes blade vibrations, limits the system’s maximum power and leads to a sharp decline in efficiency. However, the embedded multi-dimensional look-up tables typically necessitates a substantial amount of bench tests and makes extrapolation difficult. To address this, the study innovatively introduces a hybrid modeling framework that uses a small-sample physics-based model as a prior constraint and a data-driven method to compensate for physics-based prediction errors. The residual learning serves as a bridge trained with Gaussian process regression (GPR) by minimizing the root mean square error (RMSE). The hybrid model is implemented and validated on a fuel cell centrifugal compressor. Results show that, the physics-based prediction effectively constrains the shape of the characteristic curves, while the data-driven method significantly improves accuracy, especially in the choke region, with a 5% enhancement. In real-time prediction, as training samples expand, the accuracy improves further with the validation data almost all falling within the 95% confidence interval. The maximum training time is 20.96 s, indicating suitability for real-time prediction. This study highlights the potential of hybrid model of fuel-cell centrifugal compressors in engineering applications.