Min Zhang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Min Zhang has a total of 40 co-authors. Here are the top co-authors by number of shared publications:
- Juan Du
Digital Twin Research Center, Institute of Engineering Thermophysics, Chinese Academy of Sciences 1 , Beijing 100190
18 shared publications - Dun Ba
University of Chinese Academy of Sciences Advanced Gas Turbine Laboratory, Institute of Engineering Thermophysics, Chinese Academy of Sciences No. 11, Beisihuanxi Road, Beijing 100190
10 shared publications - Zhonggang Fan9 shared publications
- Xiaobin Xu6 shared publications
- Jia-Hui Qiu3 shared publications
- Yang Liu
Digital Twin Research Center, Institute of Engineering Thermophysics, Chinese Academy of Sciences , Beijing 100190, China; , Christchurch 8140, New Zealand; , Beijing 100190, China
3 shared publications
Latest Publications
- 23 September 2026
- Published by Springer Nature in Journal of Thermal Science
- Vol. 35 (6) , 2447-2465
Multi-Dimensional Simulation of Turbofan Engine Performance by Integrating 2D Model of Fan-Bypass System
- 23 September 2026
- Published by Springer Nature in Journal of Thermal Science
- Vol. 35 (6) , 2447-2465
Multi-Dimensional Simulation of Turbofan Engine Performance by Integrating 2D Model of Fan-Bypass System
- 1 June 2026
- Published by AIP Publishing in Physics of Fluids
- Vol. 38 (6)
Dynamic mode decomposition of the surge modes in a two-stage high-speed fan
Flow instabilities such as rotating stall and surge limit the performance and stability of compressors. Although dynamic mode decomposition (DMD) has been widely applied to analyze unsteady compressor flows, most existing studies are restricted to single rotor configurations or single rotational speed and primarily focus on stall or mild surge regimes. Multi-stage coupling effects, rotational-speed-dependent instability mechanisms, and deep surge conditions remain insufficiently explored. In this study, full-annulus unsteady Reynolds-averaged Navier–Stokes simulations of a two-stage transonic fan are performed under multiple rotational speeds, and DMD is applied to extract the dominant modes during the deep surge process. A comparison between DMD and fast Fourier transform shows that the maximum relative frequency error reaches 11.7% at a low-frequency mode (0.015 blade passing frequency), while the error for other modes remains within 7%, confirming the reliability of the DMD analysis. The growth rate analysis indicates that at the design speed (1.0n), the second-stage rotor becomes unstable first, whereas at 0.8n, the first-stage rotor initiates instability. The dominant modes reveal three primary types of flow disturbances, including axial disturbances, circumferential disturbances, and self-excited tip leakage flow oscillations, which together form a self-sustained feedback loop driving the instability. The flow evolution further demonstrates that rotating stall reduces flow capacity, leading to blockage, and ultimately triggering severe reverse flow associated with deep surge. This study extends previous DMD-based investigations by revealing the instability onset locations in a two-stage fan and by capturing the full process from near-stall to deep surge.
- 1 June 2026
- Published by AIP Publishing in Physics of Fluids
- Vol. 38 (6)
Dynamic mode decomposition of the surge modes in a two-stage high-speed fan
Flow instabilities such as rotating stall and surge limit the performance and stability of compressors. Although dynamic mode decomposition (DMD) has been widely applied to analyze unsteady compressor flows, most existing studies are restricted to single rotor configurations or single rotational speed and primarily focus on stall or mild surge regimes. Multi-stage coupling effects, rotational-speed-dependent instability mechanisms, and deep surge conditions remain insufficiently explored. In this study, full-annulus unsteady Reynolds-averaged Navier–Stokes simulations of a two-stage transonic fan are performed under multiple rotational speeds, and DMD is applied to extract the dominant modes during the deep surge process. A comparison between DMD and fast Fourier transform shows that the maximum relative frequency error reaches 11.7% at a low-frequency mode (0.015 blade passing frequency), while the error for other modes remains within 7%, confirming the reliability of the DMD analysis. The growth rate analysis indicates that at the design speed (1.0n), the second-stage rotor becomes unstable first, whereas at 0.8n, the first-stage rotor initiates instability. The dominant modes reveal three primary types of flow disturbances, including axial disturbances, circumferential disturbances, and self-excited tip leakage flow oscillations, which together form a self-sustained feedback loop driving the instability. The flow evolution further demonstrates that rotating stall reduces flow capacity, leading to blockage, and ultimately triggering severe reverse flow associated with deep surge. This study extends previous DMD-based investigations by revealing the instability onset locations in a two-stage fan and by capturing the full process from near-stall to deep surge.
- 14 March 2026
- Published by Springer Nature in Journal of Thermal Science
- Vol. 35 (3) , 701-716
Performance Prediction of a Two-Stage High-Speed Axial Fan with Total Pressure Distorted Inflow Based on Body-Force Model
- 14 March 2026
- Published by Springer Nature in Journal of Thermal Science
- Vol. 35 (3) , 701-716
Performance Prediction of a Two-Stage High-Speed Axial Fan with Total Pressure Distorted Inflow Based on Body-Force Model
- 1 July 2025
- Published by AIP Publishing in Physics of Fluids
- Vol. 37 (7)
An inverse aerodynamic design framework for compressor blades based on generative model
The aerodynamic design of compressors is critical to the overall performance of aeroengines. Traditional compressor design methods are typically categorized into forward and inverse design. Forward design relies on predefined geometric parameters and iteratively adjusts them to achieve target performance, while inverse design begins with the desired performance specifications and directly derives the blade geometry. However, these existing methods exhibit significant limitations, including heavy reliance on expert knowledge, high computational complexity, and limited robustness. This paper proposes a generative inverse design framework based on a diffusion-driven gradient optimization network to mitigate the drawbacks of traditional methods. By integrating the strong global exploration capability of diffusion models with the efficient local adjustment ability of gradient-based optimization methods, the framework overcomes the limitations of single neural network models in complex design tasks. It takes key aerodynamic performance of the compressor, such as mass flow rate, isentropic efficiency, and total pressure ratio, as inputs and outputs compressor blade design parameters that meet these objectives. The framework is validated through the design of three-dimensional transonic axial compressor rotor blades. Results indicate that the framework can rapidly generate multiple viable geometries that meet the specified design requirements, with a maximum relative error of less than 1.5% compared to the target performance metrics. This significantly improves design efficiency and reduces dependence on the expert knowledge of designers, offering a promising direction for the intelligent design of turbomachinery.
- 1 July 2025
- Published by AIP Publishing in Physics of Fluids
- Vol. 37 (7)
An inverse aerodynamic design framework for compressor blades based on generative model
The aerodynamic design of compressors is critical to the overall performance of aeroengines. Traditional compressor design methods are typically categorized into forward and inverse design. Forward design relies on predefined geometric parameters and iteratively adjusts them to achieve target performance, while inverse design begins with the desired performance specifications and directly derives the blade geometry. However, these existing methods exhibit significant limitations, including heavy reliance on expert knowledge, high computational complexity, and limited robustness. This paper proposes a generative inverse design framework based on a diffusion-driven gradient optimization network to mitigate the drawbacks of traditional methods. By integrating the strong global exploration capability of diffusion models with the efficient local adjustment ability of gradient-based optimization methods, the framework overcomes the limitations of single neural network models in complex design tasks. It takes key aerodynamic performance of the compressor, such as mass flow rate, isentropic efficiency, and total pressure ratio, as inputs and outputs compressor blade design parameters that meet these objectives. The framework is validated through the design of three-dimensional transonic axial compressor rotor blades. Results indicate that the framework can rapidly generate multiple viable geometries that meet the specified design requirements, with a maximum relative error of less than 1.5% compared to the target performance metrics. This significantly improves design efficiency and reduces dependence on the expert knowledge of designers, offering a promising direction for the intelligent design of turbomachinery.
- 1 April 2025
- Published by AIP Publishing in Physics of Fluids
- Vol. 37 (4)
Experimental investigation of adaptive Coanda jet control for performance improvement of a highly loaded compressor cascade
With the increase in the blade load of modern compressors for aircraft engines and gas turbines, flow separation generated in stator blade passages especially under transient flight maneuvers or off-design operating conditions is enhanced. This not only deteriorates the compressor efficiency but also blockages the flow path that may induce instability such as stall or surge. Due to this, an adaptive Coanda jet control (ACJC) system has been developed based on a compressor stator cascade. This system can execute real-time monitoring the flow condition of the stator via an incidence angle prediction model and can inject appropriate air to eliminate flow separation based on an optimal injection mass flow rate prediction model. In order to further validate the active control effects of the ACJC system, this article conducts detailed experimental tests for the stator cascades. Inflow condition matrix is elaborately arranged by considering a wide range of inlet Mach number [0.4, 0.6] and incidence angle [−3°, 5°]. Results indicate that the incidence angle predicted by the ACJC system is nearly the same as the set value. Also, the total pressure loss is dramatically decreased due to the Coanda jet with a mass flow rate predicted by the ACJC system. In addition, compared to the cascade without ACJC under the incidence angel of 5°, the total pressure loss coefficient is reduced by 14.3%, 12.9%, and 17.8% at inflow Mach number being 0.4, 0.5, and 0.6, respectively.
- 1 April 2025
- Published by AIP Publishing in Physics of Fluids
- Vol. 37 (4)
Experimental investigation of adaptive Coanda jet control for performance improvement of a highly loaded compressor cascade
With the increase in the blade load of modern compressors for aircraft engines and gas turbines, flow separation generated in stator blade passages especially under transient flight maneuvers or off-design operating conditions is enhanced. This not only deteriorates the compressor efficiency but also blockages the flow path that may induce instability such as stall or surge. Due to this, an adaptive Coanda jet control (ACJC) system has been developed based on a compressor stator cascade. This system can execute real-time monitoring the flow condition of the stator via an incidence angle prediction model and can inject appropriate air to eliminate flow separation based on an optimal injection mass flow rate prediction model. In order to further validate the active control effects of the ACJC system, this article conducts detailed experimental tests for the stator cascades. Inflow condition matrix is elaborately arranged by considering a wide range of inlet Mach number [0.4, 0.6] and incidence angle [−3°, 5°]. Results indicate that the incidence angle predicted by the ACJC system is nearly the same as the set value. Also, the total pressure loss is dramatically decreased due to the Coanda jet with a mass flow rate predicted by the ACJC system. In addition, compared to the cascade without ACJC under the incidence angel of 5°, the total pressure loss coefficient is reduced by 14.3%, 12.9%, and 17.8% at inflow Mach number being 0.4, 0.5, and 0.6, respectively.
- Open Access
- 1 December 2024
- Published by AIP Publishing in Physics of Fluids
- Vol. 36 (12)
Experimental investigation on stability-enhancement mechanism of tip air injection in an axial-flow compressor with circumferential distortions
We conduct experimental investigations of the effect of the tip air injection as a stability control method on a low-speed axial-flow compressor experiencing severe circumferential total pressure distortion. Eight Coanda-shaped injectors, uniformly distributed upstream of the rotor blade leading edge, were employed to counteract stall margin degradation caused by a flat-baffle circumferential distortion. Unsteady pressure data from the tip clearance and rotor wake regions were captured using time-resolved sensors and probes and analyzed through time-frequency transformation and proper orthogonal decomposition (POD) techniques. POD investigation under distorted inflow conditions reveals that the third-order POD mode, which characterizes unsteady tip leakage flow (TLF), plays a dominant role in triggering stall. Based on this insight, tip air injection with varying momentum ratios was applied to enhance the stall margin. The positive correlation between the injected momentum ratio and stall margin improvement was established, demonstrating the stabilizing effect of tip air injection under the circumferential distortion. The results indicate that tip air injection primarily acts on the rotor blade tip region to suppress the unsteady TLF at the center and downstream areas of the distortion, leading to an increase in tip blade load and a reduction in both the scale and number of pre-stall disturbances. However, the injection has a minimal impact on the unsteady TLF frequency band in the rotor wake region. Additionally, the POD analysis confirms that as the injection rate is increased, the energy percentage, flow field intensity, and power spectral density amplitude of the third-order POD mode are decreased, further reflecting the mitigation of unsteady TLF. The present work provides important insight into the stability control mechanisms of a tip air injection under circumferential distortion, offering design guidance for the implementation of active stability control strategies in axial-flow compressors.
- 1 December 2024
- Published by AIP Publishing in Physics of Fluids
- Vol. 36 (12)
Open AccessExperimental investigation on stability-enhancement mechanism of tip air injection in an axial-flow compressor with circumferential distortions
We conduct experimental investigations of the effect of the tip air injection as a stability control method on a low-speed axial-flow compressor experiencing severe circumferential total pressure distortion. Eight Coanda-shaped injectors, uniformly distributed upstream of the rotor blade leading edge, were employed to counteract stall margin degradation caused by a flat-baffle circumferential distortion. Unsteady pressure data from the tip clearance and rotor wake regions were captured using time-resolved sensors and probes and analyzed through time-frequency transformation and proper orthogonal decomposition (POD) techniques. POD investigation under distorted inflow conditions reveals that the third-order POD mode, which characterizes unsteady tip leakage flow (TLF), plays a dominant role in triggering stall. Based on this insight, tip air injection with varying momentum ratios was applied to enhance the stall margin. The positive correlation between the injected momentum ratio and stall margin improvement was established, demonstrating the stabilizing effect of tip air injection under the circumferential distortion. The results indicate that tip air injection primarily acts on the rotor blade tip region to suppress the unsteady TLF at the center and downstream areas of the distortion, leading to an increase in tip blade load and a reduction in both the scale and number of pre-stall disturbances. However, the injection has a minimal impact on the unsteady TLF frequency band in the rotor wake region. Additionally, the POD analysis confirms that as the injection rate is increased, the energy percentage, flow field intensity, and power spectral density amplitude of the third-order POD mode are decreased, further reflecting the mitigation of unsteady TLF. The present work provides important insight into the stability control mechanisms of a tip air injection under circumferential distortion, offering design guidance for the implementation of active stability control strategies in axial-flow compressors.
- 12 November 2024
- Published by Wiley in Journal of Turbomachinery
- Vol. 147 (1) , 1-32
Optimized Casing Treatment to Improve Stall Margin Under Circumferential Pressure Distortion in a High-Speed Axial Flow Fan
An optimized axial slot casing treatment (OPCT) was designed that experimentally improves the stall margin by 7.31% with an efficiency gain of 1.94% at a design speed of 100% in a high-speed axial flow fan. Subsequently, the OPCT was tested under the distorted inflow with a 9% composite distortion index; it can still improve the stall margin by 6.19% with an efficiency gain of 1.63% at the 100% design speed. The unsteady measurement results indicate that the location of the shock wave obviously moves forward toward the leading edge of the blade tip as the rotor blade rotates out of the distorted region and induces the stall in advance. When the OPCT is applied, the strength of the interaction between the shock wave and tip leakage flow can be suppressed, and the location of the shock wave at the distorted region is pushed backward. Thus, the trigger of the stall inception generated in the downstream of the distortion region is postponed. Meanwhile, a new phenomenon was discovered that the OPCT does not alter the instability mode of the fan, and the fan eventually surges, which is consistent with the B parameter value of 2.942. However, under the distorted inflow, the B parameter can be reduced from 2.942 to 0.863, thereby the instability mode of the fan is changed from surge to stall. It can guide the instability prediction when the boundary condition (distortion or casing treatment) of the fan is changed.
- 12 November 2024
- Published by Wiley in Journal of Turbomachinery
- Vol. 147 (1) , 1-32
Optimized Casing Treatment to Improve Stall Margin Under Circumferential Pressure Distortion in a High-Speed Axial Flow Fan
An optimized axial slot casing treatment (OPCT) was designed that experimentally improves the stall margin by 7.31% with an efficiency gain of 1.94% at a design speed of 100% in a high-speed axial flow fan. Subsequently, the OPCT was tested under the distorted inflow with a 9% composite distortion index; it can still improve the stall margin by 6.19% with an efficiency gain of 1.63% at the 100% design speed. The unsteady measurement results indicate that the location of the shock wave obviously moves forward toward the leading edge of the blade tip as the rotor blade rotates out of the distorted region and induces the stall in advance. When the OPCT is applied, the strength of the interaction between the shock wave and tip leakage flow can be suppressed, and the location of the shock wave at the distorted region is pushed backward. Thus, the trigger of the stall inception generated in the downstream of the distortion region is postponed. Meanwhile, a new phenomenon was discovered that the OPCT does not alter the instability mode of the fan, and the fan eventually surges, which is consistent with the B parameter value of 2.942. However, under the distorted inflow, the B parameter can be reduced from 2.942 to 0.863, thereby the instability mode of the fan is changed from surge to stall. It can guide the instability prediction when the boundary condition (distortion or casing treatment) of the fan is changed.
- 1 November 2024
- Published by AIP Publishing in Physics of Fluids
- Vol. 36 (11)
Fast prediction of compressor flow field based on a deep attention symmetrical neural network
Accurate and rapid prediction of compressor performance and key flow characteristics is critical for digital design, digital twin modeling, and virtual–real interaction. However, the traditional methods of obtaining flow field parameters by solving the Navier–Stokes equations are computationally intensive and time-consuming. To establish a digital twin model of the flow field in a transonic three-stage axial compressor, this study proposes a novel data-driven deep attention symmetric neural network for fast reconstruction of the flow field at different blade rows and spanwise positions. The network integrates a vision transformer (ViT) and a symmetric convolutional neural network (SCNN). The ViT extracts geometric features from the blade passages. The SCNN is used for deeper extraction of input features such as boundary conditions and flow coordinates, enabling precise flow field predictions. Results indicate that the trained model can efficiently and accurately reconstruct the internal flow field of the compressor in 0.5 s, capturing phenomena such as flow separation and wake. Compared with traditional numerical simulations, the current model offers significant advantages in computational speed, delivering a three-order magnitude speedup compared to computational fluid dynamics simulations. It shows strong potential for engineering applications and provides robust support for building digital twin models in turbomachinery flow fields.
- 1 November 2024
- Published by AIP Publishing in Physics of Fluids
- Vol. 36 (11)
Fast prediction of compressor flow field based on a deep attention symmetrical neural network
Accurate and rapid prediction of compressor performance and key flow characteristics is critical for digital design, digital twin modeling, and virtual–real interaction. However, the traditional methods of obtaining flow field parameters by solving the Navier–Stokes equations are computationally intensive and time-consuming. To establish a digital twin model of the flow field in a transonic three-stage axial compressor, this study proposes a novel data-driven deep attention symmetric neural network for fast reconstruction of the flow field at different blade rows and spanwise positions. The network integrates a vision transformer (ViT) and a symmetric convolutional neural network (SCNN). The ViT extracts geometric features from the blade passages. The SCNN is used for deeper extraction of input features such as boundary conditions and flow coordinates, enabling precise flow field predictions. Results indicate that the trained model can efficiently and accurately reconstruct the internal flow field of the compressor in 0.5 s, capturing phenomena such as flow separation and wake. Compared with traditional numerical simulations, the current model offers significant advantages in computational speed, delivering a three-order magnitude speedup compared to computational fluid dynamics simulations. It shows strong potential for engineering applications and provides robust support for building digital twin models in turbomachinery flow fields.
- 1 November 2024
- Published by AIP Publishing in Physics of Fluids
- Vol. 36 (11)
Adaptive jet flow control for performance improvement via coupling stator-blade with rotor-tip injections in a transonic compressor
With the increase in the blade load of modern compressors for aircraft engines and gas turbines, flow separation is always generated in stator and rotor blade passages. This not only deteriorates the compressor efficiency but also blocks the flow path that may induce instability such as stall or surge. In order to tackle these issues, an adaptive jet flow control system coupling stator-blade with rotor-tip injections is proposed, and the goal of this paper is to investigate its effects on the aerodynamic performance of an in-house 1.5 stage transonic compressor. The developed flow control system integrates a jet configuration on the stator-blade surface and on the rotor casing surface respectively. The first step to construct this system is designing the jet configurations, and the Coanda effect is adopted to achieve wall-attachment injection. Second, the flow control system is intended to execute real-time monitoring on the operating condition of the compressor and can inject air with appropriate mass flow rate to eliminate flow separation and rotating stall. To do this, an optimal injection mass flow rate prediction model is established via a back propagation neural network algorithm. It aids to ensure an adaptive jet mass flow rate control during the compressor operation. Then, the effectiveness of the adaptive jet flow control system is evaluated using numerical simulations, and the influence mechanism of the injection mass flow on flow fields and aerodynamic performance is analyzed. Results indicate that compared to the prototype compressor, the compressor with adaptive stator-blade and rotor-tip injections can significantly increase efficiency and improve stability margin.
- 1 November 2024
- Published by AIP Publishing in Physics of Fluids
- Vol. 36 (11)
Adaptive jet flow control for performance improvement via coupling stator-blade with rotor-tip injections in a transonic compressor
With the increase in the blade load of modern compressors for aircraft engines and gas turbines, flow separation is always generated in stator and rotor blade passages. This not only deteriorates the compressor efficiency but also blocks the flow path that may induce instability such as stall or surge. In order to tackle these issues, an adaptive jet flow control system coupling stator-blade with rotor-tip injections is proposed, and the goal of this paper is to investigate its effects on the aerodynamic performance of an in-house 1.5 stage transonic compressor. The developed flow control system integrates a jet configuration on the stator-blade surface and on the rotor casing surface respectively. The first step to construct this system is designing the jet configurations, and the Coanda effect is adopted to achieve wall-attachment injection. Second, the flow control system is intended to execute real-time monitoring on the operating condition of the compressor and can inject air with appropriate mass flow rate to eliminate flow separation and rotating stall. To do this, an optimal injection mass flow rate prediction model is established via a back propagation neural network algorithm. It aids to ensure an adaptive jet mass flow rate control during the compressor operation. Then, the effectiveness of the adaptive jet flow control system is evaluated using numerical simulations, and the influence mechanism of the injection mass flow on flow fields and aerodynamic performance is analyzed. Results indicate that compared to the prototype compressor, the compressor with adaptive stator-blade and rotor-tip injections can significantly increase efficiency and improve stability margin.
- Open Access
- 8 October 2024
- Published by Global Power and Propulsion Society in Journal of the Global Power and Propulsion Society
- Vol. 8, 390-404
Analysis on stall mechanism of axial flow compressor with rotating inlet distortion by dynamic mode decomposition
To study the impact of rotating inlet distortion on stall mechanism of an axial compressor, the full-annulus unsteady simulation results are analysed using dynamic mode decomposition (DMD). Firstly, the frequency and energy ratio of each mode at near stall condition for the compressor with uniform inflow are obtained. A dominating frequency band 0.3–0.6 BPF regarded as the characteristic frequency of unsteady tip leakage flow is observed. The distribution of eigenvalues further confirms that there are a few critical unstable modes in the flow field. With further throttling, four large-scale disturbance structures are captured by the second-order mode, whose frequency is twice as the rotor frequency. The circumferential propagation velocity of the stall cell is thus deduced to be about 50% of the rotor speed. Further, when the compressor is subjected to rotating inlet distortion, the change of stalling process is related to the relative position of distortion plate and stall cell according to the DMD mode frequency and shape. The addition of rotating distortion does not change the stall type of the compressor, but the unstable modes indicate that the number and shape of the stall cells have been varied.
- 8 October 2024
- Published by Global Power and Propulsion Society in Journal of the Global Power and Propulsion Society
- Vol. 8, 390-404
Open AccessAnalysis on stall mechanism of axial flow compressor with rotating inlet distortion by dynamic mode decomposition
To study the impact of rotating inlet distortion on stall mechanism of an axial compressor, the full-annulus unsteady simulation results are analysed using dynamic mode decomposition (DMD). Firstly, the frequency and energy ratio of each mode at near stall condition for the compressor with uniform inflow are obtained. A dominating frequency band 0.3–0.6 BPF regarded as the characteristic frequency of unsteady tip leakage flow is observed. The distribution of eigenvalues further confirms that there are a few critical unstable modes in the flow field. With further throttling, four large-scale disturbance structures are captured by the second-order mode, whose frequency is twice as the rotor frequency. The circumferential propagation velocity of the stall cell is thus deduced to be about 50% of the rotor speed. Further, when the compressor is subjected to rotating inlet distortion, the change of stalling process is related to the relative position of distortion plate and stall cell according to the DMD mode frequency and shape. The addition of rotating distortion does not change the stall type of the compressor, but the unstable modes indicate that the number and shape of the stall cells have been varied.