Rundi Qiu
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Rundi Qiu has a total of 25 co-authors. Here are the top co-authors by number of shared publications:
- Yiwei Wang
Chinese Academy of Sciences
4 shared publications - Jingzhu Wang
Chinese Academy of Sciences
3 shared publications - Yongjiu Wang
Key Laboratory for Mechanics in Fluid-Solid Coupling Systems, Institute of Mechanics, Chinese Academy of Sciences 2 , Beijing 100190
2 shared publications - Chichao Huang1 shared publication
- Chun Fan
Peking University
1 shared publication Guanghang Wang
1 shared publication
Latest Publications
- 1 September 2026
- Published by AIP Publishing in Physics of Fluids
- Vol. 38 (9)
Physics-informed super-resolution of dual-cylinder wake flow fields from coarse-grid data
High-resolution flow fields are essential for resolving wake interaction and pressure-coupled unsteady features in bluff-body flows, yet their acquisition from experiments or high-fidelity simulations remains expensive. In dual-cylinder configurations, the interaction between the cylinders can substantially alter the near-wake topology, pressure evolution, and dominant vortex-shedding characteristics, making physically reliable reconstruction from coarse data particularly challenging. In this study, physics-informed neural networks (PINNs) are employed for super-resolution reconstruction of dual-cylinder flow fields at Reynolds number 200 using direct numerical simulation data. To address the non-uniqueness of pressure reconstruction in standard PINNs, a pressure-constrained PINNs (PC-PINNs) framework is proposed by introducing sparse pressure supervision at specific locations to anchor the absolute pressure field. Furthermore, a transfer-learning-enhanced framework (TL-PINNs) is developed to accelerate convergence across related cylinder arrangements. Rather than focusing only on reconstruction accuracy, the present study examines whether the reconstructed fields preserve physically meaningful observables, including near-wake structures, pressure distributions, pressure-fluctuation histories, and dominant spectral signatures. The results show that standard PINNs can recover the major velocity structures but suffer from pressure drift, whereas PC-PINNs markedly improve pressure reconstruction and enable more reliable recovery of pressure-related wake features. TL-PINNs further reduce the training cost while maintaining predictive fidelity and physical interpretability. These findings demonstrate the potential of pressure-constrained physics-informed reconstruction for mechanism-oriented analysis of interference-induced bluff-body flows from coarse-resolution data.
- 1 September 2026
- Published by AIP Publishing in Physics of Fluids
- Vol. 38 (9)
Physics-informed super-resolution of dual-cylinder wake flow fields from coarse-grid data
High-resolution flow fields are essential for resolving wake interaction and pressure-coupled unsteady features in bluff-body flows, yet their acquisition from experiments or high-fidelity simulations remains expensive. In dual-cylinder configurations, the interaction between the cylinders can substantially alter the near-wake topology, pressure evolution, and dominant vortex-shedding characteristics, making physically reliable reconstruction from coarse data particularly challenging. In this study, physics-informed neural networks (PINNs) are employed for super-resolution reconstruction of dual-cylinder flow fields at Reynolds number 200 using direct numerical simulation data. To address the non-uniqueness of pressure reconstruction in standard PINNs, a pressure-constrained PINNs (PC-PINNs) framework is proposed by introducing sparse pressure supervision at specific locations to anchor the absolute pressure field. Furthermore, a transfer-learning-enhanced framework (TL-PINNs) is developed to accelerate convergence across related cylinder arrangements. Rather than focusing only on reconstruction accuracy, the present study examines whether the reconstructed fields preserve physically meaningful observables, including near-wake structures, pressure distributions, pressure-fluctuation histories, and dominant spectral signatures. The results show that standard PINNs can recover the major velocity structures but suffer from pressure drift, whereas PC-PINNs markedly improve pressure reconstruction and enable more reliable recovery of pressure-related wake features. TL-PINNs further reduce the training cost while maintaining predictive fidelity and physical interpretability. These findings demonstrate the potential of pressure-constrained physics-informed reconstruction for mechanism-oriented analysis of interference-induced bluff-body flows from coarse-resolution data.
- 1 June 2026
- Published by AIP Publishing in Physics of Fluids
- Vol. 38 (6)
Investigation of the air film around an axisymmetric projectile with a pressure-equalizing exhaust system
This study numerically investigates air film flow around a projectile with a pressure-equalizing exhaust system. Dimensional analysis is employed to identify the key dimensionless parameters governing flow regime transitions. The air film evolution is categorized into three distinct stages: rapid growth, periodic pulsation, and quasi-steady. The characteristics of the air film and the associated development of the flow field are analyzed. Analysis reveals that the reentrant jet strength and gas injection rate are the primary factors driving transitions in the flow pattern. The velocity vector field reveals that the upstream-propagating reentrant jet dominates interfacial instability and periodic air film shedding, governing the flow regime transition of the air film. During the rapid growth stage, the pressure gradient exhibits minor fluctuations. However, it intensifies markedly upon transition into the periodic pulsation stage. Wavelet analysis reveals that frictional drag is linked to air film geometry, while tail cavity pulsations dominate pressure drag. Pressure drag constitutes ∼95% of the total drag, driven by high-pressure pulsation from the tail cavity, which is affected by the shedding of air film convected downstream. This study provides valuable insights for the engineering design and active flow control of projectiles.
- 1 June 2026
- Published by AIP Publishing in Physics of Fluids
- Vol. 38 (6)
Investigation of the air film around an axisymmetric projectile with a pressure-equalizing exhaust system
This study numerically investigates air film flow around a projectile with a pressure-equalizing exhaust system. Dimensional analysis is employed to identify the key dimensionless parameters governing flow regime transitions. The air film evolution is categorized into three distinct stages: rapid growth, periodic pulsation, and quasi-steady. The characteristics of the air film and the associated development of the flow field are analyzed. Analysis reveals that the reentrant jet strength and gas injection rate are the primary factors driving transitions in the flow pattern. The velocity vector field reveals that the upstream-propagating reentrant jet dominates interfacial instability and periodic air film shedding, governing the flow regime transition of the air film. During the rapid growth stage, the pressure gradient exhibits minor fluctuations. However, it intensifies markedly upon transition into the periodic pulsation stage. Wavelet analysis reveals that frictional drag is linked to air film geometry, while tail cavity pulsations dominate pressure drag. Pressure drag constitutes ∼95% of the total drag, driven by high-pressure pulsation from the tail cavity, which is affected by the shedding of air film convected downstream. This study provides valuable insights for the engineering design and active flow control of projectiles.
- 1 March 2026
- Published by AIP Publishing in Physics of Fluids
- Vol. 38 (3)
Numerical study of the behaviors of ventilated supercavity under the heave motion
In the operational environment of underwater vehicles, the heave motion directly influences the ventilated supercavity flow field. However, the impact mechanism of the ventilated supercavity characteristics has not been fully revealed. Therefore, this study focuses on the transient evolution characteristics of the ventilated supercavity under the heave motion of the cavitator, aiming to fill the gap in this area of research. The numerical simulations employ the volume of fluid method to capture the gas–liquid interface, coupled with a k–ε turbulence model to describe turbulent flow characteristics. The periodic heave motion conditions of the cavitator are constructed by superimposing cosine functions and using an overlapping grid technique. After the model's validity is verified against experimental data, a series of numerical simulations is conducted at heave frequencies of 10, 5, and 2.5 Hz. The results indicate that the heave motion of the cavitator significantly alters the geometry of the ventilated supercavity, causing its surface curvature to change in a regular pattern and exhibit a wave-like distribution. Under fixed heave displacement conditions, an increase in heave frequency further intensifies the wave-like fluctuation of the cavity, and the curvature along the centerline significantly increases. Mechanism analysis shows that the heave motion of the cavitator alters the vertical velocity distribution near the cavity surfaces, leading to changes in the curvature of the upper and lower surfaces of the cavity, thus inducing dynamic evolution of its overall geometry. These findings provide key theoretical support and technical reference for the structural optimization and active control of ventilated supercavities.
- 1 March 2026
- Published by AIP Publishing in Physics of Fluids
- Vol. 38 (3)
Numerical study of the behaviors of ventilated supercavity under the heave motion
In the operational environment of underwater vehicles, the heave motion directly influences the ventilated supercavity flow field. However, the impact mechanism of the ventilated supercavity characteristics has not been fully revealed. Therefore, this study focuses on the transient evolution characteristics of the ventilated supercavity under the heave motion of the cavitator, aiming to fill the gap in this area of research. The numerical simulations employ the volume of fluid method to capture the gas–liquid interface, coupled with a k–ε turbulence model to describe turbulent flow characteristics. The periodic heave motion conditions of the cavitator are constructed by superimposing cosine functions and using an overlapping grid technique. After the model's validity is verified against experimental data, a series of numerical simulations is conducted at heave frequencies of 10, 5, and 2.5 Hz. The results indicate that the heave motion of the cavitator significantly alters the geometry of the ventilated supercavity, causing its surface curvature to change in a regular pattern and exhibit a wave-like distribution. Under fixed heave displacement conditions, an increase in heave frequency further intensifies the wave-like fluctuation of the cavity, and the curvature along the centerline significantly increases. Mechanism analysis shows that the heave motion of the cavitator alters the vertical velocity distribution near the cavity surfaces, leading to changes in the curvature of the upper and lower surfaces of the cavity, thus inducing dynamic evolution of its overall geometry. These findings provide key theoretical support and technical reference for the structural optimization and active control of ventilated supercavities.
- 1 August 2025
- Published by AIP Publishing in Physics of Fluids
- Vol. 37 (8)
Phase-field physics-informed neural networks for multiphase flow simulations in axisymmetric coordinates with applications of bubble rising in multiple regimes
While physics-informed neural networks are becoming a promising approach for multiphase flow simulations, the transition from two-dimensional to three-dimensional applications is primarily hindered by computational inefficiency. This study presents a phase-field Physics-Informed Neural Networks (PF-PINNs) specifically designed for axisymmetric multiphase flow dynamics. Adaptive time marching strategy and axisymmetric weighted sampling strategy are introduced to ensure training convergence. The framework is validated through bubble rising simulations in infinite fluid domain. Results show that the shape of the interface coincides well with numerical validation. Dynamic behaviors, including centroid trajectories, rising velocity and velocity fields are precisely captured. We also evaluate the applicable range of this method with cases over different Reynolds numbers and Bond numbers. The prediction succeeds in Re=[10, 200] and Bo=[10, 200], which demonstrates the generality of PF-PINNs in axisymmetric coordinate and the ability to capture different bubble dynamics in different regimes.
- 1 August 2025
- Published by AIP Publishing in Physics of Fluids
- Vol. 37 (8)
Phase-field physics-informed neural networks for multiphase flow simulations in axisymmetric coordinates with applications of bubble rising in multiple regimes
While physics-informed neural networks are becoming a promising approach for multiphase flow simulations, the transition from two-dimensional to three-dimensional applications is primarily hindered by computational inefficiency. This study presents a phase-field Physics-Informed Neural Networks (PF-PINNs) specifically designed for axisymmetric multiphase flow dynamics. Adaptive time marching strategy and axisymmetric weighted sampling strategy are introduced to ensure training convergence. The framework is validated through bubble rising simulations in infinite fluid domain. Results show that the shape of the interface coincides well with numerical validation. Dynamic behaviors, including centroid trajectories, rising velocity and velocity fields are precisely captured. We also evaluate the applicable range of this method with cases over different Reynolds numbers and Bond numbers. The prediction succeeds in Re=[10, 200] and Bo=[10, 200], which demonstrates the generality of PF-PINNs in axisymmetric coordinate and the ability to capture different bubble dynamics in different regimes.
- 1 May 2025
- Published by AIP Publishing in Physics of Fluids
- Vol. 37 (5)
A multi-scale online method to predict the unsteady pressure of the ventilated cavities around an axisymmetric body
During the underwater motion of an axisymmetric body, unsteady shedding of the ventilated cavity causes severe pressure fluctuations on its downstream surface, affecting the motion stability. A traditional control system relies on monitoring data for posture adjustment, but may fail due to the hysteresis effect. To address this, a multi-scale online method is proposed to predict the unsteady pressure caused by cavitation shedding. This method decomposes the unsteady pressure into two scales: large-scale pressure, predicted using a medium support vector regression (SVR) model, and small-scale fluctuating pressure, predicted via a multi-round online deployment (MROD) method. The MROD method employs an offline-trained double-layer long short-term memory network, iteratively invoked to intermittently incorporate real-time data for advanced predictions. The prediction accuracy and speed of this method are influenced by key hyperparameters, including the input sequence length, output sequence length, real-time interval between time steps, and time step interval between consecutive real-time data inputs. Results show that both MROD and SVR models exhibit high prediction accuracy and robust generalization ability for predicting the small-scale fluctuating pressure and large-scale pressure, respectively. The proposed method achieves weighted mean relative errors below 1% for both interpolation and extrapolation of unsteady pressure, demonstrating its effectiveness in predicting unsteady pressure for axisymmetric bodies under unknown operating conditions. This high-accuracy prediction ensures stable motion of the axisymmetric body in complex marine environments.
- 1 May 2025
- Published by AIP Publishing in Physics of Fluids
- Vol. 37 (5)
A multi-scale online method to predict the unsteady pressure of the ventilated cavities around an axisymmetric body
During the underwater motion of an axisymmetric body, unsteady shedding of the ventilated cavity causes severe pressure fluctuations on its downstream surface, affecting the motion stability. A traditional control system relies on monitoring data for posture adjustment, but may fail due to the hysteresis effect. To address this, a multi-scale online method is proposed to predict the unsteady pressure caused by cavitation shedding. This method decomposes the unsteady pressure into two scales: large-scale pressure, predicted using a medium support vector regression (SVR) model, and small-scale fluctuating pressure, predicted via a multi-round online deployment (MROD) method. The MROD method employs an offline-trained double-layer long short-term memory network, iteratively invoked to intermittently incorporate real-time data for advanced predictions. The prediction accuracy and speed of this method are influenced by key hyperparameters, including the input sequence length, output sequence length, real-time interval between time steps, and time step interval between consecutive real-time data inputs. Results show that both MROD and SVR models exhibit high prediction accuracy and robust generalization ability for predicting the small-scale fluctuating pressure and large-scale pressure, respectively. The proposed method achieves weighted mean relative errors below 1% for both interpolation and extrapolation of unsteady pressure, demonstrating its effectiveness in predicting unsteady pressure for axisymmetric bodies under unknown operating conditions. This high-accuracy prediction ensures stable motion of the axisymmetric body in complex marine environments.
- 1 September 2024
- Published by AIP Publishing in Physics of Fluids
- Vol. 36 (9)
Modeling two-phase flows with complicated interface evolution using parallel physics-informed neural networks
The physics-informed neural networks (PINNs) have shown great potential in solving a variety of high-dimensional partial differential equations (PDEs), but the complexity of a realistic problem still restricts the practical application of the PINNs for solving most complicated PDEs. In this paper, we propose a parallel framework for PINNs that is capable of modeling two-phase flows with complicated interface evolution. The proposed framework divides the problem into several simplified subproblems and solves them through training several PINNs on corresponding subdomains simultaneously. To enhance the accuracy of the parallel training framework in two-phase flow, the overlapping domain decomposition method is adopted. The optimal subnetwork sizes and partitioned method are systematically discussed, and a series of cases including a bubble rising, droplet splashing, and the Rayleigh–Taylor instability are applied for quantitative validation. The maximum relative error of quantitative values in these cases is 0.1319. Our results show that the proposed framework not only can accelerate the training procedure of PINNs, but also can capture the spatiotemporal evolution of the interface between various phases. This framework overcomes the difficulties of training PINNs to solve a forward problem in two-phase flow, and it is expected to model more realistic dynamic systems in nature.
- 1 September 2024
- Published by AIP Publishing in Physics of Fluids
- Vol. 36 (9)
Modeling two-phase flows with complicated interface evolution using parallel physics-informed neural networks
The physics-informed neural networks (PINNs) have shown great potential in solving a variety of high-dimensional partial differential equations (PDEs), but the complexity of a realistic problem still restricts the practical application of the PINNs for solving most complicated PDEs. In this paper, we propose a parallel framework for PINNs that is capable of modeling two-phase flows with complicated interface evolution. The proposed framework divides the problem into several simplified subproblems and solves them through training several PINNs on corresponding subdomains simultaneously. To enhance the accuracy of the parallel training framework in two-phase flow, the overlapping domain decomposition method is adopted. The optimal subnetwork sizes and partitioned method are systematically discussed, and a series of cases including a bubble rising, droplet splashing, and the Rayleigh–Taylor instability are applied for quantitative validation. The maximum relative error of quantitative values in these cases is 0.1319. Our results show that the proposed framework not only can accelerate the training procedure of PINNs, but also can capture the spatiotemporal evolution of the interface between various phases. This framework overcomes the difficulties of training PINNs to solve a forward problem in two-phase flow, and it is expected to model more realistic dynamic systems in nature.