E. Fan
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
E. Fan has a total of 17 co-authors. Here are the top co-authors by number of shared publications:
- Chih-Yung Wen
The Hong Kong Polytechnic University
12 shared publications - Lisong Shi
The Hong Kong Polytechnic University
7 shared publications - Jiaao Hao
The Hong Kong Polytechnic University
5 shared publications - Xin Li
Department of Aeronautical and Aviation Engineering, The Hong Kong Polytechnic University 1 , Hung Hom, Kowloon, Hong Kong
3 shared publications - Ben Guan
College of Aerospace and Civil Engineering, Harbin Engineering University, Harbin 150001, China
2 shared publications Ben Guan
The Hong Kong Polytechnic University, Department of Mechanical Engineering 1 , Hung Hom, Kowloon, Hong Kong, People’s Republic of China
1 shared publication
Latest Publications
- 21 November 2026
- Published by Elsevier in Combustion and Flame
- Vol. 296
Transmission of a detonation wave across a hot inert gas layer
- 21 November 2026
- Published by Elsevier in Combustion and Flame
- Vol. 296
Transmission of a detonation wave across a hot inert gas layer
- Open Access
- 2 March 2026
- Published by Springer Nature in Shock Waves
- Vol. 36 (1)
Two-dimensional numerical simulation of imploding detonations
This study aims to contribute to the understanding of imploding detonations from a numerical perspective, focusing primarily on the detailed transient and wave structures during implosion that are challenging to capture experimentally. An inviscid perfect gas model with a single-step Arrhenius reaction is employed. Imploding detonations are initiated by collisions of multiple small hot spots, and both two-dimensional circular and polygonal implosions are examined, with attention to the effects of obstacles and varying ignition pressures. For circular implosions, a slight acceleration of detonation velocity is observed at ignition, with significant acceleration occurring only in the final stages. Near the implosion center, local wave velocities exceed twice the Chapman–Jouguet velocity, yet the wave front maintains cellular instabilities until the collapse is complete. Additionally, the merging of transverse waves is observed during the implosion. In polygonal cases, a small number of “ignition edges” allows regular reflection, preserving the initial wave front geometry, while increasing the number of ignition lines leads to Mach reflection and the formation of a circular wave front. When obstacles are introduced, the detonation reflects off the obstacle, causing localized delays in the wave front that cannot be fully compensated, as transverse waves are unable to propagate sufficiently in the circumferential direction to smooth out disturbances. Similar effects are noted for implosions with non-uniform ignition pressures. It should be noted that the findings are based on a simplified model and do not account for real gas effects or additional physical processes present in actual detonation implosions.
- 2 March 2026
- Published by Springer Nature in Shock Waves
- Vol. 36 (1)
Open AccessTwo-dimensional numerical simulation of imploding detonations
This study aims to contribute to the understanding of imploding detonations from a numerical perspective, focusing primarily on the detailed transient and wave structures during implosion that are challenging to capture experimentally. An inviscid perfect gas model with a single-step Arrhenius reaction is employed. Imploding detonations are initiated by collisions of multiple small hot spots, and both two-dimensional circular and polygonal implosions are examined, with attention to the effects of obstacles and varying ignition pressures. For circular implosions, a slight acceleration of detonation velocity is observed at ignition, with significant acceleration occurring only in the final stages. Near the implosion center, local wave velocities exceed twice the Chapman–Jouguet velocity, yet the wave front maintains cellular instabilities until the collapse is complete. Additionally, the merging of transverse waves is observed during the implosion. In polygonal cases, a small number of “ignition edges” allows regular reflection, preserving the initial wave front geometry, while increasing the number of ignition lines leads to Mach reflection and the formation of a circular wave front. When obstacles are introduced, the detonation reflects off the obstacle, causing localized delays in the wave front that cannot be fully compensated, as transverse waves are unable to propagate sufficiently in the circumferential direction to smooth out disturbances. Similar effects are noted for implosions with non-uniform ignition pressures. It should be noted that the findings are based on a simplified model and do not account for real gas effects or additional physical processes present in actual detonation implosions.
- Open Access
PhyNiKCE: A Neurosymbolic Agentic Framework for Autonomous Computational Fluid Dynamics
The deployment of autonomous agents for Computational Fluid Dynamics (CFD), is critically limited by the probabilistic nature of Large Language Models (LLMs), which struggle to enforce the strict conservation laws and numerical stability required for physics-based simulations. Reliance on purely semantic Retrieval Augmented Generation (RAG) often leads to "context poisoning," where agents generate linguistically plausible but physically invalid configurations due to a fundamental Semantic-Physical Disconnect. To bridge this gap, this work introduces PhyNiKCE (Physical and Numerical Knowledgeable Context Engineering), a neurosymbolic agentic framework for trustworthy engineering. Unlike standard black-box agents, PhyNiKCE decouples neural planning from symbolic validation. It employs a Symbolic Knowledge Engine that treats simulation setup as a Constraint Satisfaction Problem, rigidly enforcing physical constraints via a Deterministic RAG Engine with specialized retrieval strategies for solvers, turbulence models, and boundary conditions. Validated through rigorous OpenFOAM experiments on practical, non-tutorial CFD tasks using Gemini-2.5-Pro/Flash, PhyNiKCE demonstrates a 96% relative improvement over state-of-the-art baselines. Furthermore, by replacing trial-and-error with knowledge-driven initialization, the framework reduced autonomous self-correction loops by 59% while simultaneously lowering LLM token consumption by 17%. These results demonstrate that decoupling neural generation from symbolic constraint enforcement significantly enhances robustness and efficiency. While validated on CFD, this architecture offers a scalable, auditable paradigm for Trustworthy Artificial Intelligence in broader industrial automation.
Open AccessPhyNiKCE: A Neurosymbolic Agentic Framework for Autonomous Computational Fluid Dynamics
The deployment of autonomous agents for Computational Fluid Dynamics (CFD), is critically limited by the probabilistic nature of Large Language Models (LLMs), which struggle to enforce the strict conservation laws and numerical stability required for physics-based simulations. Reliance on purely semantic Retrieval Augmented Generation (RAG) often leads to "context poisoning," where agents generate linguistically plausible but physically invalid configurations due to a fundamental Semantic-Physical Disconnect. To bridge this gap, this work introduces PhyNiKCE (Physical and Numerical Knowledgeable Context Engineering), a neurosymbolic agentic framework for trustworthy engineering. Unlike standard black-box agents, PhyNiKCE decouples neural planning from symbolic validation. It employs a Symbolic Knowledge Engine that treats simulation setup as a Constraint Satisfaction Problem, rigidly enforcing physical constraints via a Deterministic RAG Engine with specialized retrieval strategies for solvers, turbulence models, and boundary conditions. Validated through rigorous OpenFOAM experiments on practical, non-tutorial CFD tasks using Gemini-2.5-Pro/Flash, PhyNiKCE demonstrates a 96% relative improvement over state-of-the-art baselines. Furthermore, by replacing trial-and-error with knowledge-driven initialization, the framework reduced autonomous self-correction loops by 59% while simultaneously lowering LLM token consumption by 17%. These results demonstrate that decoupling neural generation from symbolic constraint enforcement significantly enhances robustness and efficiency. While validated on CFD, this architecture offers a scalable, auditable paradigm for Trustworthy Artificial Intelligence in broader industrial automation.
- 1 October 2025
- Published by Elsevier in Computer Physics Communications
- Vol. 318
Fire: An open-source adaptive mesh refinement solver for supersonic reacting flows
- 1 October 2025
- Published by Elsevier in Computer Physics Communications
- Vol. 318
Fire: An open-source adaptive mesh refinement solver for supersonic reacting flows
- 1 August 2025
- Published by AIP Publishing in Physics of Fluids
- Vol. 37 (8)
Simulations of detonation wave reflection over cylindrical convex surfaces with a detailed reaction model
The unsteady reflection of detonation waves over cylindrical convex surfaces is numerically investigated using a high-resolution adaptive mesh refinement technique combined with a detailed reaction mechanism. Results reveal that cellular instabilities complicate the determination of the transition point from regular reflection to Mach reflection. However, the triple-point trajectories connecting the incident detonation wave and Mach stem remain largely unaffected by these instabilities. This consistency is confirmed by reducing the radius and the initial angle of the cylindrical surfaces. Finally, no universal scaling is observed across all configurations. At larger radii, the triple-point trajectories tend to stay closer to the surface.
- 1 August 2025
- Published by AIP Publishing in Physics of Fluids
- Vol. 37 (8)
Simulations of detonation wave reflection over cylindrical convex surfaces with a detailed reaction model
The unsteady reflection of detonation waves over cylindrical convex surfaces is numerically investigated using a high-resolution adaptive mesh refinement technique combined with a detailed reaction mechanism. Results reveal that cellular instabilities complicate the determination of the transition point from regular reflection to Mach reflection. However, the triple-point trajectories connecting the incident detonation wave and Mach stem remain largely unaffected by these instabilities. This consistency is confirmed by reducing the radius and the initial angle of the cylindrical surfaces. Finally, no universal scaling is observed across all configurations. At larger radii, the triple-point trajectories tend to stay closer to the surface.
- 31 July 2025
- Published by Elsevier in Combustion and Flame
- Vol. 280
Detonation propagation in continuously curved three-dimensional ducts
- 31 July 2025
- Published by Elsevier in Combustion and Flame
- Vol. 280
Detonation propagation in continuously curved three-dimensional ducts
- Open Access
- 15 February 2025
- Published by Elsevier in International Journal of Hydrogen Energy
- Vol. 109, 1064-1080
Evolution of shock-induced reactive and inert double-layer gas cylinders
- 15 February 2025
- Published by Elsevier in International Journal of Hydrogen Energy
- Vol. 109, 1064-1080
Open AccessEvolution of shock-induced reactive and inert double-layer gas cylinders
- 1 August 2024
- Published by AIP Publishing in Physics of Fluids
- Vol. 36 (8)
Role of Atwood number in the shock-induced evolution of a double-layer gas cylinder
An A/B/C-type gas cylinder with various concentrations of SF6 (ranging from 5% to 80% in volume fraction) in the inner cylinder is constructed to investigate the dependence of the interface evolution on the Atwood number. For negative Atwood numbers, secondary vortex pairs emerge at the downstream interface of the outer cylinder following the interaction of a high-pressure triple point with the downstream interface, while a downstream jet is formed due to the generation of a notably higher-pressure zone after the transmitted shock wave traverses the convergence point. The widths and heights of both outer and inner cylinders are analyzed to quantify the interface evolution. The mechanism behind the vorticity evolution is investigated using the vorticity transport equation. The vorticity equation is introduced to investigate the mechanism of vorticity evolution. The dilatation and baroclinic terms play a dominant role in the dynamics of vorticity production. The net circulation can be predicted by linearly summing existing circulation models. Analysis of the area and mean mass fraction histories of the outer and inner cylinders shows that more ambient gas dilutes SF6 and promotes gas mixing as the Atwood number decreases.
- 1 August 2024
- Published by AIP Publishing in Physics of Fluids
- Vol. 36 (8)
Role of Atwood number in the shock-induced evolution of a double-layer gas cylinder
An A/B/C-type gas cylinder with various concentrations of SF6 (ranging from 5% to 80% in volume fraction) in the inner cylinder is constructed to investigate the dependence of the interface evolution on the Atwood number. For negative Atwood numbers, secondary vortex pairs emerge at the downstream interface of the outer cylinder following the interaction of a high-pressure triple point with the downstream interface, while a downstream jet is formed due to the generation of a notably higher-pressure zone after the transmitted shock wave traverses the convergence point. The widths and heights of both outer and inner cylinders are analyzed to quantify the interface evolution. The mechanism behind the vorticity evolution is investigated using the vorticity transport equation. The vorticity equation is introduced to investigate the mechanism of vorticity evolution. The dilatation and baroclinic terms play a dominant role in the dynamics of vorticity production. The net circulation can be predicted by linearly summing existing circulation models. Analysis of the area and mean mass fraction histories of the outer and inner cylinders shows that more ambient gas dilutes SF6 and promotes gas mixing as the Atwood number decreases.
- Open Access
- 1 February 2024
- Published by AIP Publishing in Physics of Fluids
- Vol. 36 (2)
Effects of Mach and Atwood numbers on the shock-induced evolution of a double-layer gas cylinder
The evolution of a double-layer gas cylinder under various Mach numbers (M = 1.27, 1.5, 1.7, and 2.1) and Atwood numbers (A = 0.504, 0.392, 0.194, and −0.202 for the first layer) are studied numerically. At M ≥ 1.5 and A = 0.504, a bulge is generated near the upstream pole of the outer cylinder due to the impingement of a reflected shock wave, which promotes the formation of an upstream jet. At a higher Mach number, the evolution of the jet is suppressed under the influence of a higher pressure upstream of the jet head. The compressibility effects are quantified by the widths and heights of the gas cylinders. The Atwood number is associated with nonlinear acoustic effects, and the sign of A results in a significant variation in the wave patterns. The development of vortex pairs slows down with the decreasing Atwood numbers in the scenario of A > 0 for the first layer gas cylinder, while vortex pairs emerge and propagate in both upstream and downstream directions from the outer interface in the scenario of A < 0 for the first layer. As the Mach numbers and the magnitude of the Atwood numbers are increased, the mixing of various gases is promoted by detecting the circulation and mixed mass. The net circulation can be predicted by the linear summation of the Picone and Boris model and the Samtaney and Zabusky model under various Mach and Atwood numbers.
- 1 February 2024
- Published by AIP Publishing in Physics of Fluids
- Vol. 36 (2)
Open AccessEffects of Mach and Atwood numbers on the shock-induced evolution of a double-layer gas cylinder
The evolution of a double-layer gas cylinder under various Mach numbers (M = 1.27, 1.5, 1.7, and 2.1) and Atwood numbers (A = 0.504, 0.392, 0.194, and −0.202 for the first layer) are studied numerically. At M ≥ 1.5 and A = 0.504, a bulge is generated near the upstream pole of the outer cylinder due to the impingement of a reflected shock wave, which promotes the formation of an upstream jet. At a higher Mach number, the evolution of the jet is suppressed under the influence of a higher pressure upstream of the jet head. The compressibility effects are quantified by the widths and heights of the gas cylinders. The Atwood number is associated with nonlinear acoustic effects, and the sign of A results in a significant variation in the wave patterns. The development of vortex pairs slows down with the decreasing Atwood numbers in the scenario of A > 0 for the first layer gas cylinder, while vortex pairs emerge and propagate in both upstream and downstream directions from the outer interface in the scenario of A < 0 for the first layer. As the Mach numbers and the magnitude of the Atwood numbers are increased, the mixing of various gases is promoted by detecting the circulation and mixed mass. The net circulation can be predicted by the linear summation of the Picone and Boris model and the Samtaney and Zabusky model under various Mach and Atwood numbers.
- 12 March 2022
- Published by Elsevier in Combustion and Flame
- Vol. 241
Numerical investigation on reacting shock-bubble interaction at a low Mach limit
- 12 March 2022
- Published by Elsevier in Combustion and Flame
- Vol. 241
Numerical investigation on reacting shock-bubble interaction at a low Mach limit