Research interests

 

I have been working on propulsion related research since pursuing PhD in Princeton University. In addition to conventional aerospace sciences specifically on fluid dynamics, combustion, and high-speed propulsion, recently we have found some potential routes of using novel electromagnetic propulsion based on the geomagnetic field to generate thrust with much less demands of propellant and energy, providing an alternative way to near earth and deep space transportation in near future. The related results have been published in a prestigious journal and more are being studied and prepared for submission. Propelled by strong desire of devising innovative and effective ways for long-term space travels, I am keen to devote main efforts to studies and services of relevance.

 

(Last updated: 2005; to be updated)

I have been working on the dynamics of droplet collision and flame propagation. My current interests include computational techniques and nonlinear fluid dynamics, specifically about flame instabilities, pattern formations, universalities in complex systems, and multiple-scale phenomena. I have been involved in the development of computational physics, such as immersed-boundary method for flame motion, front-tracking method and multi-scale simulation for droplet collision, and lattice Boltzmann method for multiphase flows and flame dynamics. The established computational and theoretical methodologies shall be interacted with experimental efforts to investigate the wonder of fluid physics, which will be further applied to such fresh while fervid fields as micro/nano fluidics (e.g. micro channels in fuel cells) and bio-mechanics (blood flows). In addition, considering the importance of clean energy technologies, my direction in near future may include hydrogen combustion, which is of close relevance to my background. Furthermore, new subjects specifically related to comprehension and interaction of physics and engineering sciences, such as search for new energy and space exploration, shall be of my future interests. Among these, some have been written in project proposals for funding support, either in progress or submission, and some are in collaboration either internationally or domestically. The details are described as follows.

 

Dynamics of droplet collision 液滴碰撞之動力學

The subject is concerned with the interactions between binary droplets and of droplets with the liquid/solid surfaces. The collision outcome can be as rich as ranging from merging, bouncing, merging and separation followed by satellite droplets, and splattering. This topic plays a key role in various technologies and natural phenomena, such as formation of raindrops and their interaction with the earth surface of ocean or soil, meteorite impaction, nuclear reaction, ink-jet printing technology, and spray combustion. I was experimentally observing and quantifying the possible outcomes of collision, and computationally simulating the global flow field on the basis of continuum fluid mechanics and the motion within the inter-surface gap which needs further connection with the micro mechanics relevant to rarefied gas dynamics and intermolecular forces, in the hope of well predicting the transition criteria. The specific merit of the computational effort is on the implementation and modification of a numerical algorithm that tracks moving boundaries in multiphase flows.

 

Flame dynamics火焰傳播動力學

● Front tracking for flame propagation

For this topic, the dynamics of flame fronts was computationally studied. The inherent complexity associated with the nonlinearity and multiple time scales characterizing fluid flows and chemical kinetics imposes severe difficulty in advances in such crucial fields as turbulent combustion. However, by treating the flame as a propagating front with infinitesimal thickness while still coupled to the external parameters accounting for the burning conditions, the complication and computational effort can be substantially reduced. Encouraged by the advancement of the computational scheme for somewhat prohibitive combustion problems, I can thus apply this method to study the underlying physics of the interaction between flames and vortical flows, and the intricate evolutions of flame-front instabilities.

 

● Nonlinear flame dynamics and pattern formation

In combustion, due to density variation across the flame front, there exists an inherent hydrodynamic instability known as Darrieus-Landau (D-L) instability. It leads to nonlinear evolution of flame fronts shown as unsteady motions of cells or stable movement of large-scale wrinkles, depending on the configuration and parametric conditions. Specifically, the former is associated with chaotic formation of merging and dividing of elemental cells, which is analogous to self-replicating patterns in reaction-diffusion systems and biological evolutions. Another flame-front instability is the Rayleigh-Taylor (R-T) instability, which is caused by buoyancy-induced driving force in stratified density fields. Different from D-L instability, it does not generate cellular structure of specific dimension and thereby the resulted nonlinear development of flame surface is more disordered. I am interested in the transition between the two instabilities as gravity is increased. Though both regimes have been studied respectively by different groups of researchers (i.e. D-L instability by combustion scientists and R-T instability by astrophysicists), the intermediate regime that spans the range and covers both effects is rarely specified. An immediate interest is the evolution of thermonuclear flame that is initiated in supernovae of Type Ia. It has been argued that the creation of turbulence triggered by R-T instability leads to acceleration of flame speed and subsequent transition from deflagration to detonation, whereby explosions of supernovae can be accomplished, according to the scenario of delayed detonation theory. On the other hand, it was reported that the fractal dimension in spherically expanding flames without gravitational effect is close to that of turbulence and thus D-L instability could be a key mechanism for self generation of turbulence. As a consequence, after a thermonuclear flame is ignited, it may have been corrugated and accelerated by D-L instability, through which the flow field can be turbulized, before R-T instability becomes dominant.

 

Computational fluid dynamics 計算流體力學

● Front tracking and multi-scale simulation for droplet collision

To further investigate the transition behaviors and criteria for various regimes in droplet collision mentioned earlier, it shall be helpful by simulating the hierarchical dynamics of the narrow gas gap between the impinging interfaces of droplets or liquid films, which is critical for the occurrence of merging. It was found that whether the droplets merge or bounce away depends on the balance between the impact inertia and the resisting pressure inside the gap. As such, during the stage when the gap is compressed, different computational schemes may need to be employed or coupled with theoretical modeling, since the gap scale spans a broad range when the fluid dynamics is dominated by continuum mechanics, rarefied gas, and intermolecular forces between the liquids. As simulated by the front tracking method that is based on continuum Navier-Stokes dynamics with incompressibility, the colliding droplets are stopped at certain distance and bounce away after some time. I suggested that the missing physics related to the rarefied gas and compressibility could bring the interfaces further toward each other to a sufficiently small distance whereby the van der Waals force dominates and merging would occur. To simulate the process and obtain the criteria for transitions between various regimes, I shall try the connections between models for different scales such as continuum, kinetic, or molecular schemes.

 

● Lattice Boltzmann simulation for multiphase flows and flame dynamics

Lattice Boltzmann method (LBM), based on a mesoscopic treatment for kinetic dynamics, is becoming a powerful tool to simulate fluid flows, specifically for its handling with complex geometry and multiphase flows as well as high parallelization. For multiphase flows, I have qualitatively simulated the collision events between droplets based on previous theories that employ pseudo inter-particle forces. However, for a quantitative description of the various regimes such as bouncing and coalescence, well connection between the physical and parametrical conditions of the scheme needs to be implemented. For example, the intermolecular attraction due to van der Waals force shall be correlated with the coefficients of inter-particle forces in LBM; this may necessitate both theories and experiments. Furthermore, issues such as evaluation of surface tension in multiphase flows with multiple components and rarefied fluid effect need to be addressed. LBM can also be developed to simulate flame dynamics as another choice to the full-field numerical simulation and front-tracking method. For simplicity, a reduced one-step reaction could be adopted to form a flame front, which separates burned and unburned gases with distinct flow properties as that performed for the flame-sheet approach. Once it can be established for low dimensional cases, three dimensional configurations shall be able to be extended straightforwardly. As a benefit, the suitability of LBM for parallel computing would facilitate the simulation of large complex systems.

 

● Extensive applications

In view of the merits, the front-tracking and lattice Boltzmann methods are promising for such hot research fields as micro/nano fluidics and bio-mechanics. Particularly, the essence of the former technique, i.e. the immersed-boundary method which smears the singular sources carried by the interfaces and enables the flow field with distinct fluid properties to be solved in a unified domain, has been applied extensively in biological systems to simulate the evolution of fluid flows driven by moving interfaces such as the flow in the heart and blood vessels. In lieu of the detailed tracking for the interfaces, the latter, however, provides a prospect to overcome the substantial increase of computational efforts when myriads of elastic objects such as blood cells are encountered. These sophisticated interactions between flow and interfaces would be practically investigated as long as appropriate interfacial conditions such as surface tension are implemented. Thereby the computational methodologies promise high potential and flexibility for the mesoscopic simulations and further extensions.