Dr Jun Xia
Reader

Summary

Dr Jun Xia obtained his BEng and MSc at Zhejiang University, China. He studied at the University of Southampton for his PhD on direct and dynamic large-eddy simulation studies of flame suppression by water mists and sprays, followed by postdoctoral research on diluted combustion at the same institution. In addition to better understanding the interactions between inert dispersing evaporating droplets and a diffusion flame in these engineering applications using high-fidelity simulations, fundamental differences in modelling framework between these burning systems and fuel-spray combustion were discussed. He then joined the Centre for Advanced Powertrain and Fuels of Brunel University as an academic.

Dr Xia is a certified software engineer and interested to better understand multi-physics engineering flow dynamics and transport phenomena in energy storage and systems, usually multiphase, resorting to high-fidelity simulation supported by high-performance computing and physics-guided machine learning, which helps to develop physics-based subgrid models.

One of his main research interests is fuel droplet(s) and spray dynamics, including flow and combustion. Interface-capturing numerical techniques, which combine sharp-interface-retaining level-sets and mass-conserving volume-of-fluid, have been further developed to better understand the puffing and microexplosion dynamics of an emulsion droplet and a droplet group, and their effects on fuel/air mixing and burning under convective heating. Recently, the capability of the code has been extended to cope multicomponent droplets, by incorporating more realistic evaporation of multicomponent liquid which takes into account liquid-component activities (therefore non-ideal liquid which is usually important for liquid mixtures with components different in chemical structure and molecular size), making it ready to develop models for complex spray processes embedding disruptive secondary breakup and atomisation such as microexplosion. Other major efforts include attempts on developing an integrated simulation tool for sprays in dense, transitional, and dilute spray regimes, to diminish the impact of uncertainties of upstream boundary conditions on spray modelling, which is important for predicting spray combustion dynamics and emissions, especially minor species on ppm levels. Graphics processing units were considered to speed up computing in spray solvers. Lattice Boltzmann was further developed for inside-injector, cavitating flows at low Reynolds number but realistic gas/liquid density ratio, interacting with an idealised moving needle valve.

In addition to gas-liquid two-phase flows, Dr Xia’s research has also been on gas-solid two-phase reacting flows. In collaboration with leading overseas groups, we further developed high-fidelity simulation techniques to investigate solid-fuel, i.e., coal and/or biomass, burning and especially alkali-metal minor-species emissions, incorporating radiation and pyrolysis models that are both important in the context. Chemistry tabulation has been developed to predict, with turbulence offline, alkali-metal emissions from a turbulent pulverised coal flame, which was quantitatively characterised by turbulence-resolving simulation. An important knowledge gap is alkali-species emissions of particles during the burning. We therefore have further developed lattice Boltzmann methods to simulate a burning porous char particle, aiming to better understand the emission from a subgrid point-source fuel particle in macroscopic high-fidelity simulation of turbulent combustion of pulverised solid-fuels and their mixtures.

Under the support of the EPSRC, microscopic molecular dynamics simulation has been used to investigate underground CO2 storage in an exploited or depleted oil reservoir, specifically the properties of a three-phase dodecane droplet and the impacts of CO2 and H2O on the droplet or film in the context of oil recovery. We also quantified transport and thermodynamic properties of CO2/H2 mixtures in a variety of compositions under typical under-surface thermodynamic conditions, with H2 as impurity in deposited CO2 at one end and with CO2 as cushion gas in H2 storage at the other, in porous aquifers or depleted ones. Clearly identified by molecular dynamics with anisotropic diffusion of supercritical species under these conditions, a recurrent neural network was also developed to predict the transition between anomalous and normal self-diffusion. With these knowledge gaps filled, we are getting ready for macroscopic modelling of geological flows under the impact of CO2 and/or H2 to guide underground CO2/H2 storage.