Numerical computation of fluid properties at nano/meso scales.

Engineering systems are increasingly being developed with dimensions within the micro to nano scale. Mature simulation schemes are available for large scale systems (> 0.5mum) in the form of continuum mechanics, and for small scale systems (< 50nn). However, there is to simulation scheme that covers the middle, meso scale, range between them. The work presented in this thesis focuses on the development of a computational framework focused on fluid systems on the nano- meso scale, with characteristic dimensions between 50nm and 500nm. Existing methods approach the meso scale either with approximated molecular behaviour from the 'top down', or directly modelling molecular physics from the 'bottom up'. Top down approaches have the disadvantage of only including known behaviour with some statistical variations to approximate chaotic behavior. Bottom up approaches model the fluid from a molecular physics model, but fail to capture bulk fluid behaviour and are computationally expensive. The approach developed in this thesis, covers the middle ground between continuum and molecular simulation scales. A molecular physics model is used to govern the behaviour of the fluid, and is surrounded by a set of meso scale boundary conditions, providing an accurate and efficient fluid model. Bulk fluid behaviour is extracted in the form of ensemble property distributions in a versatile grid-like implementation, allowing the fluid properties to be calculated from first principles accurately and efficiently. Each part of the developed method is validated separately. The physics model is compared with published results of simulations at molecular scales, as there is insufficient information for meso scale fluid systems. The bulk ensemble property collection scheme is fully explored by means of a parametric study. Case studies are presented to highlight how bulk fluid properties, such as velocity, temperature and pressure, can be examined as distributions in time and space over the flow field in channel flow systems. The approach developed in this thesis opens the door to accurate and efficient meso scale fluid simulation. This work has also identified the next step to widen and improve the abilities for meso scale fluids to be fully investigated