Applicability Of Experimentally Derived Local Droplet Distribution To The Fully Lagrangian Approach

It is computationally challanging to simulate sprays by fully resolving the transport equations at the droplet scale [1]. Instead, recent work has identified that the fully Lagrangian approach (FLA) has the potential to reduce computational time in spray simulations [2,3], while capturing complex processes in the flow such as droplet accumulation and droplet trajectory crossing [4]. In the FLA, the dispersed droplet phase is treated as a continuum [5]. All droplet parameters, including concentration, are
calculated along the chosen trajectories by solving systems of ordinary differential equations. This approach to modelling sprays has the advantages of Lagrangian particle tracking and continuous formulations for the particle-gas flow. The method has been applied to the case of polydisperse evaporating sprays by extending the original FLA to account for droplet sizes; this method is known as the generalised FLA(gFLA). For the closure of the gFLA, local droplet size and velocity distributions at each location are required as the initial conditions. This information is used to obtain the new size distributions as the flow evolves. In previous work [2,3], due to lack of experimental data, the
initial distributions for droplet size were assumed to be log-normal and the same at all locations. Furthermore, it was assumed that the initial velocity is the same for all the droplets.
The present work focuses on the development of a new experimental approach to characterising sprays, which enables us to obtain local droplet size and velocity distributions at each location for all instances of time. A flat fan spray nozzle was chosen to continuously inject water under steady flow conditions
at room temperature.Theback-lit images of the spray were captured by a high speed camera at 84000 frames per second. A microscopic lens was attached to the camera which resulted in a field of view of 2.23 × 5.95mm^2 and a resolution of 4.5 μm/pixel.The entire spray plume was captured by traversing the nozzle relative to the camera. In this way, a grid of multiple image batches were acquired. The structures in the atomised region were identified and tracked using an in-house image processing code. The statistical analysis of the data was used to obtain local droplet size and velocity distributions, which is required for the gFLA.