Spray ignition is a complex combination of physical and chemical processes. Therefore there is no way to simulate all processes without significant simplifications. Dynamic and thermal effects of droplets in a two phase flow are respected through averaged properties of mass, momentum and energy. When modeling mixing of fuel vapor with the surrounding gas, the differences between diffusion rates of components as well as the finite penetration of the diffusion fluxes are not taken into account. Last seen in the fact that fuel vapor is uniformly mixed with the gas filling within the calculation cell, whose dimensions are several orders of magnitude larger than the droplet. As a consequence, the chemical reactions in the two-phase medium are represented by volume processes in the computational cell. Meanwhile, it is well known that the surrounding of an individual droplet or a group of droplets is characterized through significant inhomogeneity. Under such circumstances, self-ignition of fuel vapor cannot be considered as a process in a well stirred reactor of a size of the computational cell.
In this work a new approach for spray ignition simulation is presented. The simulation is split into two tasks. A CFD package is used to calculate spray injection and subsequent two-phase flow in the given conditions. From this two-phase flow the droplet track data and the gas phase parameters are extracted to serve as variable boundary conditions for the second task, the one dimensional single droplet ignition simulations, named “Spraylet”, which are fully transient and based on comprehensive chemistry. The obtained ignition delay times can be transformed into spatial distributions of ignition probability. While the CFD part is handling spray formation, turbulence, temperature, pressure and global vaporization, the spraylet calculations take care of droplet related physics and chemistry.
The spraylet model is based on the transient differential energy and mass conservation equations in the liquid and gas phases with variable physical properties. Also the concept of multi-component diffusion is applied. The effects of the presence of neighboring droplets in the flow, usually referred to as “spray” effects, is approximated by taking modifications of the conditions at the outer boundary of the computational domain into account. Thus the model considers the finite rates of diffusion and chemical reactions, as well as spray effects, and allows for spray ignition simulations to predict the most probable instants of ignition of the individually calculated droplet trajectories.
The validation of the spraylet-based simulation is performed by comparison with experimental data obtained in the hot-wind-tunnel. N-heptane as liquid fuel was injected into cross flow of air with a pressure of 5 bar and a temperature of 800 K.