Laser optical methods
Combustion processes of defined droplet chains in a natural gas envelope flame or spray flames are investigated with the help of modern laser optical methods. The results of these experiments can be compared with computer simulations that model the vaporisation, mixing and combustion process.
For practically relevant combustion arrangements, a turbulent, pre-mixed free jet flame with natural gas as fuel was implemented. Liquids can be fed in via a droplet chain generator. In addition, sprays can be fed in via nozzles.
The majority species and the temperature in the gas phase can be measured using Raman spectroscopy. The methodology is very complex, but meets the requirements and offers the advantages of non-invasive in-situ measurement. Any intervention in the flame using mechanical components, would lead inevitably to a modification of the chemical and physical processes and therefore to a considerable error. For Raman spectroscopy, two Nd:YAG lasers emitting at 532 nm with pulse energies of approx. 600 mJ and a pulse length of approx. 6 ns are used on the excitation side. One of the techniques used in order to avoid plasma formation in the measured volume, is pulse widening from approx. 6 ns to approx. 30 ns via two optical delay paths. An achromatic lens, a spectrometer and an ICCD camera are used on the detection side. The lens and the spectrometer are adjustable in all spatial directions, in order to optimise imaging properties and collection efficiency.
Modelling and simulation of droplet and spray combustion
Multiphase processes too, can be simulated in detail in reliance on the general conservation equations. Due to the massive computing time, once again drastic simplifications are required. Thus for example, the combustion of individual droplets or droplet groups can be simulated in detail by restricting the situation to simple geometries. This can include, for example, simulations of single droplets in which conservation equations for the processes in the gas phase and in the droplet are solved, which are linked to the phase boundary through balance equations. In particular, it is possible to determine the ignition delays, which are important for practical applications and which can sometimes deviate greatly from those in a pure gas phase reaction. This is shown in the illustration above for the fuels methanol, n-heptane and iso-octane as a function of the temperature of the air surrounding the droplets. The simulations are based on a detailed gas phase model (detailed reaction mechanisms, detailed transport models), and on a detailed description of the processes within the droplet and at the phase boundary. It is a characteristic fact that the ignition delays are longer than those under the same conditions in a homogeneous stoichiometric gas mixture.
The use of detailed simulations in the description of technical combustion systems (e.g. diesel engine combustion) is associated with an unmanageable computing time. The complexity of the mathematical modelling depends crucially on the properties of the fuel spray. A statistical formulation is obtained by considering the probability density function for the spray, i.e. the disperse phase. In analogy with the PDF transport equation for the gas phase, the underlying spray equation can be solved by Monte Carlo methods. This is shown in the illustration below. It is, however, difficult to determine the exchange terms between the disperse liquid phase and the gas phase, which are effected by the evaporation velocity, the fine structure of the reaction zone around the flame and possibly even by processes in the droplet (convection, heat conduction). Here too, detailed simulations permit an accurate insight into the coupling of flow, transport and reaction processes. For example, simulating the ignition and combustion of individual droplets but also of droplet groups is possible by using detailed models for the chemical reaction and the molecular transport processes (see above). The results of such simulations can then be used as libraries of combustion scenarios for describing the spray combustion.
Stochastic distribution of the stochastic particles in the modelling of a turbulent spray flame. Particle size represents droplet size (not to scale).