In the quest to mitigate demand for conventional petroleum-derived transportation fuels and reduce their associated emissions of various pollutants, there are an increasing number of alternative fuels are being proposed. Employing such alternatives necessitates a comprehensive understanding and accurate measurement of their combustion characteristics for effective commercial deployment. Particle Image Velocimetry (PIV) is amongst the advanced experimental techniques now in use to improve our understanding of combustion. It was therefore installed and employed in the course of the present work. Such a technique can directly measure key combustion characteristics with high accuracy, under both laminar and turbulent conditions.
This PIV technique was employed first for measuring laminar burning velocities during flame propagation in spherical explosions, by the measurement of flame speed and gas velocity just ahead of the flame. Measurements made in this way are compared with those obtained solely from the flame speed method, which is based on the flame front propagation speed and the ratio of unburned to burned gas densities. Different values arose between the two methods. The principal reason was the common assumption in the flame speed method that the burned gas density is at the equilibrium, burned gas, adiabatic temperature value. When allowance is made for the effects of flame stretch rate and Lewis number on this density, the differences in burning velocities are significantly reduced. Burning velocities and Markstein numbers have been measured for methane, i-octane, ethanol, and n-butanol over a range of equivalence ratios at atmospheric pressure and, in the case of n-butanol, also over a range of pressures. In measuring Markstein numbers, there is a dependency upon the isotherm employed for the measurement of the stretch rate. This aspect was studied by comparing measurements with two different isotherms. It was concluded that the measured PIV flame measurements might under-estimate the Markstein numbers by about 12%.
The PIV technique was employed also to measure the turbulence characteristics of the flow in fan-stirred vessel, using dry air in the absence of phase change and chemical reaction. Since a knowledge of the aerodynamic characteristics of the turbulent flow enables better analysis of the flame/turbulence interactions. Spatial and temporal distributions of mean and root mean square, rms, velocity fluctuations are investigated, as well as integral length scales, L, Taylor microscales, λ, and Kolmogorov length scales, η, in the fan speed range, 1,000-6,000 rpm. Mean velocities are about 10 % of the turbulence velocity, u'. Importantly, turbulence is close to homogeneous and isotropy in the central volume, although this volume decreases with increasing fan speed. Its radius and other characteristics are expressed in terms of the fan speed. Relationships are presented for the variations of u' and L with fan speed, temperature and pressure. A novel relationship between the autocorrelation function and integral length scale is obtained, for when Taylor’s hypothesis is invalid.
Finally, changes induced in the turbulent flow fields by methane/air flames were measured at different experimental conditions. In measuring turbulent burning velocity in spherical explosions, allowance must be made for the transient changes in the rms turbulent velocity, to which the flame is exposed. This rms turbulent velocity was measured a head of flame front. The influences of pressure, temperature and equivalence ratio on its value were investigated and a novel empirical expression obtained.
