No abstract

The opposed jet configuration presents a canonical geometry suitable for the evaluation of calculation methods seeking to reproduce the impact of strain and re-distribution on turbulent transport in reacting and non-reacting flows. The geometry has the advantage of good optical access and, in principle, an absence of complex boundary conditions. Disadvantages include low frequency flow motion at high nozzle separations and comparatively low turbulence levels causing bulk strain to exceed the turbulent contribution at small nozzle separations. In the current work, fractal generated turbulence has been used to increase the turbulent strain and velocity measurements for isothermal flows are reported with an emphasis on the axis, stagnation plane and the distribution of mean and instantaneous strain rates. Energy spectra were also determined. The instrumentation comprised hotwire anemometry and particle image velocimetry with the flows to both nozzles seeded with 1 μm silicon oil droplets providing a relaxation time of 3 μs. It is shown that fractal grids increase the turbulent Reynolds number range from 48-125 to 109-220 for bulk velocities from 4 to 8 m/s as compared to conventional perforated plate turbulence generators. Low frequency motion of the order 10 Hz could not be completely eliminated and probability density functions were determined for the location of the stagnation plane. Results show that the fluctuation in the position of the stagnation plane is of the order of the integral length scale, which was determined to be 3.1±0.1 mm at the nozzle exits through the use of hot-wire anemometry. Flow statistics close to the fractal plate located upstream of the nozzle exit were also determined using a transparent glass nozzle.

The opposed jet configuration presents an attractive canonical geometry for the evaluation of burning properties of turbulent flames with past studies typically limited to low Reynolds numbers. Fractal grid generated turbulence was used to remove the low turbulence level limitations associated with conventional perforated plate generators with the turbulent Reynolds number range moved from 50-120 to 130-318. Optimal grid configurations were determined with particular emphasis on reducing the impact of the flow upstream of the turbulence generators in order to facilitate simpler boundary conditions for computational studies. The resulting flow structures were analysed using proper orthogonal decomposition and conditional proper orthogonal decomposition. Velocity and reaction progress variable statistics, including conditional velocities and scalar fluxes, are reported for fuel lean methane, ethylene and propane flames approaching extinction. The instrumentation comprised particle image velocimetry with the flows to both nozzles seeded with 1 µm silicon oil droplets or 3 µm Al2O3 particles. Probability density functions were determined for the instantaneous location of the stagnation point to eliminate the possibility of low frequency bulk motion distorting velocity statistics. Probability density functions of flame curvature were determined using a multi-step flame front detection algorithm with estimates of the turbulent burning velocity provided along with a discussion of alternative determination methods. The data sets show that fractal grids generate multi-scale broadband turbulence and present an opportunity for a systematic evaluation of calculation methods for premixed turbulent flames that undergo a transition from non-gradient to gradient turbulent transport while approaching extinction.

Twin premixed turbulent opposed jet flames were stabilized for lean mixtures of air with methane and propane in fractal grid generated turbulence. A density segregation method was applied alongside particle image velocimetry to obtain velocity and scalar statistics. It is shown that the current fractal grids increase the turbulence levels by around a factor of 2. Proper orthogonal decomposition (POD) was applied to show that the fractal grids produce slightly larger turbulent structures that decay at a slower rate as compared to conventional perforated plates. Conditional POD (CPOD) was also implemented using the density segregation technique and the results show that CPOD is essential to segregate the relative structures and turbulent kinetic energy distributions in each stream. The Kolmogorov length scales were also estimated providing values ∼0.1 and ∼0.5 mm in the reactants and products, respectively. Resolved profiles of flame surface density indicate that a thin flame assumption leading to bimodal statistics is not perfectly valid under the current conditions and it is expected that the data obtained will be of significant value to the development of computational methods that can provide information on the conditional structure of turbulence. It is concluded that the increase in the turbulent Reynolds number is without any negative impact on other parameters and that fractal grids provide a route towards removing the classical problem of a relatively low ratio of turbulent to bulk strain associated with the opposed jet configuration.

No abstract