Investigation on spatial and temporal characteristics of scalar dissipation rate in non-reacting and reacting turbulent jets
This thesis analyses spatial and temporal characteristics of scalar dissipation rate (SDR) in statistically steady and decelerating axisymmetric jets, using Direct Numerical Simulations (DNS). The SDR represents the mixing rate of jet fluid with the ambient. In turbulent chemically reacting flows, SDR is an essential modeling parameter which controls ame ignition and extinction. The simulated DNS dataset is thoroughly validated for transport and turbulent mixing characteristics of the steady-state jet. Validation includes evaluating budget terms for turbulent kinetic energy and mixture fraction variance transport equations. The equations are balanced and agree with experimental measurements. This provides confi dence in the data and allows for further investigation on turbulent mixing quantities such as SDR. Near-field mixing up to 10 diameters from the nozzle is described by temporally and spatially-coherent layers of high SDR values. The layers are aligned at angles of 45 to 75 with respect to the axial direction. Alignment angles are consistent with other experimental observations. Individual components of the SDR are also inspected, regarding coherence. Radial component of SDR is shown to have the same angular alignment characteristics. Axial component of SDR is not aligned to any direction, but rather shows irregular structures. The azimuthal component of SDR shows the same angular alignment characteristics as the radial component, but also shows an azimuthal alignment which is not present in the other components. Further downstream in the jet, spatial characteristics of SDR become self-similar. Self-similarity appears when the SDR is normalised by its centreline value and the radial coordinate is scaled by its axial location. In the self-similar region, centreline SDR scales with the downstream distance, x⁻4. The normalised pro les of the SDR show a monotonic increase near the jet axis, followed by a steep decrease, after an off-centreline peak. Self-similar characteristics also appear for radial, axial and azimuthal components of SDR. Axial components show a plateau region near the centreline, followed by a steep decrease with increasing radii. Pro files of the radial component are similar to the total SDR, and pro files of the azimuthal component are similar to their axial counterparts. A budget analysis of the mean SDR transport equation reveals that two terms are dominant: the fi rst is production due to local curvature effects on the scalar fi eld, and the second is destruction by scalar fi eld stretch due to velocity gradients. The two terms are balanced at the leading order. Next, the transport equation for SDR due to scalar fluctuations was analysed. Two dominant budget terms are linked to (i) production by curvature effects on scalar fi eld fluctuations and (ii) destruction by turbulence-scalar interaction. For both mean and fluctuation equations, convection, diffusion and turbulent transport terms are at least one order of magnitude smaller than dominant terms. The SDR characteristics of the decelerating jet were analysed. As the jet is stopped, the high dissipation coherent structures of SDR in ltrate towards the so-called potential core. Note that the SDR is null inside the potential core in the steady-state jet. After a transient time, centreline pro les of ensemble-averaged SDR have the following characteristics. A decelerating wave propagates downstream at approximatively half of centreline axial velocity. Upstream of the decelerating wave, the pro le is proportional to the axial distance and inversely proportional to time squared. Downstream of the decelerating wave, the pro le remains same as for steady-state. Furthermore, increased entrainment of ambient fluid during deceleration leads to an overall increase in radial profi les for SDR and its directional components. In the mean SDR transport equation, the production term and the destruction term, which were dominant in the steady-state, are no longer the only significant cant terms. Two additional terms have an influence at leading order: the temporal term of SDR and the turbulent transport term. A new DNS solver for reacting flows is developed. This is done by coupling the High Performance Solver for Turbulence and Aeroacoustic Research (HiPSTAR) with the open-source CANTERA chemical kinetics suite. Development is motivated by HiPSTARs ability to solve the ow governing equations in cylindrical coordinates. This reduces computational cost up to 30% for ow problems which are naturally tailored for a cylindrical grid. The new solver supports detailed chemical mechanisms. Zero-dimensional and one-dimensional validation cases are run with HiPSTAR-CANTERA.