Application of Multilateral Jet for Partial Pre-Mixing

Abstract

The feasibility of multilateral jet mixing to produce partially premixed flames is studied in this thesis. Partially premixed flames are ubiquitous in many practical combustion systems, whilst controlled partially premixed flames are more stable and resilient towards extinction. It is hypothesized that, better control of flames is possible through multilateral jet mixing, in particular when reactants’ composition is variable. The technique of using multilateral jet for mixing has not been tested thoroughly and the fundamental fluid mechanics associated with the resulting flows is still unclear. The overall aim of this study is to understand the fundamental flow characteristics of multilateral jet in a confined cross-flow and to explore the feasibility and limitations of this technique to better control the stability of turbulent jet flames. Hence, this thesis aims to explore: the flow structures and characteristics inside and outside the nozzle; the side-jets mixing modes and parameters; mixing efficacy; and the flame structures in the near-field of the nozzle exit. The study was conducted in two parts: isothermal water-based flow studies; and turbulent reacting flows studies. Experimental campaigns under isothermal conditions were used to investigate the flow structures and regimes that can be produced by varying the side-jets to primary flow momentum ratio (MR). The experiments were conducted with nozzles consisting of four side-jets (4SJ), equi-spaced and located one primary diameter upstream of the nozzle exit, with the nozzle placed in a closed loop water tunnel. This study employed Planar Laser Induced Fluorescence (PLIF) and Particle Image Velocimetry (PIV), two non-intrusive laser diagnostic techniques, to study the mixing and flow fields, respectively. The flow Reynolds number, based on the primary jet diameter, ranged from 1300 to 6500. Different mixing regimes were identified to correspond to different momentum ratios. These regimes are: streaming flow; impinging flow; and backflow regime. The PIV results show that the side-jets in the streaming flow regime does not alter the primary flow field significantly. The impinging side-jets forma stagnation point upstream, which diverts primary flow over the stagnation region. Increasing the momentum ratio further leads to the backflow regime, which shows flapping characteristics upstream. The effects of the side-jets momentum ratio on the near-field flow, downstream of the primary jet nozzle exit were also investigated. This study shows that with the increase in momentum ratio, the coherent large-scale vortices roll-ups in the near-field become less apparent and more random multi-scale vortices are observed. Furthermore, increasing the momentum ratio further increases the centreline turbulence intensity and velocity decay of the jet in near-field. The existence of the side-jets alters the velocity and secondary flow distribution (dye) profile at the nozzle exit. These profile modification and increase in centreline turbulence and velocity decay persist to approximately two primary diameters downstream before morphing to a Gaussian profile, consistent with that of a round jet. Similar experiments were also conducted using the same primary flow with three side-jets (3SJ) mounted one primary diameter upstream of the exit plane, and with momentum ratios varied. This study shows that for both the 3SJ and 4SJ configurations and at low flow momentum ratio, Counter-rotating Vortex Pairs (CVPs) appear. When the side-jets penetrate the primary flow centreline, axis-switching of the CVPs are observed. Commercially available Computational Fluid Dynamics (CFD) package ANSYS CFX was used to further interrogate the vortices after axis switching and determined that the vortices are rotational, which advect flow in the nozzle, both towards and away from the flow centreline. This study also shows that classical Jets in Cross-Flow (JICF) scaling methods are not suitable for scaling the trajectory of the side-jet in a confined flow. The trajectory of the individual JICF are affected by: existence of adjacent jets; confinement geometries; and restrictions posed by the primary flow centreline. The backflow length for both the 3SJ and 4SJ can be scaled to the MR and the number of side-jets. A constant for the scaling, k is identified as 0.18, for the dye mixture fraction scaling method and 0.16 for the velocity scaling method. It is also found that the 4SJ configurations generally show higher turbulence and vorticity than the 3SJ due to the increase in primary flow blockage ratio. Experiments were also conducted on partially premixed flames of natural gas and air. The primary nozzle was constructed of a stainless steel long-pipe with 25.4mm inner diameter. Both 3SJ and 4SJ nozzle configurations were examined under the influence of different flow MR. The PIV on the centreplane, downstream from the nozzle exit for 4SJ shows that the velocity profiles are symmetrical and similar for the planes in the side-jets’ axis and 45° offset. This similarity seem to happen despite the obvious differences in the mixing profile generated at the nozzle exit shown in earlier studies. Flame photography was conducted with a standard DSLR at different exposure time. The physical flame length for the lifted flames, with momentum-ratio matching the impinging flow regime is markedly shorter than that in the backflow regime. Transitioning from lifted flames to attached flames shows a reduction in OH* emission, usually associated with reduction in temperature, at the flame front. Furthermore, the study also shows that flow cases with higher momentum-ratio are more stable than the lower momentum-ratio case despite similar air-to-fuel ratio. The various studies conducted in this thesis have shown that multi-lateral jet mixing is a feasible, simple and effective technique to partial premix reactants. The momentum ratio is able to provide additional control to stabilize the turbulent flame independent of equivalence ratio. More work is needed to better optimize this technique. In particular, a further understanding of the reactive scalars distribution in the generated flames; the development of high fidelity predictive models of these flames and the adaptability of the nozzle to different fuels’ compositions, are all needed if this technique is to be developed further.