No abstract

No abstract

This work investigates the effects of premixed combustion kinematics in pre-chamber volumes on the development of emitted hot jets from the igniter. The effects of fuel type, orifice diameter, and ignition location are evaluated experimentally, with high-speed OH* and CH* chemiluminescence imaging, and computationally with Large-Eddy Simulations (LES). The imaging experiments allowed for simultaneous viewing of combustion processes within a quartz chamber and of the developing jet flow. Results from these experiments provided insight on the temporal evolution of the jet relative to the growth of an ignited kernel within the chamber, as well as information on the emission or lack of emission of radical species from the chamber. Computational results provided data on the temporal behavior of the pressure within the chamber and profiles of the high velocity flow through the orifice. These results, combined, have shown that dependent on the strain rate and effective orifice size, local quenching of radical species at the orifice occurs which fundamentally change whether hot products, reactive layers, or both are present in the turbulent jet emission. The dynamic structure and composition of the turbulent jet controls its relevance as an effective ignition source.

Large-bore natural gas engines may use pre-chamber ignition. Despite extensive research in engine environments, the exact nature of the jet, as it exits the pre-chamber orifice, is not thoroughly understood and this leads to uncertainty in the design of such systems. In this work, a specially-designed rig comprising a quartz pre-chamber fit with an orifice and a turbulent flowing mixture outside the pre-chamber was used to study the pre-chamber flame, the jet, and the subsequent premixed flame initiation mechanism by OH* and CH* chemiluminescence. Ethylene and methane were used. The experimental results are supplemented by LES and 0D modelling, providing insights into the mass flow rate evolution at the orifice and into the nature of the fluid there. Both LES and experiment suggest that for large orifice diameters, the flow that exits the orifice is composed of a column of hot products surrounded by an annulus of unburnt pre-chamber fluid. At the interface between these layers, a cylindrical reaction zone is formed that propagates in the main chamber in the axial direction assisted by convection in the jet, but with limited propagation in the cross-stream direction. For small orifice diameters, this cylinder is too thin, and the stretch rates are too high, for a vigorous reaction zone to escape the pre-chamber, making the subsequent ignition more difficult. The methane jet flame is much weaker than the one from ethylene, consistent with the lower flame speed of methane that suggests curvature-induced quenching at the nozzle and by turbulent stretch further downstream. The velocity of the jet is too high for the ambient turbulence to influence the jet, although the latter will affect the probability of initiating the main premixed flame. The experimental and modelling results are consistent with ongoing Direct Numerical Simulations at ETH Zurich.

The response of three flames with different degrees of premixedness (fully premixed, non-premixed with radial, and non-premixed with axial fuel injection) to acoustic oscillations is studied experimentally. The flames were imaged using OH* chemiluminescence and OH planar laser-induced fluorescence at 5 kHz. In addition to a flame kinematics analysis, the amplitude dependence of the transfer function was calculated. The dominant spatial structures of the heat release and their periodicity were examined using the proper orthogonal decomposition (POD) method. The Non-Premixed system with Radial fuel injection (NPR) showed the highest response to acoustic forcing, followed by the fully premixed and the Non-Premixed system with Axial fuel injection (NPA). In addition, the response of the nonpremixed system with radial fuel injection was greater than that of the fully premixed system for various bulk velocities U, global equivalence ratios φ, forcing amplitudes A, and forcing frequencies f . In the fully premixed system, the heat release modulation was mainly through flame surface area modulation, while in the NPR system, both the flame area and the equivalence ratio modulations were found to be important mechanisms of the heat release oscillations. About 70% of the energy of the total fluctuations in the NPR case was contained in the first four POD modes, a percentage that decreased with overall equivalence ratio, but only this dropped to about 40% for the NPA flame. The frequency spectra of the coefficients of the POD modes exhibited peaks at the forcing frequency, with increasing broadband contributions in higher modes and for the NPA flame.ARTICLE HISTORY