This study is focused on the streamer-to-spark transition generated by an overvoltage nanosecond pulsed discharge under atmospheric pressure air in order to provide a quantitative insight into plasma-assisted ignition. The discharge is generated in atmospheric pressure air by the application of a positive high voltage pulse of 35 kV to pin-to-pin electrodes and a rise time of 5 ns. The generated discharge consists of a streamer phase with high voltage and high current followed by a spark phase characterized by a low voltage and a decreasing current in several hundreds of nanosecond. During the streamer phase, the gas temperature measured by optical emission spectroscopy related to the second positive system of nitrogen shows an ultra-fast gas heating up to 1200 K at 15 ns after the current rise. This ultra-fast gas heating, due to the quenching of electronically excited species by oxygen molecules, is followed by a quick dissociation of molecules and then the discharge transition to a spark. At this transition, the discharge contracts toward the channel axis and evolves into a highly conducting thin column. The spark phase is characterized by a high degree of ionization of nitrogen and oxygen atoms shown by the electron number density and temperature measured from optical emission spectroscopy measurements of N + lines. Schlieren imaging and optical emission spectroscopy techniques provide the time evolution of the spark radius, from which the initial pressure in the spark is estimated. The expansion of the plasma is adiabatic in the early phase. The electronic temperature and density during this phase allows the determination of the isentropic coefficient. The value around 1.2-1.3 is coherent with the high ionization rate of the plasma in the early phase. The results obtained in this study provide a database and the initial conditions for the validation of numerical simulations of the ignition by plasma discharge.

International audienceThis article presents a comparison between experiments and Large-Eddy Simulation (LES) of a spark ignition engine on two operating points: a stable one characterized by low cycle-to-cycle variations (CCV) and an unstable one with high CCV. In order to match the experimental cycle sample, 75 full cycles (with combustion) are computed by LES. LES results are compared with experiments by means of pressure signals in the intake and exhaust ducts, in-cylinder pressure, chemiluminescence and OH Planar Laser Induced Fluorescence (PLIF). Results show that LES is able to: (1) reproduce the flame behavior in both cases (low and high CCV) in terms of position, shape and timing; (2) distinguish a stable point from an unstable one; (3) predict quantitatively the CCV levels of the two fired operating points. For the unstable case, part of the observed CCV is due to incomplete combustion. The results are then used to analyze the incomplete combustion phenomenon which occurs for some cycles of the unstable point and propose modification of the spark location to control CCV

No abstract

This paper describes a compressible Large Eddy Simulation (LES) used to investigate cyclic variations for nonreacting flow in an optical single cylinder engine setup. The simulated operating point is part of a large experimental database designed to validate LES for cycle-to-cycle prediction, and constitutes a first step towards the realization of fired operating points. The computational domain covers almost the whole experimental setup (intake and exhaust plenums, intake and exhaust ducts, cylinder) to account for acoustic phenomena. The assessment of the computation is performed in two regions of the domain: the intake and exhaust duct predictions are compared to the results of a Helmholtz solver and the experiment (pressure transducers and Particle Image Velocimetry (PIV)) while the in-cylinder dynamics are compared to PIV measurements. The ability of the B. Enaux · V. Granet (B) · O. Vermorel 154 Flow Turbulence Combust (2011) 86:153-177 developed methodology to capture the correct level of cycle-to-cycle variations is demonstrated considering in-cylinder pressure and velocity fields predictions. Cycleto-cycle variations in velocity are highlighted and localized using a proper orthogonal decomposition analysis.

No abstract

No abstract

An automated spectral fitting algorithm (named Thermo NO-LIF) designed for the extraction of temperature information from experimental NO-LIF spectra is presented and analysed. With the aid of the Thermo NO-LIF, the high-accuracy (0.5%) of multi-line NO-LIF thermometry are demonstrated in the burnt-gases region of a near-adiabatic Bunsentype premixed flame. The results are compared with the temperature measured using N 2 spontaneous Raman scattering (SRS). The current study also applies and analyses the algorithm and the thermometry technique in a premixed flat-flame under various subatmospheric pressures. The results suggest that choosing a long excitation scan range with as few overlapping lines as possible can maximise the performance of the multi-line NO-LIF thermometry. The uncertainty arising from the overlapping lines is especially crucial at atmospheric pressure. Additionally, the current study compares the performance of multi-line NO-LIF thermometry using narrowband and broadband lasers. Thermo NO-LIF is available to anyone interested in applying multi-line NO-LIF thermometry.

No abstract