Characterisation of a spark ignition system by planar laser-induced fluorescence of OH at high repetition rates and comparison with chemical kinetic calculations

This study reports the application of a novel, high speed laser-detector system for the time-resolved study of flame propagation in a well-controlled spark ignition system. The ignition system allowed full and reproducible control over the energy deposited during breakdown and the ensuing arc discharge of the spark plasma. Ignition was performed in a closed vessel which was filled with stoichiometric mixtures of methane and air. Four sequential snapshots of twodimensional OH distributions were recorded during single ignition events by the use of planar laser-induced fluorescence (PLIF). From these OH distributions flame front velocities have been extracted with an accuracy of better than 2%. Onedimensional numerical simulations of the ignition event including detailed chemistry and transport processes have been performed. Experimental results and results from the simulations have been compared to each other with respect to flame front velocities as well as spatial concentration profiles of OH radicals. In general a good agreement was obtained. In this way the ignition system was carefully characterised. PACS: 07.50.-e; 07.60.-j; 47.11.+j In many technical applications an electrical spark initiates a combustion process. A detailed knowledge of the chemistry and physics of the ignition and the subsequent flame propagation in its different modes is necessary to optimise, for example, the design of igniters used in lean-operation I.C. engines [1–3]. On the other hand, this knowledge is for example of great importance concerning the prevention of accidental ignitions in the chemical industry [4] or in atomic power plants [5]. The present study reports on a detailed investigation and characterisation of spark-ignited combustion events. The flame propagation in premixed methane–air mixtures under precisely defined conditions was studied by means of planar laser-induced fluorescence (PLIF) of the OH radical. Furthermore, the system was investigated by numerical simulations and detailed comparisons with the experimental results were carried out. The conditions under which the present investigations were performed are sufficiently general to aid an improvement in the understanding of a wide variety of problems of technical and academic interest. A typical electrical spark of commercial ignition systems can be divided into three phases [6]. First, a short breakdown phase is initiated which creates a conducting plasma channel between two electrodes. Temperatures in this phase have been reported to reach up to 60 000 K [6]. Owing to the high pressure increase inside the plasma channel a shock wave is generated which leaves the activated volume within approximately 1 μs. This shock wave heats the surrounding gas it passes through [7–9]. After the formation of the plasma channel, the charge stored in the electrodes (which act like a capacitor) is discharged very rapidly (in the order of ns) [10]. The high conductivity of the plasma channel causes a drop in electrical resistance and the arc phase is initiated. This second phase is characterised by a much lower potential drop and temperature (up to 6000 K [6]). When the cathode fall is building up, the glow phase begins. Typically temperatures decrease to approximately 3000 K in this third phase of the spark ignition process. The conversion efficiency from electrical to thermal energy of the gas phase decreases from the breakdown to the glow phase [11] because of heat loss to the electrodes. At the surface of the plasma channel a layer of highly reactive radicals and atoms is formed at high temperatures giving rise to a growing flame kernel. A few hundred μs after the breakdown the surface temperature is lower than 3000 K [6]. For this range the “high-temperature chemistry” of hydrocarbon combustion applies for the description of the ensuing ignition process [12]. The further development of this initial flame kernel is of course influenced by a variety of parameters. Depending on stoichiometry, heat losses to the electrodes, and the turbulence intensity of the load the flame kernel may extinguish, a self-sustaining flame propagation (deflagration) may ensue, or a detonation may develop. Spark ignition has been investigated by a number of researchers in closed vessels [6, 7, 13–15] and in flow reactors [16, 17]. An excellent survey of early investigations is given in [18]. In more recent work [6] the initiation and subsequent flame propagation by breakdown, arc, and glow phase were investigated for a variety of different CH4–air mixtures