5th Asia-Pacific Conference on Combustion, The University of Adelaide, Adelaide, Australia 17-20 July 2005 Multi-dimensional effects on Low Strain Rate Flame Extinction in Methane/Air Counterflow Non-premixed Flames J. Park 1 , C.B. Oh 2 , K.T. Kim 1 , J.S. Kim 1 and A. Hamins 3 1 School of Mechanical & Automotive Engineering Sunchon National University, 315 Maegok, Sunchon, Jeonnam, South Korea 2 Thermo-Fluid System Department, Korea Institute of Machinery and Materials, 171 Jang-Dong, Yuseong-Gu, Daejeon, 305-343, South Korea Building and Fire Research Laboratory , National Institute of Standards and Technology, 100 Bureau Drive, STOP 8663, Gaithersburg, MD 20899-8663 Abstract Flame structure and extinction mechanism of counterflow methane/air non-premixed flame diluted with nitrogen are studied by NASA 2.2 s drop tower experiments and two- dimensional numerical simulations with finite rate chemistry and transport properties. Extinction mechanism at low strain rate is examined through the comparison among results of 0-g experiment, 1D and 2D simulations with a finite burner diameter. A two-dimensional simulation in counterflow flame especially with a finite burner diameter is shown to be very important in explaining the importance of multidimensional effects and lateral heat loss in flame extinction, effects that cannot be understood using a one-dimensional flamelet model. Extinction mechanism at low strain rate is quite different from that at high strain rate. Low strain rate flame is extinguished initially at the outer flame edge, the flame shrinks inward, and finally is extinguished at the center. It is clarified from the overall fractional contribution by each term in energy equation to heat release rate that the contribution of radiation fraction with 1D and 2D simulations does not change so much and the overall fractional contribution is decisively attributed to radial conduction (“lateral heat loss”). The experiments by Maruta et al.[12] can be only completely understood if multi- dimensional heat loss effects are considered. It is, as a result, verified that the turning point, which is caused only by pure radiation heat loss, has to be shifted towards much lower global strain rate in 0-g flame. 1 Introduction Laminar non-premixed flames in a counterflow configuration have been studied extensively over the past 30 years. This configuration is useful because it minimizes enthalpy losses to the ducts, reduces multidimensional effects, and allows easy control of the reactant mixtures and the flow field. Because it is a relatively simple and well-defined system, many analytical and computational models have been developed to study flame extinction, structure, kinetics and radiation for this configuration [1-8]. One limitation of the counterflow configuration is that as the global strain rate decreases, buoyant instabilities, enthalpy losses and flame curvature can become significant due to the effects of gravity [9]. For this reason, a majority of studies have focused on high global strain rate flames in 1-g where buoyant effects are negligible. By performing experiments in 0-g, low strain rate flames can be attained which are free from buoyant instabilities and curvature effects. Several studies to date have used this strategy to study low strain rate counterflow flames [10-12]. The first comprehensive extinction measurement of very low strain rate flames in 0-g was reported by Maruta et al. [12]. The extinction of methane-air flames with N 2 added to the fuel stream was investigated using the 10 s drop tower located in Japan. The minimum methane concentration required to sustain combustion was measured to decrease as the strain rate decreased until a critical value was observed. As the global strain rate was further reduced, the required methane concentration increased. This behavior was denoted as a "turning point" and was attributed to the enhanced importance of radiation heat losses at low strain rates. The previous study [12] showed that radiation fractions at extinction conditions of low strain rate flames were less than 0.2. The result was, however, a direct outcome of 1D simulation based on calculations with point sources of infinite boundary where effects of finite duct diameter were not considered. To verify the extinction behavior addressed to radiation heat loss, further extended studies with a two-dimensional simulation may be required. The goal of this work is to determine the maximum mole fraction of nitrogen necessary for extinction of methane-air flames at any global strain rate as well as to examine the near extinction flame structure of these flames. Main concerns in the present study is to clarify the extinction mechanism especially at low strain rates through the comparison among results of 0-g experiment, 1D simulation, and 2D simulation. Especially multi-dimensional effects in experiments with a finite burner diameter are verified to affect flame structure and extinction mechanism considerably at low strain rate flames. 2 Experimental Methods A 15 mm diameter stainless steel counterflow burner was enclosed in a 25 L cylindrical chamber for 0-g experiments. The experimental hardware was mounted to a standard NASA drop rig “A” frame. The distance between the burner duct exits was 15 mm. To control each of the reactant gas flows, a very fast response time (≈50 ms) pressure controller was used in series with a critical flow orifice. The total system response time was equal to the flow control response time plus the residence flow time from the mixing tee to the flame zone. The system response time was estimated to be equal to 0.0665 and 0.6397 seconds for strain rates of 50 s -1 and 7 s -1 respectively. In the experiments described here, the global strain rate was varied from 7 s -1 to 50 s -1 . Each suppression measurement was performed by increasing the agent flow rate and simultaneously decreasing the fuel flow rate, while maintaining a constant global strain rate. 345