Research on ignition, flame growth and flame propagation in engine-like turbulence has produced widely varying correlations between turbulence parameters and flame speed. Some previous work has shown that the burning velocity observed in a given turbulence level depends on the flame size as well as the turbulence intensity and scale. This explains some of the previous experimental discrepancy and emphasizes the importance of measuring flame growth and turbulence effects over the range of interest for a given modelling requirement. This paper reports on an experimental study of flame growth from ignition sparks in spatially uniform, decaying turbulence similar to that found in engine combustion chambers. High speed schlieren video and pressure trace analyses were used to study 3-dimensional turbulent flame growth in a constant volume, cubical combustion chamber. Lean methane-air mixtures of 60%, 70%, 80%, 90% and 100% stoichiometric compositions were ignited at 1 atm and 300 K. Schlieren images of flame growth were recorded on high speed video at 2000 frame per second while the combustion chamber pressure rise was recorded concurrently. The combined high speed schlieren video and pressure trace analyses enables the study of flame growth from ignition spark until the end of combustion. Pre-ignition turbulence was generated by pulling a perforated plate across the chamber. The ignition-time turbulence intensity ranged from 0 to 2 m/s with integral scale of 8, 4, 2 or 1 mm depending on the plate hole size. The assumed turbulence intensity during flame propagation was continually adjusted to account for viscous decay and rapid distortion.
Laminar flame growth calculated from the pressure trace using a multi-zone thermodynamic model agreed well with the flame growth deduced from high speed schlieren video. Results showed that laminar flames grew at a quasi-steady rate determined by the laminar flame speed and chamber temperature and pressure. Turbulent flames were found to accelerate from the ignition spark until the flame was quenched by the walls, by which time it was many times larger than the turbulence length scale. Burning velocity results show that, for a given flame size, the normalized turbulent burning velocity, St/Sl, is approximately proportional to the normalized rms turbulence intensity, u'/Sl; where St, is the turbulent burning velocity, Sl, is the laminar burning velocity and u' is the rms turbulence intensity. The slope of this roughly linear relation increases with increasing flame size and decreases with increasing integral scale. In short, a given level of combustion chamber turbulence becomes progressively more effective in enhancing the burning velocity as the turbulent flame grows. For a fixed turbulence intensity and flame size, faster burning can also be achieved by using smaller scale turbulence.