Experimental Investigation of the Variations of Early Flame Development in a Spark-Ignition Direct-Injection Optical Engine

David L. S. Hung,Hao Chen,Min Xu,Jie Yang,Hanyang Zhuang
DOI: https://doi.org/10.1115/1.4027256
2014-01-01
Journal of Engineering for Gas Turbines and Power
Abstract:Experiments under two intake air swirl levels (swirl ratios of 0.55 and 5.68) were conducted in order to investigate the early flame development of combustion in a single-cylinder spark-ignition direct-injection engine. The engine was equipped with a quartz insert in the piston, which provided an optical access to its cylinder through the piston. The crank angle resolved combustion images through the piston window and in-cylinder pressure measurements of 250 cycles were simultaneously recorded for both swirl levels at a specified engine speed and low load condition. The early development, size, and spatial characteristics extracted from the flame images were analyzed as a function of crank angle degrees after the ignition. The experimental results revealed that the early flame development was strongly influenced by the highly directed swirl motion of intake-air into the combustion cylinder. The location of the start of the flame kernel relative to the spark plug position also changed intermittently at different swirl levels. While the structure of the early flame was found to be similar for both swirl levels, the starting location of the flame showed a vast difference in how the flame progressed. In general, the flame kernel was formed two crank-angle degrees after spark timing for the high swirl level, which was four crank-angle degrees earlier than that of the low swirl case. For the low swirl flow, the early combustion showed more cycle-to-cycle variation in terms of both the flame size and centroid location. It was quantitatively shown that increasing the swirl ratio from 0.55 to 5.68 could reduce the cycle-to-cycle variation of the early flame structure, resulting in about three to four crank-angle degrees advance of the peak pressure location and a 1% improvement for the coefficient of variation (COV) of the indicated mean effective pressure (IMEP).
What problem does this paper attempt to address?