Influence of the spark plug mounting angle on spark ignition process at different flow velocities

Influences of the spark-plug mounting angle on the spark ignition of gasoline-air mixtures in high velocity flows were investigated in a combustion chamber connected at the end of a rapid compression machine. Gasoline-air mixture with an air excess ratio of 1.9 was combusted. A spark plug was mounted on the combustor wall at angles of −30°, −15°, 0°, 15°, and 30°. Flow velocity at the spark plug was varied from 5 to 20 m/s by adjusting the driving pressure of the piston. Time resolved spark-discharge voltage and current were measured with a voltage and a current probe, and the discharge energy was calculated using them. Behaviors of the discharge channel were observed optically with a high-speed video camera in addition to simultaneous measurements of OH* chemiluminescence from the flame kernel with the other high-speed video camera. The results showed that the averaged induction time of flame kernel initiation increased monotonously with an increase in the plug mounting angle from −30° to 30° in flows with velocities of 5 and 10 m/s. However, the increasing trend became less obvious with velocities of 15 and 20 m/s, involving with larger variation in induction time. Discharge duration became shorter with an increase in the plug mounting angle regardless of the same setup. Negative mounting angle resulted in frequent short-circuits/restrikes at the early phase. On the other hand, positive mounting angles resulted in a longer discharge channel length, higher energy deposition before the first restrike and longer induction time of the first restrike. Although the discharge channel at positive mounting angles elongates along with the flow, the early flame kernel was developed toward the center of the swirl. Thereafter, flame propagated outward from the central region, where the flow speed was lower than the downstream area. Negative mounting angle led to a more oval flame kernel shape and the propagation closer to combustion chamber wall, which showed a shorter induction time of flame kernel initiation in the cases of 5 and 10 m/s. However, fragments of the flame kernel tended to be formed due to frequent short-circuits with the velocities of 15 and 20 m/s. This resulted in the large variation of the induction time.