Feasibility of Detecting Low-Speed Pre-Ignition (LSPI) and Suppressing the Subsequent Knock in High-Performance Spark Ignited Gasoline Engines, 03-R8428
Inclusive Dates: 10/21/14 – 02/20/14
Background — The spark-ignition engine has been known to exhibit several different abnormal combustion phenomena. Most of these abnormal combustion issues have been addressed with improved engine design or control schemes. However, in the next generation of SI engines, where turbocharging and direct injection technology are employed to increase brake power at low engine speeds for improved efficiency, a new phenomenon of abnormal combustion, described as low-speed pre-ignition (LSPI), has been exhibited. Cylinder pressure data and heat-release rates for normal combustion, light knocking conditions and LSPI are shown in Figure 1. Because of the violent nature of LSPI, damage to engine components can occur within only a few engine cycles. However, because LSPI events occur only sporadically and in an uncontrolled fashion, the causes for this phenomenon have not been successfully explained and solutions to suppress or prevent it are desperately needed.
Approach — The approach was to first detect the pre-ignition with cylinder pressure. Once detected, fuel would then immediately be injected with the goal of quenching the flame. This would suppress the severe knock event associated with LSPI. The fuel-air mixture would be reignited late in the cycle to minimize hydrocarbon emissions. This strategy is depicted in Figure 2.
Because LSPI is a somewhat random phenomenon, knock was induced by advancing ignition timing. The resulting knock intensity was very consistent. This enabled the resulting knock intensity to be compared with and without the suppression strategy enabled. Using this method, no difference could be found with the strategy enabled.
Accomplishments — While the goal of project was unsuccessful, it was recognized early on in the project before a large portion of the funding was spent. At the same time, other useful knowledge was gained. This project provided experience making control decisions within a fraction of a crank angle degree, which enables same cycle engine control. Typically decisions about injection and ignition timing are done within a loop that executes at a rate of about 10 ms. At 2,000 rpm, this results in decisions being made every 120 CA°. This ability has proved beneficial in other projects requiring triggering for optical equipment based off of combustion metrics. This helped enable combustion to be visualized for LSPI engine cycles.