Search Results

Diesel fuel injection control systems, which are one of the expected solutions for meeting emission regulations, and achieving economical fuel consumption, consist of many parameters. To optimize this system, we have developed an adaptive multi-objective genetic algorithm (MOGA) method for real-time optimization. This method provides Pareto approximation set for the required objectives. Thus engine specifications can be precisely measured and minimize evaluation time. Pre-selection, which is an acceleration method for MOGA, is employed. By using this method we achieved the faster search of three-stage injection sets and PID controller gain.

Spark knock suppression with hydrogen addition was investigated at two engine speeds (2000 rpm and 4800 rpm). The experimental results showed that the spark knock is strongly suppressed with increasing hydrogen fraction at 2000 rpm while the effect is much smaller at 4800 rpm. To explain these results, chemical kinetic analyses with a two-zone combustion model were performed. The calculated results showed that the heat release in the end gas zone rises in two stages with a remarkable appearance of low temperature oxidation (LTO) at 2000 rpm, while a single stage heat release without apparent LTO process is presented at 4800 rpm due to the shorter residence time in the low temperature region.