Optimization on Fuel Injection Parameter of Heavy-Duty Diesel Engine Based on Multi-Objective Genetic Algorithms

Cao Lu, Lou Diming, Lin Haoqiang, Zhu Zewei (United Automotive Electronic Systems Co., Tongji University)

CHALLENGE -  Injection parameters have significant influence on economy, combustion characteristics and emission performance of diesel engines. Different performances of diesel engines tend to have contradictory relationships. In order to secure a better compromising result, multiple performance objectives need to be comprehensively considered and injection parameters need to go through combination optimization. A six-cylinder, exhaust valve turbocharged and direct injection common rail diesel engine is selected as the testing machine.

SOLUTION -  By using the GT-power software to establish the computing model of diesel engines and employing the modeFRONTIER multiobjective genetic algorithm (MOGA), this study carried out optimal research on performance indicators of diesel engines such as BSFC, NOX, soot, HC and CO, achieved the optimal solution to combination of injection parameters, and analyzed the relationship between optimal objective and fitness.​


BENEFITS - The optimization results achieved by using modeFRONTIER show hat the brake specific fuel consumption (BSFC), NOX emission, soot, HC emission and CO emission are respectively 212.23g/(kW·h), 0.188g/(kW·h), 4.53g/(kW·h), 0.045g/(kW·h) and 11.31g/(kW·h) when weighting factor of the five indexes are respectively 0.05, 1, 0.05, 1, 0.5. The corresponding optimum combination of injection parameters including rail pressure (112MPa), main injection timing (-10°CA), pilot interval (6.9°CA) and pilot amount (2.1mg) is also obtained.​