基于响应面和遗传算法的柴油机瞬态过程喷油参数优化

摘要

基于台架试验数据,利用响应面法建立了某工程机械用柴油机瞬态过程喷油参数与性能的近似高精度模型,基于此模型采用遗传算法对瞬态过程喷油参数分别进行离线优化研究.结果表明:采用单目标优化确定的燃油消耗率(BSFC)、NOx比排放量和颗粒质量(PM)比排放量的优化极限分别可达180.23 g/(kW·h),8.92 g/(kW·h)和0.011 8 g/(kW·h),相对原机可降低多达4.5%,34.0%和37.3%.双目标优化的Pareto解集表明,相比于同时优化BSFC和NOx比排放量,BSFC和PM比排放量更容易同时得到优化.采用权重因子适应度函数的三目标优化结果对应的BSFC,NOx比排放量及PM比排放量分别为184.70 g/(kW·h),12.62 g/(kW·h)和0.012 2 g/(kW·h),较原机分别降低2.1%,6.6%和35.3%.改进优化模型后,性能优化Pareto解集对应的BSFC和PM比排放量水平都非常接近其优化极限,但NOx比排放量相对其优化极限仍然较高.%Based on the bench test data, an approximate high-precision model of fuel injection parameters and performance during transient process for a diesel engine applied in engineering field was established by using response surface method.Then genetic algorithm was used to optimize injection parameters offline.Finally, the best optimized values of brake specific fuel consumption(BSFC), NOx and PM emission by single objective methods were 180.23 g/(kW·h), 8.92 g/(kW·h) and 0.011 8 g/(kW·h), which decreased by 4.5%, 34.0% and 37.3% respectively.Pareto solution of double objective optimization showed that BSFC and PM emission were easier to optimize simultaneously comparing with BSFC and NOx emission.Triple objective optimization results of BSFC, NOx and PM emission based on the fitness function of weight factor were 184.70 g/(kW·h), 12.62 g/(kW·h) and 0.012 2 g/(kW·h), which decreased by 2.1%, 6.6% and 35.3% respectively.With improved optimization model, the correspondent BSFC and PM emission of Pareto solutions for performance optimization were close to limit values of single objective optimization, but NOx emissionwas still high.