基于局部线性模型树的高压共轨柴油机排放模型

Emission Model of High-pressure Common Rail Diesel Engine Based on Local Linear Model Tree

为研究面向闭环控制的柴油机在线排放模型,以1台高压共轨、涡轮增压中冷柴油机的转速、扭矩、空燃比、燃烧始点、燃烧重心、燃烧终点、最高燃烧温度、最大缸内压力等运转和燃烧的各项参数为基础,运用局部线性模型树对排放物HC,CO,CO2,NOx和烟度进行了仿真研究。研究结果表明,以转速、扭矩、空燃比为输入时,CO,CO2,NOx的仿真结果与试验值具有较好的一致性,以转速、扭矩、空燃比、燃烧重心为输入时,HC、烟度的仿真结果与试验值具有较好的一致性。各排放的期望响应与仿真输出的平均误差在10%以内,线性回归相关系数达到0.96以上。各个排放物的仿真过程单独进行时,可以得到较好的仿真效果。因此,局部线性模型树模型适用于高压共轨柴油机排放物的仿真。

In order to research the online emission model of the closed-loop control diesel engines ,the hydrocarbon (HC ) , carbon monoxide (CO) ,carbon dioxide (CO2 ) ,nitrogen oxides (NOx ) and soot of a high-pressure common rail ,turbocharged and inter-cooled diesel engine were simulated with the local linear model tree based on the operation and combustion parameters such as the speed ,torque ,air-fuel ratio ,combustion timing ,combustion center ,combustion termination ,maximum combus-tion temperature and maximum in-cylinder pressure .The results indicate that the simulated results of CO ,CO2 ,and NOx emis-sion and the simulated results of HC and soot emission have good consistency with their experimental values when using the speed ,torque and air-fuel ratio as the input and using those parameters and combustion center as the input respectively .The average error between expected response and simulated output is lower than 10% and the linear regression correlation coeffi-cient is beyond 0 .96 .Good simulation effect achieved when each of the emissions was simulated separately .Therefore ,the lo-cal linear model tree is appropriate for the emission simulation of high-pressure common rail diesel engine .