项目名称: 基于灰箱模型的柴油机微粒排放虚拟传感器研究
项目编号: No.51266015
项目类型: 地区科学基金项目
立项/批准年度: 2013
项目学科: 能源与动力工程
项目作者: 何超
作者单位: 西南林业大学
项目金额: 56万元
中文摘要: 微粒是柴油机排放的主要污染物之一,尤其是细小微粒,对人体健康危害极大。随着柴油机-后处理器系统复杂程度的增加,基于闭环控制的集成排放管理系统成为研究热点,其中排放传感器的研发是关键。本课题拟以柴油机微粒排放传感器为研究对象,以高压共轨柴油机电控单元ECU信号为基础,结合影响柴油机-后处理器系统排放微粒质量和数量的物理化学规律,运用人工神经网络、遗传规划、符号回归等智能算法,建立柴油机微粒质量和数量实时排放的灰箱模型,并以排放实验数据进行校验,从而构建柴油机微粒排放虚拟传感器。该传感器可用于柴油机-后处理器系统的闭环控制以及自诊断系统(OBD)的排放监测,为超低排放柴油机的开发以及在用车排放管理提供理论支持和科学依据,为控制柴油机细小微粒排放奠定技术基础。
中文关键词: 柴油机;微粒;虚拟传感器;局部线性模型树;遗传规划
英文摘要: Particulate matter (PM) is the most important pollutant of diesel engine, especially the fine particles, which is very harmful to human health. With the increasing complication of diesel engine - aftertreatment system, many concerns were focused on the integrated emission management system, in which the development of emission sensor is very crucial. The PM virtual sensor of diesel engines will be studied in this project based on the signals of electronic control unit (ECU) of high-pressure, common rail diesel engine. The physical and chemical factors which influence the PM mass and number emissions will be investigated on a diesel engine with aftertreatment. The intelligent algorithms, such as artificial neural networks, genetic programming and symbolic regression, will be applied to build a gray-box model of PM mass and number emissions. Then the gray-box model will be validated by the experiment data in order to develop the virtual PM sensor of diesel engines. The PM sensor can be used on the closed-loop control of diesel engine - aftertreatment system and on-board diagnostics (OBD) of diesel engines. The research results of the project will provide academic support and scientific foundation for the development of ultra-low emission diesel engines and in-use compliance, as well as the technologies for control
英文关键词: diesel engine;particulate matter;virtual sensor;LOLIMOT;genetic programming