刘文龙,高莹,尤岳林,等.基于高阶模型与扩展卡尔曼滤波的柴油机选择性催化还原系统输出状态观测[J].内燃机工程,2025,46(2):82-93.
基于高阶模型与扩展卡尔曼滤波的柴油机选择性催化还原系统输出状态观测
Output State Observation of Diesel Engine Selective Catalytic Reduction System Based on Higher Order Model with Extended Kalman Filtering
DOI：10.13949/j.cnki.nrjgc.2025.02.010
关键词:柴油机  选择性催化还原  扩展卡尔曼滤波  观测器
Key Words:diesel engine  selective catalytic reduction（SCR）  extended Kalmanfilter（EKF）  observer
基金项目:国家自然科学基金项目（52076102）
作者单位E-mail
刘文龙 吉林大学 汽车仿真与控制国家重点实验室，长春 130025 wenlongliu1012@163.com 
高莹* 吉林大学 汽车仿真与控制国家重点实验室，长春 130025 gaoying@jlu.edu.cn 
尤岳林 吉林大学 汽车仿真与控制国家重点实验室，长春 130025 youyl22@mails.jlu.edu.cn 
姜长文 吉林大学 汽车仿真与控制国家重点实验室，长春 130025 jiangcw21@mails.jlu.edu.cn 
华韬轶 吉林大学 汽车仿真与控制国家重点实验室，长春 130025 tyhua21@mails.jlu.edu.cn 
夏博聪 吉林大学 汽车仿真与控制国家重点实验室，长春 130025 xiabc22@mails.jlu.edu.cn 
摘要点击次数: 130
全文下载次数: 105
摘要:简化选择性催化还原（selective catalytic reduction， SCR）高阶模型方程并优化了化学反应参数，并结合SCR高阶模型设计了可变时间步长扩展卡尔曼滤波（variable time-step extended Kalman filter， VTS-EKF）观测器。假设两种观测情况：情况1为下游NOx传感器故障时，观测储氨覆盖率和下游NOx体积分数；情况2为下游NH3传感器故障的情况下，观测储氨覆盖率和下游NH3体积分数。通过动态工况台架试验验证了这两种情况的观测效果。结果显示：情况1中储氨覆盖率观测值与模型计算值的决定系数为0.99，下游NOx体积分数模型计算值和试验值的平均绝对误差（mean absolute error， MAE）为；情况2中储氨覆盖率观测值与模型计算值的为0.99，下游NH3体积分数VTS-EKF观测值和试验值的MAE为。结果表明，SCR高阶模型结合VTS-EKF，可实现SCR输出状态的精准观测。
Abstract:The high-order selective catalytic reduction（SCR） model equations were simplified， and the chemical reaction parameters were optimized. And a variable time-step extended Kalman filter （VTS-EKF） observer was designed in conjunction with the high-order SCR model. Two observation scenarios were considered. Case 1 was involved in observing the ammonia storage coverage and the downstream NOx volume fraction when the downstream NOx sensor failed， while Case 2 was involved in observing the ammonia storage coverage and the downstream NH3 volume fraction when the downstream NH3 sensor failed. The observation effects of these two cases were verified through dynamic condition bench tests. The results showed that the coefficient of determination （） between the observed and model calculated values of ammonia storage coverage in Case 1 was 0.99， and the mean absolute error （MAE） between the model calculated and experimental values of downstream NOx volume fraction was . In Case 2， the between the observed and modeled values of ammonia storage coverage was 0.99， and the MAE between the VTS-EKF observed and experimental values of downstream NH3 volume fraction was . The results demonstrate that the SCR high-order model combined with VTS-EKF can accurately observe the SCR initial state.
查看全文  HTML   查看/发表评论