摘要
基于某重型柴油机进行选择性催化还原(SCR)的后处理系统匹配,开发了SCR后处理控制器,采用高性能芯片MC9S12XEQ512进行了控制器硬件设计,并完成性能测试,对SCR系统控制策略进行了深入研究,利用催化器下游NO x传感器和废气流量计的测量值作为期望输出,采用有导师学习方式的BP神经网络开发了催化器上游NO x排放估算模型,并基于此模型设计稳态工况添蓝计量控制策略,提出了催化器载体瞬态温变滞后修正算法和瞬态工况NH3泄漏控制策略.结果表明:设计的SCR控制系统瞬态工况的鲁棒性好,NO x转化率控制在60%左右,NH3泄漏量最大值控制在2.5×10-5以下;设计的控制器应用到目标发动机,经ESC和ETC试验验证排放达到国Ⅳ标准.
To match the heavy diesel engine with selective catalytic reduction （SCR） postprocessing sys tem, the SCR postprocessing system controller was developed. SCR controller hardware was designed based on advanced chip MC9S12XEQ512 to realize performance test. The control strategy of SCR system was investigated in detail. Using the downstream NO3concentration and exhaust gas mass flow measured by NOx sensor and flowmeter as expected output, the raw NOx emission model was established with a teacher of BP neural networks for target diesel engine. Based on the BP model AdBlue dosing control strategy under steady state conditions, the heat hysteresis correction algorithm of catalyst substrate and the NH3 leak control strategy were proposed under transient operation conditions. The test results show that the proposed SCR control strategy has good robustness under transient conditions with NOx conversion ra tio of 60% and maximal NH3 leak quantity less than 2.5 x 10^ 5. Using the proposed controller in target engine, the ESC and ETC test validation shows that the heavy duty diesel engine can satisfy ChinaIV emission regulations.
出处
《江苏大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2014年第1期7-13,共7页
Journal of Jiangsu University：Natural Science Edition
基金
国家科技支撑计划项目(2011BAG06B04)
国家国际科技合作专项基金资助项目(2012DFA11180)
关键词
柴油机
选择性催化还原
添蓝
BP神经网络
SCR电控单元
diesel engine
selective catalytic reduction
AdBlue
BP neural networks
dosing control unit （DCU）
