基于神经网络的预测算法在CNG发动机空燃比控制中的应用研究  

Research on Application of Air-Fuel Ratio Control of CNG Engine Based on Neural-Network Predictive Algorithm

作　　者：刘一鸣[1] 陈永全[1] 花志远[1] 

机构地区：[1]合肥工业大学机械与汽车工程学院,安徽合肥230009

出　　处：《小型内燃机与摩托车》2012年第6期46-49,共4页Small Internal Combustion Engine and Motorcycle

摘　　要：针对天然气发动机本身的非线性特点和空燃比传输延迟的特性,提出了一种基于神经网络预测的空燃比控制策略,利用Matlab/Simulink建立控制器的算法模型,以dSPACE公司的MicoAutoBox为算法实施平台,在NQ150N型天然气发动机上进行了实验测试,实验结果表明,与普通PID控制算法相比,基于神经网络预测的控制算法稳态性能优良,能明显改善过度工况空燃比的控制效果。For air-fuel ratio signal of a CNG engine,there exist transmission delay and nonlinearities,which affects the control accuracy of air-fuel ratio using directive air-fuel ratio sensor.A neural-network predictive algorithm to air-fuel ratio is provided in this paper.The control algorithm was built under the Matlab/Simulink environment and implemented on NQ150N CNG engine with MicroAutoBox hardware of dSPACE.The actual test results in experiments show that compared with the PID algorithm,this method has better steady performance and better control effect under transient conditions.

关 键 词：神经网络预测控制 空燃比 DSPACE CNG发动机 

分 类 号：TK43[动力工程及工程热物理—动力机械及工程]