汽油机动力性的前馈神经网络模型  

Research on the Model of the Feedforward Neural Network About Power Characteristic of Automotive Gasoline Engines

作　　者：吴晓红[1] 蔡惠京[1] 

机构地区：[1]中山学院,广东中山528402

出　　处：《系统工程与电子技术》2002年第4期79-81,96,共4页Systems Engineering and Electronics

基　　金：中山市科技计划项目基金资助课题

摘　　要：如何优化汽油机的动力性 ,是研究者致力追求的目标。利用多层前馈神经网络的反向传播算法特性 ,建立了汽油机动力性与点火正时、空燃比、发动机转速、油门开度等之间关系的神经网络模型 ,用以分析、预测、优化汽油机动力性能 。In this paper, we investigate the problem about the power characteristic of automotive gasoline engines applying multilayer feedforward neural networks. The neural network is trained by a large number of test data. A gasoline engine performance model based on the neural network is developed. The inputs of the neural network model are gasoline RPM, throttle position, air-fuel ratio and ignition timing, the outputs are gasoline power, torque, specific fuel consumption, the emissions of HC,CO and Nox. The function, prediction and reliability of the neural network model are verified with test data.

关 键 词：前馈神经网络 BP学习算法 汽油机 动力性 

分 类 号：TK41[动力工程及工程热物理—动力机械及工程] TP183[自动化与计算机技术—控制理论与控制工程]