基于BP神经网络的电控LPG-柴油双燃料内燃机碳烟排放的预测  

Emissions Prediction for Electronically Controlled LPG-diesel Engine Based on Back Propagation Neural Network

作　　者：李捷辉[1] 黄建军[1] 朱磊[1] 

机构地区：[1]江苏大学汽车与交通工程学院,江苏镇江212013

出　　处：《拖拉机与农用运输车》2006年第4期56-58,共3页Tractor & Farm Transporter

基　　金：国家自然科学基金资助项目(50276026;50276021);江苏省自然科学基金资助项目(BK2001092)

摘　　要：建立了一个基于BP神经网络的碳烟排放模型,用它来预测电控LPG-柴油双燃料的碳烟排放。模型将转速、油泵齿条位置、油门开度、LPG喷气时间、冷却水温度、转矩作为输入变量。本文采用准牛顿反向传播算法,对试验样本数据进行训练,并对此模型进行了泛化检验,将预测结果与试验结果进行了比较。此排放模型可以指导双燃料内燃机的排放性能优化试验,减少试验工作量。Based on the BP neural network,this paper established a particulate emission model for the electronically controlled LPG-diesel dual fuel engine, This model took the parameters of engine speed, fuel pump rack position, pedal, LPG injection time,cooling water temperature and torque as the input variables. Using BFGS arithmetic to train the experimental data and verify the extensive ability of the model, the comparison between the prediction and experimental data was carried out. The emission model can be used to guide the experimentation of emission performance optimization and to reduce the experimental workload.

关 键 词：LPG-柴油双燃料内燃机 BP神经网络 排放模型 准牛顿反向传播算法 

分 类 号：TK40[动力工程及工程热物理—动力机械及工程]