王雪翠;滕勤;谈建;杨瑜;

合肥工业大学机械与汽车工程学院

    摘要：基于排放测量数据,采用量化共扼梯度法(SCG)建立了煤层气发动机的BP神经网络排放预测模型,检验结果证实了模型的准确性。为全面了解煤层气发动机的排放特性和制定合理的排放控制策略,借助于该模型,通过模拟研究对发动机的排放性能进行了预测,并分析了预测结果。

    关键词： 煤层气发动机;;废气排放;;预测模型;;量化共轭梯度法

    Abstract: A BP neural network model for predicting exhaust emission of the coal-bed gas engine under steady-state operation condition was established based on exhaust emission experiment data using Scaled Conjugate Gradient(SCG) method.The results show that the simulation data and the measured data was in good agreement and the model was accurate enough.In order to know the emission performance of the coal-bed gas engine and get proper control strategy,the model was used to predict emission performance of the coal-bed gas engine by means of simulation study based on this model.And some prediction results are presented.

    Keywords: Coal-bed Gas Engine;Exhaust Emission;Prediction Model;Scaled Conjugate Gradient Method

参考文献：

[1]左承基,钱叶剑,顾红兰.煤层气发动机和排放模型及实验[J].农业机械学报,Vol.36,2005(9):1~4.

[2]周斌.一种预测内燃机稳态排放特性的方法[J].西南交通大学学报(自然科学版),Vol.39,2004(2):135~138.

[3]孙俊,高孝洪.BP网络在双燃料发动机排放预测中的应用研究[J].武汉理工大学学报(交通科学与工程版),Vol.31,2003(3):350~353.

[4]刘震涛,费少梅.基于RBF网络的双燃料发动机排放模型[J].农业机械学报,Vol.33,2002(4):1~4.

[5]俞明,林冬燕,孙国斌等.基于广义回归神经网络的发动机排放预测模型[J].华南理工大学学报(自然科学版),Vol.31,2003(1):51~56

[6]李军,秦大同.基于BP神经网络技术在汽车发动机排放检测中的应用[J].交通标准化,2004(12):81~83.

[7]左承基,郭威,钱叶剑等.火花点火式变组分煤层气发动机的工作稳定性和排放特性[J].内燃机学报,Vol.21,2003(5):329~332.

[8]董长虹.Matlab神经网络与应用[M].北京:国防工业出版社,2005:73~74.

[9]Moller A F.A scaled conjugate graident algorithm for fast super-vised learning.Neural Networks,Vol.6,1997:525~533.

[10]Craig Garlin Carmichael.A study of the accuracy,complete-ness and efficiency of artificial neural networks and related in-ductive learning techniques[D].lowaStateUniversity,2001:11~13.