基于GMDH网络的船用增压柴油机性能预测及仿真-《集美大学学报（自然科学版）》

文章信息/Info

Title:: Performance Prediction and Simulation Analysis for 4190ZLC Marine Diesel Engine Based on GMDH Algorithm Neural Network

Author(s):: LIU LeiHUANG Jia-liang; (Marine Engineering InstituteJimei UniversityXiamen 361021China)

Keywords:: 4190ZLC marine diesel engine; GMDH neural network; performance prediction; computer simulation

摘要:: 采用数据分组处理（Group Method of Data Handing,GMDH）的神经网络分类方法，建立4190ZLC船用四冲程增压柴油机性能预测的数学模型.针对船用中速柴油机运行状况，考虑到其影响运行状态的因素，结合实验数据进行4190ZLC船用柴油机性能的预测及仿真分析.该模型解决了神经网络结构较大，计算耗时较长的问题.将该模型与BP（Back-Propagation,BP）前馈神经网络仿真结果进行比较，结果表明其仿真效果好于BP神经网络模型，并且该神经网络能较好地满足柴油机性能预测仿真的需求

Abstract:: A mathematical model of the 4190ZLC four-stroke turbocharged marine diesel engine for its performance prediction was established,using GMDH (Group Method of Data Handing) algorithm neural network classification methods.Taking into account factors that affect engine磗 performance and being combined with experimental data,the performance prediction and simulation analysis for 4190ZLC marine diesel engine under the medium-speed marine diesel engine operating conditions were carried out.The problems of large architecture and long calculation time for the neural network were solved by the model.Compared with BP(Back-propagation)，the feed-forward neural network model the simulation results of GMDH algorithm neural network classification model were better.The GMDH algorithm neural network model can well,meet the requirements of the diesel engine performance prediction simulation