The Study Of The Emissions Prognostication On Marine Diesel Engine Based On GRNN

With the continuous development of the global water transportation industry,the massive utilization of ships caused serious environment pollutions.In order to reduce harmful emissions of ships effectively,the International Marine Organization formulated series relevant laws and regulations.This study aims to forecast the emissions of ships,and to provide important basis for pollution reduction.Break through the traditional method,this paper introduced the idea of the GRNN to seek for a new mean for prediction of marine diesel engine exhaust emissions.Based on MATLAB software,established generalized regression neural network prediction model for marine diesel engine on exhaust emissions.The realization of the GRNN prediction without complex test equipment and conditions,and has nothing to do with the mathematical model of marine diesel engine.It only needs a small amount of test data.This method can efficiently reduce the cost.It is a kind of accurate and effective method.This article is mainly to discuss how to make choices of the GRNN to predict exhaust emissions for marine diesel engine based on the advantages.(1)This paper expounded the relevant restriction of laws and regulations,about the diesel engine exhaust emissions,the composition of marine diesel engine emissions,the harm and generating mechanism,and some basic knowledge about neural network.At the same time this paper emphatically introduced the GRNN model structure and the theoretical knowledge.(2)This paper introduced the relevant emission test system and important test instruments,and choosed the best cam working segment which presented best efficiency to measure the test data.Then using experimental data obtained in the MATLAB environment to establish a network of marine diesel engine GRNN emission model for predicting and forecasting,verifying the accuracy of the prediction model through simulation and analysis of network generalization ability.(3)Make sure the emission prediction interface based on Java for marine engine is able to run in the background,as well as to output and display the correct emission data through the GRNN emission model for marine engine built on Java.