The Prognostication Of NO_x-emission On Marine Diesel Engine Based On BP Neural Network

Marine diesel engine plays a very important role in navigation economy as dynamical equipment. But the emission-pollution produced by it cannot be ignored, and the trend of the pollution is more and more serious. In order to control the emission-pollution, IMO has promulgated 《THE PROTOCOL OF 1997 TO AMEND PARPOL 73/78 RESOLUTIONS ADOPTED AT 1997 CONFERENCE OF CONTRACTING GOCERNMENTS TO MARPOL 73/78》 ,and this protocol has become effective from May.19.2005. The author of this dissertation did some study on marine diesel engine NOx-emission based on the protocol.Firstly, it expatiates the mechanism of the marine diesel engine NOx producing and the international control code;introduces the test processes, notices and calculation methods of the marine diesel engine NOx-emission carried currently.Secondly, it briefly introduces the artificial intelligence and BP neural network, and then imposes the nonlinearity function of the BP neural network on the testing of the marine diesel engine NOx -emission. Upon the characteristics of the BP neural network, carries the test design, practice and data measurement, and a series of locale test. Among the testing process, brings in the variable edge uniform distribution and experimental design (U-D design), in order to make the test date more scientific and precise. Based on these, under the help of MATLAB program, uses the data gained from the training network to build BP neural network, and then forecasts the diesel engine NOx-emission and analyses the ability of the network.Finally, upon the instance of the test and BP neural network, discusses the method of optimizing the BP neural network in order to simulate the marine diesel engine NOx-emission.