Abstract
The diesel engine has a high thermal effiency, but the serverity of emmision regurations is increasing. In such a circumstance, various exhaust decrease devices is studyed. But sevral exhaust device is affected by a fuel property. In this study, we wxamined possibility of fuel kind distinction by a neural network work system useing diesel fuel, kerosene and heavy oil. The neural network work system was able to give correct answers more than 80% by learning over 5000 cycles.
