Analysis of the Pre-Injection System of a Marine Diesel Engine Through Multiple-Criteria Decision-Making and Artificial Neural Networks

Abstract

The present work proposes several pre-injection patterns to reduce nitrogen oxides in the Wärtsilä 6L 46 marine engine. A numerical model was carried out to characterise the emissions and consumption of the engine. Several pre-injection quantities, durations, and starting instants were analysed. It was found that oxides of nitrogen can be noticeably reduced but at the expense of increasing consumption as well as other emissions such as carbon monoxide and hydrocarbons. According to this, a multiple-criteria decision-making (MCDM) model was established to select the most appropriate parameters. Besides, an artificial neural network (ANN) was developed to complement the results and analyse a huge quantity of alternatives. This hybrid MCDM-ANN methodology proposed in the present work constitutes a useful tool to design new marine engines.