Journal of Marine and Engineering Technology, no.1, pp.24-33, 2021 (Peer-Reviewed Journal)
In this study, performance and emission predictions of four stroke, naturally aspirated, water cooled single cylinder diesel engine were carried out with created artificial neural networks. Partial load experiments of the engine have been conducted and power, specific fuel consumption values and CO2, CO, NOx emissions were recorded. Obtained values were used in modeling studies. In the study, 59 data were used, 80% of this data was used as training and 20% as test data. The data are modeled with multi layer, Back-Propagation (BP) and Radial Basis Function artificial neural network. According to the results obtained, the model predictions are consistent with the experimental results and it has been observed that emission, power and specific fuel consumption can be predicted with limited data of the engine such as speed and load. Also, best results for estimation of emissions as one of the most important problems of diesel engines, are obtained from BP compared to RBF.