关键词: 船用燃料油, 黏度, 预测模型, 逻辑模型

Abstract: The feasibility to predict the viscosity of heavy marine fuel oil using the existing viscosity prediction models was studied. Several common viscosity physical models were compared and selected the best one based on the analysis of experimental data. The Cragoe model was modified based on the heavy marine fuel oil database and the viscosity reduction data by dilution, the impact of the blending mechanism on the relative error of the prediction model was analyzed. The results showed that for the heavy fuel oil blending components commonly used in the current market, the Cragoe viscosity model has minor prediction errors due to the fact that the Cragoe viscosity prediction model is not limited by the viscosity ratio of the component oils, and is the best model suited for heavy marine fuel oils. The proposed correction model can further reduce the error in predicting the viscosity of heavy marine fuel oil. The analysis of multi-component blending results showed that if the profile of the viscosities distribution of the components is in a gradient form, the prediction error can be reduced. In addition, when the residual oil and dilute components are blended, the complexing effect of asphaltenes affects the prediction accuracy of the model to some extent.

Key words: marine fuel oil, viscosity, prediction model, logic model