Document Type : Research Paper
Faculty of Chemical and Petroleum Engineering, University of Tabriz, Tabriz, Iran
Mechanical Engineering Department, King Fahd University of Petroleum and Minerals
Typical models are employed to estimate the product yields of delayed coking units using complicated and multistep calculations. In current study, a new first-order mathematical model have been proposed to estimate delayed coking products yield utilizing the Volk’s model as the baseline. The modified coefficients of Volk's model for industrial level are 0.634, 0.589, 1, and 1.116 for gas, gasoline, gasoil, and coke yield prediction, respectively. In Compare to other models, the proposed model showed very close and similar trend with industrial data in yield prediction, and the average error for gas production was 0.25%. For the gasoline, almost all of the other models have overestimated efficiency. However, current model prediction was obtained close to the industrial data with average error of 14 % that is almost three times better than the Volk’s model prediction (which was the most accurate model previously). The industrial data for the gasoil was underestimated by all previous models. However, the average error of proposed model for prediction of gasoil yield was 13% while other models’ estimation error is much higher. For the coke production, this newly developed model is the most accurate one compared to other predictive models.