Predicting Coke Characteristics from the Properties of the Raw Materials and the Coking Conditions

Abstract

In oil refining, it is a priority to intensify coking. The coking of heavy petroleum residues permits the production of coke with different types of microstructure. That is of great industrial importance. The quality and yield of petroleum coke depend on the composition of the raw materials employed and on the coking conditions. This article describes the functional structure of software that employs statistical analysis of experimental data in the synthesis of polynomial regression models capable of predicting how the yield and microstructure of petroleum coke depend on the coking conditions and the properties of the initial hydrocarbons. The models are verified by using the Fisher test for statistical analysis of calculation results and data derived from a series of experiments conducted by means of the pilot coking plant in the Department of Petrochemical and Coal-Chemical Production Technology at Saint Petersburg State Institute of Technology. Testing at the Institute’s world-class laboratory has confirmed the performance of the software and its suitability for research at oil refineries with coking systems.