Abstract
Most methods for viscosity calculation of heavy hydrocarbon mixtures do not consider the temperature effect. The temperature effect defines the characteristics of heavy hydrocarbon; it represents the molecular composition of heavy hydrocarbon. Hence, in this paper, a novel iterative method of predicting the viscosity of a mixture of heavy hydrocarbon is developed by considering the viscosity-temperature behavior of the heavy hydrocarbon from experimental samples of processed heavy hydrocarbons freshly obtained from a refinery. This method is highly beneficial for assessing viscosity arising from the blending of various heavy oils. The samples are graded by a viscosity index measured by ASTM D2270 (ASTM International, West Conshohocken, PA, 2010, 1-6), where the viscosity indices equal to 128, 132, and 138 are the lower (Grade C), normal (Grade B), and higher grades (Grade A), respectively. The viscosity of the samples is measured by ASTM D445 (ASTM International, West Conshohocken, PA, 2001, 1-10). These viscosities are differentiated based on the boiling point cuts produced by the product fractionation units ranging from 4 to 6 cSt and the boiling points within the range of 400 to 500 °C. The proposed iterative model is developed based on the viscosity-temperature relation established by MacCoull (ASTM International, West Conshohocken, PA, 2021, 1-6; KuranoInt. J. Thermophys.1992, 13 (4), 643-657), with subsequent refinements grounded in empirical findings derived from experimental observations. The proposed model has been evaluated and compared to other established methods such as the linear, Arrhenius, Jones and Bingham, Kendal and Monroe, Reid, and Refutas models for samples mixed in binary combinations at varying ratios. Results show that the proposed iterative model offers extra flexibility and has better precision in terms of the viscosity measurement of the heavy oil mixtures compared to those of the other established models. It is observed that the linear, Kendal and Monroe, and Arrhenius models overestimated the viscosities of the heavy oil mixtures. In contrast, the Reid and the Jones and Bingham models underestimated the heavy oil mixtures. Therefore, the proposed model is very useful especially for very stringent viscosity applications and mixtures of hydrocarbons with a bigger difference of viscosity. The proposed model has been validated in real industrial base oil plants, which include a wider range of mixing ratios. From the results, it has been shown that the proposed model remains valid within the testing condition range during the real plant validation run.