Measurement and Modeling of Asphaltene Precipitation from Crude Oil Blends

详细信息    查看全文

• 作者：Asok Kumar Tharanivasan ; William Y. Svrcek ; Harvey W. Yarranton ; Shawn D. Taylor ; Daniel Merino-Garcia ; Parviz M. Rahimi
• 刊名：Energy & Fuels
• 出版年：2009
• 出版时间：August 20, 2009
• 年：2009
• 卷：23
• 期：8
• 页码：3971-3980
• 全文大小：462K
• 年卷期：v.23,no.8(August 20, 2009)
• ISSN：1520-5029

文摘

A previously developed regular solution model was adapted to predict the onset and amount of asphaltene precipitation from crude oil blends diluted with pure n-alkanes or a mixture of toluene and n-heptane. Tests were conducted on nine different crude oils, a gas oil, and their blends. Oils and blends were characterized in terms of SARA (saturates, aromatics, resins, and asphaltenes) fractions. The mass fraction of each SARA fraction in the blends was confirmed as a weight average of the respective fraction in constituent oils. Asphaltenes were subdivided into fractions based on the gamma function to account for the distribution of aggregates resulting from self-association. To model the asphaltene onset and yield, liquid−liquid equilibrium was assumed between a heavy (asphaltenic) and a light (nonasphaltenic) phase. The distribution of asphaltenes in unblended crude oils was determined by fitting its asphaltene yield data when diluted with n-heptane. The fitting parameter in the model was the average aggregation number of asphaltenes in the source oils. Two approaches were tested to calculate the distribution of asphaltenes in crude oil blends. In the first approach, asphaltenes were assumed to interact, and the final molar mass distribution was determined from gamma function using the average aggregation number of constituent oils. In the second approach, no interaction was assumed, and the final distribution was calculated as a sum of the individual distributions. It was found that the best predictions of the onset and yield data were obtained by using the second approach. The mass fraction of n-heptane required to initiate precipitation was predicted with an average absolute deviation of 0.53% or less in all cases.