Experimental investigation and thermodynamic modeling of asphaltene precipitation during pressure depletion and gas injection at HPHT conditions in live oil using PC-SAFT EoS

Abstract

Due to the high costs associated with asphaltene precipitation and deposition, the study of this flow assurance problem has always been a major issue in the petroleum industry. We studied the asphaltene phase behavior in four Iranian live oil samples. First, asphaltene onset pressure and the amounts of the precipitated asphaltene during pressure depletion and gas injection at high-pressure high-temperature (HPHT) conditions were measured, experimentally. Then, thermodynamic modeling of asphaltene precipitation was performed using the perturbed chain form of the statistical associating fluid theory (PC-SAFT). In previous researches, PC-SAFT equation of state (EoS) along with the standard method of characterization was applied to light crude oil having low asphaltene content only, while in this work, three of the samples were heavy oil with high asphaltene content. To perform the thermodynamic modeling, the oil samples were characterized by two different methods. In the first method, the adjustable parameters of the model, which are the aromaticity of aromatics + resins and the aromaticity of asphaltene, were adjusted separately. In the second method, both of the parameters of the model were tuned simultaneously, as one of the novelties of this research. Then the amounts of precipitated asphaltene were modeled using the aforementioned characterization methods, and the results were compared to the experimental data. According to the results, simultaneous adjustment of the model parameters in the heavy oil samples could improve the average absolute deviation (AAD) by 5% - 12%. However, no remarkable improvement was observed for the light oil sample. For two of the oil samples, the proposed model showed a good agreement with the experimental data as the calculated AAD were 14% and 26%. For the other oil samples, the trend of the experimental data was predicted by the model, but the values of the experimental data could not be matched. We were able to solve this problem by introducing a confidence factor for experimentally measured data. By considering this weight factor, in one of the samples, the AAD was reduced by about 70% and reached 24%, and in other oil sample, the error decreased by 90% and reached 31%.