Predictive modeling of Demulsification Efficiency in Brazilian Crude Oil

During oil processing, chemical demulsification plays a crucial role in promoting the separation of produced water. However, the laboratory evaluation of demulsifier efficiency under different operational conditions is predominantly conducted through trial and error, leading to high consumption of time and resources for drawing conclusive insights on the process. Aiming to overcome this issue and introduce a systematic approach to demulsification studies, this work employed Design of Experiments (DOE) as a methodological tool to evaluate the efficiency of a commercial demulsifier formulation, composed of 67% alkoxylated resin, 17% ethylene oxide/propylene oxide (EO/PO) block copolymer, and 17% xylene as solvent. The process temperature and demulsifier concentration were the independent variables studied, while demulsification efficiency (%DE) was defined as the response variable. Demulsification efficiency was assessed using the Bottle Test, conducted over a 60-minute period with a heavy oil sample from the Sergipe Basin – Brazil. The laboratory results allowed the development of a predictive and statistically significant mathematical model of the response variable. In the experimental domain investigated, temperature was the statistically significant variable, and the best operational condition was identified at 75 ºC and 300 ppm formulation concentration, achieving 90.34% efficiency.