Publications

The Delta field is mature and currently facing a significantly declining production rate since it was at peak production in 1992 with 290,000 barrels of oil per day through a 1-million-barrel steam flood injected per day. Onward, some operational challenges continue to increase. Furthermore, the integrated Delta Field facility system generates more complications in operations such as production measurement, treatment, steam generation, steam distribution injection, water disposal, and its injection network. Steam distribution dynamic operations and well-stimulation activities, for example, cause swings and changes in temperature-fluid characteristics. Furthermore, heavy oil from the Delta Field is hard to treat because of its high viscosity, emulsion stabilizers such as asphaltenes, organic acids, and generated sand. Simplifying the situation, managing BSW (Basic Sediment & Water) has become more difficult as operations have become more complicated, and the area has been producing for more than 30 years since peak production. The BSW prediction tool development utilized existing data infrastructure to assist processing facilities in managing BSW by taking necessary activities more effectively than in previous practices. When there was no online prediction tool, operators responded after the BSW increased. The BSW prediction tool can predict BS&W values with an outstanding R2 value (0.945) by utilizing an AI (Artificial Intelligence) and Data Science (DS) model. Nevertheless, there are some challenges during data integration and preparation. The problem was data supply for training the model, data acquisition technique, and measurement capability. Fortunately, the final DS/AI Model can provide an early warning up to 2 days before a potential BSW upset. The prediction enhances safety performance, particularly in treating high-risk steam systems with immediate judgments, reducing errors, and increasing efficiency. The average percentage of upset case frequency reduces from 15.08% to 1.05% per year. In conclusion, this technology application results in better BS&W management. This paper will offer an outline of recent collaborative efforts for using data science and machine learning-artificial intelligence as a solution alternative to manage BSW quality difficulties associated with crude oil process separation challenges in Indonesia's steam flood operation Delta Field. The project team concludes that DS-AI adoption can be a promising alternative for unlocking further optimization potential in oil and gas processing by leveraging enormous growing data.